- BWG Session Redundancy and High Availability Infrastructure
- Subscriber Management
- DHCP and AAA
- Dynamic Synchronization
- Configuring Session Redundancy
- Configuration Example
- Authentication
- Accounting
- Subscriber IP Address
- QoS
- Statistics and Counters
- BWG Load Balancing
- Data Path and GRE
- Version Control
- Limitations
- Switchover
Configuring the Cisco Broadband Wireless Gateway
This feature module explains and discusses the feature set for the Cisco Broadband Wireless Gateway (BWG). Additionally, this feature module explains how to configure those features, and provides sample configurations when appropriate.
This chapter contains information on the following features:
•Ethernet Convergence Sublayer (CS)
–Ethernet CS—Data and Control Both on R6
–AAA Access for Unauthenticated Subscriber
•Support for MS/Host with Statically Assigned IP
–Network Admission of an Authenticated User
–Support of Un-Authenticated User
•IP Address Allocation Using DHCP
–Configuring IP Address Allocation
–Support of Multiple Hosts Behind a SS
•Service Flow Creation and Management
–Multiple Service Flow Creation
–Mapping of Service Flows to DiffServ Classes
–Configuring Service Flows on the BWG
–Configuring Service Flow Packet Classification
•AAA Accounting Start-Stop-Interim
•Keepalive Support for R6 Interface
–BWG Session Redundancy and High Availability Infrastructure
–Configuring Session Redundancy
–QoS
–Configuring Cisco IOS SLB for Load Balancing
–Configuring the BWG for Load Balancing
•Mobile to Mobile Traffic Steering New
Ethernet Convergence Sublayer (CS)
The Wimax Ethernet Convergence Sublayer (CS) allows a WiMAX network to provide Ethernet Service directly to customers. In comparison with the IP CS, it allows IEEE 802.3 frames (carrying higher layer IP datagrams) to be encapsulated in the 802.16 PDUs. In the Cisco BWG 1.1 release, only two options of Ethernet CS are implemented: BS local switch (enterprise customers), and L2-L3 bridging conversion (residential customers),
Note L2-L2 bridging on the BWG is not currently available.
As a part of Ethernet CS requirement, the BWG will support statically assigned IP addresses for the CPE/MS/host. This includes the following sub-features for the BWG:
•Enforce a maximum number (8) of active hosts per subscriber/CPE
•Auto-learning of L2/L3 details of the hosts through ARP, or any uplink packet from the host.
•Static host IP verification with Framed-Route from AAA
•A mechanism to age/bump out the idle host.
Ethernet Convergence Sublayer (CS)—R6 Control Only
In the Cisco 12.4(15)XL1 release, the WiMAX R6 interface between the BWG and Base Station (BS) is used for signaling only; the BWG instructs the BS to switch traffic locally. In the absence of GRE encapsulation, the BWG does not receive the bearer traffic from the BS. Instead, the BS must switch or forward the ethernet frames to an externally connected L2 Switch.
Note The R6 Control Plane Only feature should only be used for the Ethernet CS in the Cisco BWG R1.1 release.
Since the data packets no longer go through the BWG in this configuration, the BWG needs the following adjustments:
•Session Idle Timer: The BWG needs to ensure that a session remains open even if no packet is received or sent for the subscriber.
•When AAA sets the session timer to 0, it indicates it is infinite.
•Accounting: The BWG only performs "Accounting Start" and "Accounting Stop", which corresponds to the service flow creation and deletion. Interim periodic accounting updates are not performed since BWG does not receive bearer traffic in this scenario.
In cases where the BWG does not receive the de-registration request from BS, the absolute session timer is used to de-register the MS on the BWG in order to remove hanging sessions.
The following steps illustrate the sequence of events for the R6 Control Only feature:
Step 1 SS sends a registration message to the BS. The BS forwards it to the BWG.
Step 2 The BWG asks for profile from the AAA server with user_name MSID@proxy-realm. Proxy-realm is configured in the user-group.
Step 3 AAA sends the SLA profile back with the following information:
–SS is identified as a business based on the SLA profile (which is configured as encap-type none on the BWG); in this case that traffic is locally switched at the BS identifies this as an Enterprise.
–The VLAN ID (unique for every business SS connected to same BS). For example, the VLAN ID = 250.
Step 4 The BWG sends the Service Flow profiles info to the BS, along with VLAN ID and the Packet Classifications rules. The BS further sends the uplink Packet Classification rules to the SS.
Step 5 Upstream traffic is initiated: this assumes that the CPE router will send uplink .1q tagged traffic, with following VLANs:
–VLAN 10 (Sales)
–VLAN 20 (Voice)
Step 6 If the PCR is such that the SS will put traffic from VLAN 10 into Service Flow 1, this service flow type is BE. If the PCR is such that the SS will put traffic from VLAN 20 into Service Flow 2, this service flow type is UGS. The SS sends traffic to the BS.
Step 7 The BS assigns another .1q tag on incoming traffic for enterprise x (for example VLAN ID = 250). The inner tag is left unchanged. Traffic is switched to the L2 network, and does not go to the BWG.
Step 8 For downstream traffic, the following occurs:
–The BS receives traffic from the L2 switched network.
–Enterprise traffic does not come through BWG.
–The PCR will tell the BS on what SF traffic has to be sent.
–The BS strips the entire outer tag and forwards the remaining packet traffic over the air by SF1 (BE) or SF2 (UGS).
Step 9 The SS receives the traffic and forwards it to the switch. The inner tag is left unchanged.
Additionally, in this scenario the BS, rather than the BWG, terminates the uplink/downlink service flows. The BWG sends the downlink classifiers to the BS so that the BS can choose a proper downlink service flow for a packet. The BS needs the downlink classifiers to select the 802.16e air link connection ID/service flow. The BWG also signals the BS which VLAN tag to use for uplink traffic.
While the BWG design allows the flexibility to perform BS local switching on a flow-by-flow basis, for Release 1.1 all service flows should be configured to be either collectively BS-local switched, or BWG switched for a particular subscriber.
Configuring R6 Control Only
This section provides information on how to configure the R6 Control Only feature on the Cisco BWG. To enable R6 Control Only, perform the following tasks:
Note For BWG Release 1.1, the same vlan should be configured in the same SLA profile.
Note When the vlan is downloaded from AAA, it will overwrite the vlans locally configured for all service flows.
The BWG controls the BS's local switching through Data Path Encapsulation Type (NONE) and Data Path ID (Priority + VLAN ID) in the R6 DP Registration Request message. Note that the VLAN ID defined here could be overwritten from AAA. The VLAN Priority (the 3 most significant bits in VLAN tag) comes from DSCP/Precedence defined for the service flow. If DSCP/Precedence is not locally defined, it is calculated based on WiMAX QoS Data Delivery Service Type used for the service flow.
Ethernet CS—Data and Control Both on R6
To align with the WiMAX Forum NWG standard, Cisco's R6 interface supports Ethernet CS with data and control. As noted earlier, only the L2-L3 bridging option for the BWG is supported in this release.
The L2 uplink traffic comes from the host behind the CPE/MS, and these packets are encapsulated 802.16 PDUs and sent to the BS by the CPE through the R1 interface. The BS then encapsulates the Ethernet frame into the GRE packet (the GRE tunnel was established during R6 signaling exchange), and sends it to the BWG. The BWG receives the GRE packet through the R6 data path. After stripping the GRE header and L2 header, the inner IP packet is forwarded to a proper interface configured in the intended VRF.
When the BWG receives the downlink packets, it finds the related host where the L2 information related to the particular host is saved. The L2 information, along with the L3 information carried in the packet, is used to compare with the configured classifiers. As a result, a proper service flow is selected. Once a service flow is selected, the saved L2 information for the host is used to encapsulate the received IP packets.
The host's L2/L3 information is dynamically learned by the BWG through the following mechanisms:
•DHCP procedure from host
•ARP Request from host
•Any uplink packet from host
VLAN to VRF Mapping for L2-L3 Bridging
This feature allows that the uplink L2 traffic with a particular VLAN ID is mapped to a VRF routing domain to forward the IP packets. On the other hand, the downlink IP traffic from a particular VRF will be sent to the MS encapsulated with the same VLAN ID. Note that BWG is designed to support at most one VLAN ID for each host behind MS.
Packet Fragmentation Handling
The virtual template interface on the BWG must be configured with a MTU so that no GRE fragmentation is needed for the R6 data path. The preferable MTU value is less than 1440.
For downlink packets larger than the MTU configured for the virtual template interface, they will be fragmented by IOS, and in doing so, IOS expects to clear the DF bits in the original packets. The BWG receives two IP packets in this scenario. The two IP packets are GRE encapsulated separately, and sent to the BS. The BS transparently passes the two packets to the host where they are assembled and delivered to an application.
For a large uplink packet, the BS clears the DF bits, fragments it into two packets, and sends them to the BWG. The BWG does not reassemble the packets, but forwards them separately.
DHCP Option 82 Enhancement
In order to allow the BWG to build the L2 header for downlink DHCP packets, the entire L2 header is coded in the Option 82, which is reflected back from the DHCP server. This sub-option only applies between the BWG and DHCP server.
DSCP Calculation/Marking/Signaling
For downlink traffic, the DSCP marking is performed by the BWG. The outer IP DSCP value is obtained in the order of precedence from:
•DSCP/Precedence set for the flow.
•Calculated from SF's Data Delivery Service Type as shown in the following table.
|
|
|
UGS |
46 |
|
ERT-VR |
38 |
|
RT-VR |
30 |
|
NRT-VR |
22 |
|
BE |
0 |
Note The inner IP's DSCP value is no longer used for marking the outer IP's DSCP.
Note The inner IP value is also marked.
Note The SF QoS to DSCP mapping table is used only when the DSCP is not configured for the SF.
For uplink traffic, the BWG obtains the DSCP value in the same way as for the downlink service flow. The uplink DSCP value is signaled to the BS in the Data Path Info TLV.
Consider that DSCP marking of inner user (uplink) packets cannot be trusted; therefore, the same uplink DSCP value signaled to BS for a service flow is also used to re-mark the R3 uplink packets.
Ethernet Frame Support
The BWG can support Ethernet II, LLC (802.2) and Ethernet SNAP frame types for different hosts. However, one particular host must use the same frame type during its session. In other words, one particular host can only transmit packets with the same Ethernet frame type.
Ethernet VLAN Support
Ethernet packets with a VLAN tag (specifically 802.1Q and Q-in-Q) are also supported.
Ethernet FCS (CRC)
There should be no CRC for both uplink and downlink packets over GRE towards the BWG.
Limitations
The following CS Ethernet limitations exist in Cisco BWG release 1.1
•Layer 2 broadcast/multicast packets from the MS other than ARP and DHCP are dropped at the BWG.
CPE Management
The CPE Management feature allows the BWG to use AAA to centralize the management of subscribers/CPEs for the overall Cisco WiMax solution. Considering the potential congestion and multiple AAA proxies in a real deployment, it can take up to 10 seconds to recieve the RADIUS Access Accept message from AAA. Therefore, the related R6 protocol state machine should be designed to tolerate this AAA response delay.
The unauthenticated CPE management function has been moved to AAA and the BWG. Specifically, it includes the following features:
•User Domain Group Re-assignment
•Service-Level-Agreement (SLA)
•VLAN ID
•Mobile Capability (non-nomadic with a home BS list, and nomadic)
•Static IP Allowed (### whether static IP is allowed for the CPE)
•CPE Type
•CPE Settings
•CPE Auto-Provisioning
AAA Access for Unauthenticated Subscriber
For Base station/CPE that does not support EAP-based authentication, the BWG provides a PPP/PAP method of authentication using RADIUS. Such CPE generally get classified as Unauthenticated CPE. And for un-authenticated users, we do not get the user name from the CPE.
In this case, the user name, realm and password will be formed based on the following CLI:
wimax agw user group-list wimax
user-group unauthenticated
aaa authentication method-list xxxx
proxy realm sprint.com passwd ciscoway
sla profile-name silver
!
Note The aaa authentication method-list xxxx command indicates if the RADIUS Access Request is initiated from the BWG for the group. If the CLI is not configured, AAA query is not required. The respective authentication method list is type PPP, which enables the BWG to perform PAP-based authentication.
Note Reauthentication for PAP users is not supported. The session will be de-registered if reauthentication is attempted by a CPE/MS.
Note The proxy realm sprint.com password ciscoway command instructs the BWG how to populate the RADIUS Access Request message. If configured, the user name will be constructed as mac@realm (for example, mac @sprint.com). If the realm is not configured, the user name will be mac. If not configured, cisco is used as the password. The method list configured using aaa authentication method-list xxxx is not used by the BWG in Access-Requests, since PAP authentication uses default method list configured globally by the aaa authentication ppp default command.
Note These two CLIs are applicable for other user groups (EAP users) as well. However, configuring proxy-realm for EAP users serve no purpose.
The reply from the AAA server contains the user's real domain name, which is used for selecting a local user group. The above scheme should not break EAP-authenticated users. In other words, the BWG should allow EAP and non-EAP authenticated users to coexist. For authenticated users, the user name is acquired from CPE through EAP identity request. EAP uses NAI in an Access request to the AAA. If the response from the AAA includes the Service Level Agreement (SLA) Profile Name and the User Domain Name for EAP users, the result from the AAA will override those determined earlier.
New AAA Attributes
To support the CPE management, the following new AAA attributes have been introduced. These new attributes may be returned in the RADIUS Access-Accept message. These attributes are all optional, and are AVPs under cisco_vsa.
Note The above AAA attributes are intended for PAP users, but they should work equally well for EAP users.
Configuring the SLA Profile
To configure the Service Level Agreement on the BWG, perform the following tasks"
In the above configuration, the user group's SLA profile acts as a default. The AAA return SLA profile name should take precedence.
Multiple SLA Profile Support Per Subscriber
Prior to this release, SLA profile configuration on the BWG allowed WiMax service providers who defined the service packages efficiently to categorize service offerings (flow types) in a reasonable number of profiles.
For example, the following flows (flow-voice, flow-data, flow-video, flow-hd-video, flow-premium-voice, flow-premium-data) could be defined in 2 packages and sold to subscribers on a package basis.
SLA-regular
•flow-voice,
•flow-data
SLA-premium
•flow-premium-voice,
•flow-premium-data
•flow-hd-video
However, if service providers wanted to offer services in all possible combinations from the above list of flows, they would have to configure as many SLA profiles on BWG.
This feature allows service providers to configure SLA profiles similar to following and package the services to customers in any desirable combination.
SLA-regular data
•Flow-data
SLA-regular voice
•Flow-voice,
SLA premium video
•Flow-hd-video
SLA-premium data flow
•Flow-premium-data
The BWG supports flow creation based on multiple SLA profile names received from AAA. This is in addition to the existing support for a single SLA profile name. AAA can send a list of SLA profile names, separated either by comma, space or semicolon.
The BWG will parse the list to choose valid, configured SLA profiles to create the service flows for session.
The BWG uses the following sequence of events to create the flows:
•BWG selects only configured SLA profile names from list.
•BWG selects only the unique flows from across the list of SLA profile names. For example, it takes "Union" across all the flows defined under received SLA profile names (with the exception of ISF).
•The current implementation needs only one ISF for successful flow creation. As a result, the BWG selects the ISF based on the order of SLA profile name in list of SLA profile names received from AAA. Even if different ISF profiles are configured for SLA profiles received, it selects the ISF from the first valid, configured SLA profile from the received list.
•The BWG discards all the remaining SLA profiles and SF profiles if the maximum allowed service flows limit is reached.
•The BWG does not check the validity (for example, if SF is configured correctly or not) of the SF under the SLA profile at the time it parses the list received from AAA. As a result, if the SF is found to be invalid or un-configured when the BWG is about to open that SF, then that SF gets dropped and BWG continues with rest of the SF profiles.
•If ISF is not present in SLA profiles coming from the AAA, the session is cleared.
Here is an example of the configuration:
wimax agw sla profile silver
service-flow pre-defined isf profile isf
wimax agw sla profile platinum
service-flow pre-defined secondary 1 profile sec4
wimax agw sla profile gold
service-flow pre-defined isf profile isf2
service-flow pre-defined secondary 1 profile sec
service-flow pre-defined secondary 2 profile sec2
service-flow pre-defined secondary 3 profile sec3
If the SLA list received from AAA is "silver, gold, or platinum" then the session's SLA profile name is set to "silver, or gold", and the following four flows are created on the BWG:
•isf
•sec4
•sec
•sec2
If the above configuration SLA list received from AAA is "platinum, unconfigured, gold", then the profile name of the session is set to "platinum, gold", and the following four flows are created:
•isf2
•sec
•sec2
•sec3
The "unconfigured" SLA is discarded by the BWG.
Here is another example:
wimax agw sla profile silver
service-flow pre-defined isf profile isf
wimax agw sla profile platinum
service-flow pre-defined secondary 1 profile sec
wimax agw sla profile gold
service-flow pre-defined isf profile isf2
service-flow pre-defined secondary 1 profile sec
service-flow pre-defined secondary 2 profile sec2
service-flow pre-defined secondary 3 profile sec3
If the SLA list received from AAA is "platinum, unconfigured, gold", then the profile name of session will get set to "platinum, gold", and the following four flows are created:
•isf2
•sec (based on the order in which the SLAs were received, this flow was taken from SLA platinum)
•sec2
•sec3
The "unconfigured" SLA is discarded by the BWG.
In the absence of SLA from AAA, the BWG creates flows according to the SLA profile defined under the session's user group. The BWG still supports the single SLA profile sent by AAA.
User Auto-Provisioning
There are occasions when a user is admitted into the network for a short while even if AAA does not have provisioning for him/her. To enable this feature, the related unauthenticated group should be properly configured. When enabled, the session timer in the user group should be configured to a small value so that the user's free use of the network is limited.
To configure auto-provisioning, perform the following task:
Note Auto-provisioning is not supported for EAP users. It does not take effect when configured with any user group other than the unauthenticated.
Note Auto-provisioning for hosts with static IP and IPCS is not supported.
Support for MS/Host with Statically Assigned IP
The BWG's DHCP mechanism for Ethernet CS is similar to that of IPCS except for an additional sub-option (L2 header) for option 82. Through the DHCP procedure, the BWG is able to capture the host's L2 header (including frame type) and IP address. In addition to the DHCP mechanism, the BWG supports the MS/host with a statically assigned IP address.
The static IP address is handled differently depending if IPCS or Ethernet CS is used.
IP CS
Authenticated Subscriber
For authenticated subscribers, the Framed-Route from AAA response can be used for downlink traffic routing. This feature is already supported in the BWG.
The BWG learns about static hosts in IP CS case through uplink data packets. The BWG does not have any L2 header information, so these static hosts are created without MAC IDs. The aging mechanism described below in the Ethernet CS section also applies to IP CS static hosts.
Note Unauthenticated subscribers are not supported.
Ethernet CS
There are two mechanisms for the BWG to learn the L2 header info and IP address from the statically configured host. The BWG creates its host entry (indexed by table ID—VRF and IP address) and L3 routing entry. The static hosts created in the BWG are limited to 8 active hosts per subscribers. However, a 9th host can be accepted if one of the existing static hosts' idle periods has exceeded a threshold value. In this case, the static host which has been idle the longest is bumped out by the new arrival. The host idle threshold is 75% of session idle timer. A static idle host is removed in the BWG only when a new host is detected by the BWG. DHCP hosts are not aged out.
All of the static IP that are learned are verified with the Framed-Route from the AAA. Currently the BWG supports one framed-route per CPE.
Note You should not disable the route aggregation feature for the user-group. This will prevent too many static host routes to be created or deleted from the routing table, which degrades the performance.
ARP Request from Host
The BWG can always intercept the ARP request packet to learn the host's IP and L2 header info (including frame type). The creation of a static host is subjected to the limits of active hosts per subscriber. If the BWG cannot admit the host, the ARP request will not receive a reply.
The BWG handles ARP packets similar to those for DHCP messages. The ARP reply packet is returned to the same service flow where the request originated. The BWG's MAC address (interface on which the ARP request was received) is used in the ARP reply message to the MS.
The BWG also implements an ARP-rate processing limiting feature to prevent itself from being overwhelmed. ARP packets are dropped if their rate of arrival exceeds what the BWG can handle. Currently the throttling mechanism on the BWG allows one upstream ARP request every 5 seconds.
To be a true ARP proxy, the BWG must be able to intercept a pass-through ARP request packet. These packets have a Target Protocol Address (TPA) different from BWG's IP. In the ARP reply packet, the TPA in the request packet is used.
Uplink Packet from Host
When an uplink packet is received by the BWG without its host entry in the Cisco Express Forwarding (CEF) path, it is rerouted to the process path where it is processed. As a result, a new host may be created. If the host cannot be created (subject to the 8 active host limit), the packet is silently dropped.
The BWG implements a throttle mechanism to prevent rerouting of the packets to the process path from overwhelming the BWG. Currently the throttling mechanism on the BWG reroutes 1 packet every 5 seconds.
Limitations
•A host that was bumped out of the network is not able to receive downlink traffic without first sending some uplink traffic. When this occurs, it indicates that the CPE actually has more than 8 hosts behind it.
•Upstream gratuitous ARPs are dropped by the BWG, and the BWG does not learn from these packets.
Other Enhancements
To support the Cisco BWG Release 1.1 and above features, the BWG is enhanced in the areas of the R6 Protocol, MIB, statistics, SR and CLI.
R6 Protocol Enhancement
The existing R6 protocol is enhanced to accommodate the Ethernet CS requirements.
R6 Attachment Request
CS Capability
The CS capability in the Registration context in the message must be set to the proper value to indicate Ethernet CS by the BS or the BS simulator.
Note The CS Capability included in this message can indicate a single CS type, or a set of CS types with a bitmap supported by the MS. In any case, it shows the MS's capabilities. Please refer to 802.16 for the details CS capability definition.
R6 Attachment Response
This message is sent from the BWG to the BS. The changes to this message are equivalent to the R6 Attachment Request message.
The CS Capability bitmap in this message is encoded to indicate that the BWG supports both IPCS and Ethernet CS.
R6 Data Path Registration Request
This message is sent from the BWG to the BS for normal R6 Data Path setup. The MSINFO TLV contains the following sub-tlvs:
•Anchor GW/DPF ID = IPv4 address of the ASN gateway
•CPE Settings—(optional)
•SF INF
–SFID
–CS Type—Specifies IP CS, Eth CS, or VLAN CS for the flow
–Packet Classification Rule
Classifier Rule Priority
Ethernet Src MAC—(optional) for uplink SF, and are sent for downstream in case of control only scenario
Ethernet Src MAC Mask
Ethernet Dest MAC—(optional) for uplink SF only
Ethernet Dest MAC Mask
EtherType IP
VLAN-ID
VLAN Priority Range
–Data Path Info
Data Path ID—GRE Key or VLAN ID
Data Path Encap. Type
DSCP
The CS Type applies to the particular service flow. It is the intersection between the MS's indicated CS capabilities (in the Attachment Request), as well as the BWG's service flow CLI configuration. If both the CPE and the BWG support both IPCS and Ethernet CS, and the BWG configures both with the same precedence, IPCS is selected.
The Ethernet Src MAC and Ethernet Dest MAC address should be included if they are not configured as any in the corresponding BWG CLI configuration. New tag values conform to the Cisco R6 specification.
The Ethernet and VLAN related classifiers will not be signaled for an IPCS service flow. For Ethernet CS, the classifier VLAN-ID and VLAN Priority Range will be signaled to BS if they are locally configured.
The "Data Path ID" is an existing TLV. The VLAN Priority value (the most significant 3 bits of VLAN Tag) for the VLAN Tag comes from the DSCP/Precedence set for the flow, or derived from the QoS Data Delivery Service configured for the flow.
The "Data Path Encapsulation Type" TLV indicates if the encapsulation type should be used to transport the R6 bearer traffic.
0 = None
1 = GRE (Default)
2 = IP-in-IP
3 = VLAN
The tag value for "Data Path Encap Type" must conform to the Cisco R6 specification. For Ethernet CS R6 control-only scenario, this TLV should be set to "none". When "Data Path Encap. Type" is set to "none", the BS interprets the Data Path ID as a VLAN ID. In its absence, use the R6 data path with GRE to transport the data traffic.
The DSCP values signaled to the BS for uplink SF come from either the BWG locally configured for the SF, or, if it is not present, derived from the QoS Data Delivery Service configured for the flow.
In addition, the ISF Path Registration Request signals the newly defined CPE Settings TLV to the BS if it is available through AAA. The CPE settings are under the MS INFO TLV.
R6 Data Path Registration Response
This message is sent from the BS to the BWG during normal data path setup. The changes to the message are equivalent to the R6 Data Path Registration Request message. All the new TLVs defined for Data Path Registration Request messages are available for the Path Registration Response message.
Other R6 Changes
To be consistent, all the R6 signaling UDP packets have a DSCP value of 0x48. In addition, the UDP checksum is already calculated in Release 1.0 and above.
The BWG enforces a limit for the number of service flows to 4, and the number of active hosts to 8 per subscriber. If you attempt to exceed the limit the BWG will fail.
Note The BWG will not allow more than 4 service flows to be configured per SLA profile.
EAP Authentication
The BWG acts as an EAP relay and is agnostic to the EAP method. EAP transport is done between the BWG and the base station as a control exchange. The base station functions as an EAP-relay, converting from Pair-wise Master Key version 2 (PKMv2) to the EAP messages over to the BWG. The BWG is an EAP pass-through, and any key that generates EAP methods is supported in the system.
PKMv2 is used to perform over-the-air user authentication. PKMv2 transfers EAP over the IEEE 802.16 air interface between the MS and the base station. The base station relays the EAP messages to the Authenticator in the BWG. The AAA client on the Authenticator encapsulates the EAP message in AAA protocol packets, and forwards them through one (or more) AAA proxies to the AAA server in the CSN of the home NSP. In roaming scenarios, one (or more) AAA brokers with AAA proxies may exist between the Authenticator and the AAA server. All AAA sessions always exist between the Authenticator and AAA server, with optional AAA brokers providing a conduit for NAI realm-based routing.
Network Admission of an Authenticated User
The following series of events illustrates how the network admits an authenticated user.
1. The authenticator (in BWG) initiates EAP authentication procedure with MS after receipt of Pre-Attachment-Ack message from the Base Station.
2. The authenticator sends EAP Request/ Identity message over Authentication Relay protocol (AuthRelay-EAP-Transfer) to BS.
3. The BS relays the EAP Request/ Identity payload in the PKMv2 EAP-Transfer/ PKM-RSP message to the MS.
4. The MS responds with EAP Response/ Identity message providing NAI. This message is transferred to BS over PKMv2 EAP-Transfer/ PKM-REQ message.
5. The BS relays EAP payload received in PKMv2 EAP-Transfer to the authenticator over Authentication Relay protocol (AuthRelay-EAP-Transfer message).
6. The EAP payload is forwarded to MS' Home AAA server via Visited AAA server (authenticator analyzes the provided NAI for resolving the Home-AAA server location). Authenticator sends EAP Request/ Identity message over Authentication Relay protocol (AuthRelay-EAP-Transfer) to BS.
7. In order to deliver EAP payload received from BS, to AAA server, authenticator forwards EAP message through the collocated AAA client using RADIUS Access-Request message (EAP payload is encapsulated into RADIUS "EAP message" attribute(s).
8. The EAP authentication process (tunneling EAP authentication method) is performed between the MS and the authentication server through the authenticator in the BWG.
9. The EAP payload returned from the AAA server in a RADIUS Access-Challenge message is transferred to the base station in an AuthRelay-EAP-Transfer message. There may be multiple EAP message exchanges between the EAP supplicant, located at the Mobile Subscriber Station, and the EAP Authentication Server, located at the AAA server.
10. The authenticator sends the Key Change Directive message to the base station to indicate completion of the EAP authentication process. The key is computed by BWG using the Master Secret Key (MSK) it received from AAA (in an Access Accept). The Key Change Directive contains the MSINFO TLV with the AK Context sub-TLV, and also the EAP Payload TLV indicating EAP success.
11. In the case of an authentication failure indication is received from the AAA server the subscriber is de-registered from the network using the Normal Mode Network-Initiated Network Exit procedure.
12. The base station acknowledges receipt of Key Change Directive message with a Key Change Acknowledgement message.
13. The base station sends the result of authentication to the Mobile Subscriber Station using a PKMv2 EAP-Transfer message.
Support of Un-Authenticated User
Support of un-authenticated users is required in the following scenarios, and can be used for pre-paid systems, or emergency calls.
•The Mobile Subscriber (MS) can choose to indicate NULL Authentication. This may be a specific type of MS, such as an MS that is limited to emergency calling. This type of MS will indicate NULL Authentication support in the SBC_REQ. The BS relays this through the NetEntry MS State Change Request to the BWG.
•Based on local policy, the BWG can choose to skip authentication, and allow a subscriber to enter the network.
•When the BWG is configured to enable NULL Authentication using the CLI, any Subscriber Station (SS)/MSS requesting NULL authentication will be mapped to a NULL-AUTH user group. DHCP requests from these SS/MSS will only be sent to the configured DHCP server. This enables the operator to control address allocation to the unauthenticated users, as well as apply any restrictions for such users. In addition, Access Control Lists may be configured that would restrict the traffic from the SS/MSS only to certain destinations.
Cisco's BWG Release 1.1 supports EAP authentication. If the BS/CPE does not support EAP authentication then the CPE gets categorized as unauthenticated. The BWG supports a very basic PAP type of authentication for such CPE where the CPE is authenticated based on its MAC ID and password. The RADIUS server should be pre-configured to provision the CPE.
Configuring AAA to Enable PAP Authentication on BWG
To configure the BWG to enable PPP/PAP authentication, perform the following task in global configuration mode:
Here is an example:
router(config)#aaa authentication ppp default group radius
Here is sample show output to verify the configuration
Router#show running-config | include aaa
aaa new-model
aaa authentication ppp default group radius
aaa authorization config-commands
aaa authorization network default group radius local
aaa authorization configuration default group radius
aaa accounting update periodic 1
aaa accounting network agw start-stop group radius
aaa session-id common
AAA Access for CPE Using PAP Authentication
For the BWG to initiate PAP Access-Request to RADIUS server for an unauthenticated user-group, perform the following task:
Here is an example configuration:
router#wimax agw user group-list wimax
user-group unauthenticated
aaa authentication method-list xxx
sla profile-name gold
timeout idle 100000
!
Proxy Realm for CPE Using PAP Authentication
For un-authenticated users, the BWG does not receive the user name from the CPE. In this case, the BWG provide a mechanism to specify proxy realm and password.
To configure the user name, realm and password, perform the following task:
Here is an example of the configuration:
router#wimax agw user group-list wimax
user-group unauthenticated
aaa authentication method-list agw
proxy realm cisco.com password cisco
sla profile-name gold
timeout idle 100000
!
In this configuration, the BWG sends and Access-Request to the RADIUS server with username set to MACID@cisco.com and password set to cisco.
If the proxy realm is not configured, the BWG sends an Access-request to the RADIUS server with username set to MACID and password set to cisco (the default password).
Auto-Provisioning for Unauthenticated CPE Using PAP Authentication
Auto-provisioning provides flexibility, and when configured on the BWG, allows CPE the network entry even if AAA does not have provisioning for that CPE.
Note Auto-provisioning is limited to CPE using PAP authentication.
To configure the unauthenticated user group for auto-provisioning, perform the following task:
Here is an example of the configuration:
router(config)#wimax agw user group-list wimax
user-group unauthenticated
aaa authentication method-list agw
proxy realm cisco.com password cisco
sla profile-name gold
timeout idle 100000
user auto-provisioning
Configuring Authentication
This section provides information on how to configure authentication and authorization on the Cisco BWG. To enable authenticated calls between the BWG and a subscriber, perform the following tasks on the BWG:
•Configuring AAA for Accounting Types
Configuring AAA for Accounting Types
To configure accounting types on the BWG, perform the following tasks:
Configuring Authorization
To configure authorization on the BWG, perform the following task:
Configuring Authentication
To configure authentication on the BWG, perform the following task:
RADIUS Server
To configure the RADIUS server host on the BWG, perform the following task:
Configuring User Groups
To configure a user group on the BWG, perform the following tasks:
Note AAA server group can be linked with the method list configurations so that different AAA servers can be configured, and thereby map to different user-groups.
Verifying the Configuration
The authentication method of a subscriber displays whether the call was authenticated with EAP, or unauthenticated for the respective user group (any, unauthenticated, domain specific).
For an authenticated call, the Auth Policy and AK Context is also displayed.
To verify your authentication configuration, use the following commands:
|
|
|
---|---|---|
Step 1 |
router# show wimax agw subscriber msid |
Displays subscriber authentication information. |
Configuration Examples
Here is sample output for subscriber information for an unauthenticated call:
Router>sh wimax agw subscriber msid 1000.0003.0000
Connection time 000:01:05
Auth policy 0X0(0)
Subscriber address 2.2.0.9, type IPv4, organization IETF
Subscriber address method Dynamic, source DHCP relay
Subscriber address assigned on flow downlink ID 17
Subscriber address prefix len allocated 32, aggregate 32
Subscriber address traffic sent 0 packets, 0 bytes
Subscriber address traffic received 0 packets, 0 bytes
Subscriber address DHCP XID 2391, server 0.0.0.0, htype 1
Subscriber address DHCP client ID 1000.0003.0000, length 6
Subscriber address DHCP Refresh time 86400 seconds
Number of sessions 1
Session details:
FSM in state Ready(7) on last event Rx Attach Ack(14)
Authentication method unauthenticated
Associated user group **unauthenticated**
Signalling address local 2.2.2.2, remote 10.1.1.82
Signalling UDP port local 2231, remote 2231
Idle for inbound 00:01:10, outbound 00:01:10
Ingress Address filtering 0 packets, 0 bytes
Number of flows 1
Flow details ISF(0)
FSM in state SF Ready(4) on last event Up(1)
Transaction ID used 0X8001(32769)
Data ID local 0x9(9), remote 0x2(2)
Data address local 2.2.2.2, remote 10.1.1.82
Data traffic sent 2 packets, 656 bytes
Data traffic received 2 packets, 1208 bytes
Accounting last record sent Interim(3)
Idle for inbound 00:01:10, outbound 00:01:10
Service Flow information Downlink:
Identifier 17
QoS information:
Data-delivery-service real-time-variable-rate
Minimum traffic-rate-reserved 4, Maximum latency 1
Here is sample output for subscriber information for an authenticated call:
Router>sh wimax agw subscriber msid 1000.0002.0001MSID 1000.0002.0001
Connection time 000:01:08
Auth policy 0X12(18), Single-EAP, CMAC
AK Ctx method C-MAC(1), Lifetime 65535
AK Ctx Seq No. AK 0, PMK 0
AK Ctx C-MAC key count 1
Number of TIDs 1
TID Key 10.1.1.82/2.2.2.2/1000.0002.0001
Peer TID 0X4(4)
FT MS State Change(9), MT Attachment Request(8)
Our TID 0x8004(32772)
Subscriber address 2.2.0.8, type IPv4, organization IETF
Subscriber address method Dynamic, source DHCP relay
....
Subscriber address DHCP Refresh time 86400 seconds
Number of sessions 1
Session details:
FSM in state Ready(7) on last event Rx Attach Ack(14)
Username eap-md5-u@eap-md5.com
Authentication method EAP
AAA session-id length 7, 0x30313233414243
AAA termination-action 1
Reauthentication attempts from subscriber 0, ASNGW 0
Associated user group **any**
Signalling address local 2.2.2.2, remote 10.1.1.82
Signalling UDP port local 2231, remote 2231
Idle for inbound 00:01:09, outbound 00:01:09
Absolute timeout 1500, remaining 00:23:49
Idle timeout 600 (both), remaining 00:08:50
Ingress Address filtering 0 packets, 0 bytes
Number of flows 1
Flow details ISF(0)
FSM in state SF Ready(4) on last event Up(1)
Transaction ID used 0X8004(32772)
Data ID local 0x8(8), remote 0x1(1)
Data address local 2.2.2.2, remote 10.1.1.82
Data traffic sent 2 packets, 705 bytes
Data traffic received 2 packets, 1208 bytes
Accounting last record sent Interim(3)
Idle for inbound 00:01:09, outbound 00:01:09
Service Flow information Downlink:
Identifier 15
Security Key Exchange
After EAP authentication of the subscriber, the BWG computes the respective Access Keys (AKs) for each Base-Station. The BWG also caches the PMK for the duration of the authentication, and recomputes additional AKs when the SS/MSS moves to another BS.
Release 1.0 and above supports Re-Authentication triggered from the mobile, and generates a new PMK.
IP Address Allocation Using DHCP
Cisco BWG Release 1.0 and above supports external Dynamic Host Configuration Protocol (DHCP) server-based address allocation.
Note The only mechanism to assign addresses to the SS/MSS is based on DHCP.
The SS/MSS can use DHCP to allocate IP addresses. For Release 1.0 and above there is no MIP or PMIP, because the BWG is only targeting fixed and portable. The DHCP relay is resident in the BWG, and interacts with a DHCP server, provided when the user-groups are on different VRF.
Overlapping of addresses with usergroup is allowed only with VRF.
After successful authentication and setup of the Initial Service Flow, the MS triggers DHCP to acquire an IP address. The DHCP server is configured on the BWG per user-domain group. The DHCP messages are transported transparently over the R6 data path between the BS and the BWG. The addresses can be allocated by the corresponding DHCP server pertaining to the user-domain group. Overlapping addresses across different user-groups are supported. using loop back might be the ideal way, however if the "dhcp gateway address" is not configured the IP of Virtual Template will be used as the gi-address
The initial service flow does not permit any data traffic except DHCP packets. After address allocation is successfully completed, the appropriate classifiers are installed that correspond to the IP address assigned to the SS/MSS.
In order to support multiple hosts behind a Subscriber Station, multiple DHCP requests from subscriber stations will be supported. These requests can be received on the same or alternate service flows.
Configuring IP Address Allocation
To configure IP address allocation using an external-based DHCP server, perform the following task:
Here is a sample configuration:
interface Loopback102
ip address 102.0.0.1 255.255.255.0
!
user-group domain eaptls.com2
aaa accounting method-list AAA-ACC1
aaa authentication method-list AAA-AUTHN1
dhcp gateway address 102.0.0.1
dhcp server primary 27.0.0.8
service-flow pre-defined isf profile sf3
service-flow pre-defined secondary 1 profile sf4
vrf VRF_2
The DHCP server and gateway also can be configured under User Group. If you do not configure DHCP server or gateway address under the user group, the global configuration method is used.
Multiple Host Support
Multiple hosts behind an SS can be supported for IPCS, using DHCP Relay option 82, or option 82 - subscriber ID.
Subscriber-id sub-option of Option 82 could be set to the MSID of the MS/SS and the Circuit-id sub-option can be set to the downlink service flow identifier. A remote ID can be set to the SS/MSS's username for an authenticated user, and the VPNID can be set to the user's VRF name if configured. This includes the new sub option 200 for the L2 header
For example, the DHCP server can allocate a unique IP address for each MAC, to support a multi-host scenario.
Now, the subscriber ID will have the username, and the remote ID will have the MACID of the user.
Note For Release1.0 and above, relay cascading is not supported.
Note The maximum number of hosts allowed behind an MS is 8.
Support of Multiple Hosts Behind a SS
Multiple hosts are also supported over a single SS/MSS
Step 1 CPE (SS) undergoes initial network entry and authentication, and a bearer path is created.
Step 2 A basic R6 bearer path between the BS and the BWG is created. The basic R6 shares a GRE key for uplink/downlink, which may be mapped to the SFID and the corresponding airlink connection.
Step 3 All uplink and downlink packets are sent and received by the CPE for all the hosts on the same service flows (R6 bearer) at the BWG.
DHCP Option 82
DHCP option 82 is applicable for subscribers as well as host. This option is sent in any DHCP messages for any host or subscriber.
Multiple hosts can also be supported using the DHCP option 82. The Relay Agent Information option is inserted by the DHCP relay agent when it forwards client-originated DHCP packets to a DHCP server. Servers that recognize the Relay Agent Information option can use the information to implement IP address, or other parameter assignment policies. Additionally, the L2 header in case of Ethernet CS is also being inserted in option 82
DHCP options 82 appends subscriber id + remote id + circuit id. This is then sent in all DHCP messages toward the server. In case of VRF, VPN ID is also sent. If the DHCP server is not Option 82 aware, and does not echo back the option 82, the BWG drops the messages from DHCP server.
This feature is valuable because it allows you to do the following:
•Identify each subscriber
•Perform subscriber management
•Assign IP addresses based on subscriber info
•Set access control, QoS and security policies
Here is the sequence of events that occur for the DHCP Option 82 feature:
Step 1 Hosts set the client identifier field to the MAC address in the DHCP message.
Step 2 DHCP message communication is done only over ISF for procuring the CPE's IP address, and can be done on any of the flows for procuring the host's IP address. The DHCP packets from BWG are sent out on the same flow as the incoming DHCP message from the host.
Step 3 The BWG inserts the option 82 fields for use by the DHCP server. Option 82 shall be inserted into all DHCP messages towards the DHCP server. For the list of options to insert refer to Table 2-3
Step 4 The DHCP Server could allocate IP address using any of the options in the Option 82 field of the incoming DHCP packet. Once the IP address is allocated, the BWG learns the assigned IP address by monitoring the responses and maps it to the R6 bearer. This process is repeated for each host, and the address is tracked and mapped to the same R6 bearer.
Step 5 The BWG will monitor all DHCP messages, and ensure that the option 82 fields are inserted.
Table 2-3 lists the DHCP Server Options.
DHCP Option 82 Enhancement in Release 1.1
In order to allow the BWG to build the L2 header for downlink DHCP packets, the entire L2 header is coded in the Option 82, which is reflected back from the DHCP server. This sub-option only applies between the BWG and DHCP server.
Service Flow Creation and Management
802.16 supports multiple service flows for a given SS. The service flows are identified by mapping a set of classification rules over the packet bearer. Each service flow is a unidirectional flow and can have a different quality of service treatment, both on the airlink and on the network.
In Cisco BWG Release 1.0 and above, service flow creation is supported only when initiated by the network. This service flow creation will provision the classifiers on the SS/MSS as well.
Additionally, pre-provisioned service flow templates are configured on the BWG locally. AAA support for downloading the Service Flow Profile ID is not supported on the BWG.
Service Flows
The BWG manages the service flows for each SS/MSS. Release 1.0 and above only supports network triggered service flows. The BWG allocates SFID for each service flow, and triggers service flow creation. Each service flow also has its respective datapath (for example, GRE key, and the packets corresponding to each service flow are transported accordingly).
All pre-provisioned flows are assumed to be available for the lifetime of the SS/MSS session, and are not deleted.
Multiple Service Flow Creation
When the control plane comes up, the BWG requests the creation of the Initial Service flow with the base station. DHCP IP address allocation and flow creation go in parallel in BWG Release 1.1.
The flows are created one after the other in parallel to DHCP allocation occuring on the ISF. Launching the creation of a service flow happens only after the successful creation of the preceding one. A service flow "fails" to be created if it fails after "x retries" of registration request for this SF. If a SF "fails" to be created, it is bypassed (in case it is a secondary SF), or the session is torn down (if it is the inital SF).
For Release 1.0 and above, the BWG supports creating 4 service flows; the initial service flow, and 3 secondary service flow.
If a secondary SF creation fails, then the next flow is attempted and session continues without the failed SF.
Configuring BWG Service
To enable BWG services, use the following commands beginning in global configuration mode:
Sample Configuration
Here is a sample configuration to clone the Virtual Address:
!
interface Virtual-Template1
ipaddress 2.2.2.2 255.255.0.0
encapsulation agw
ip mtu 1440
no keepalive !
The Gi address is picked from the Virtual Address by default. You can use the user-group configuration to override the Gi address.
Verifying the Configuration
To verify that BWG services are enabled, and to display MS State Change and Data Path statistics, use the show wimax agw statistics command in privileged EXEC mode:
Message type Deregistration Request(4/0x4)
Number of messages sent 1
Number of messages received 11
Number of messages resent 0
Message type Deregistration Response(5/0x5)
Number of messages sent 6
Number of messages received 1
Number of messages resent 10
Message type Deregistration Ack(6/0x6)
Number of messages sent 1
Number of messages received 5
Number of messages resent 0
Message type Registration Request(12/0xC)
Number of messages sent 6
Number of messages received 0
Number of messages resent 0
Message type Registration Response(13/0xD)
Number of messages sent 0
Number of messages received 6
Number of messages resent 0
Message type Registration Ack(14/0xE)
Number of messages sent 6
Number of messages received 0
Number of messages resent 0
Message function type Context Delivery(4/0x4)
Message type Context Delivery Request(1/0x1)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Context Delivery Report(2/0x2)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message function type Auth Relay(8/0x8)
Message type EAP Start(1/0x1)
Number of messages sent 0
Number of messages received 2
Number of messages resent 0
Message type EAP Transfer(2/0x2)
Number of messages sent 56
Number of messages received 56
Number of messages resent 0
Message type Key Change Directive(5/0x5)
Number of messages sent 8
Number of messages received 0
Number of messages resent 0
Message type Key Change Confirm(6/0x6)
Number of messages sent 0
Number of messages received 2
Number of messages resent 0
Message type Key Change ACK(7/0x7)
Number of messages sent 2
Number of messages received 8
Number of messages resent 0
Message type CMAC Key Count Update(8/0x8)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type CMAC Key Count Update Ack(9/0x9)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message function type MS State Change(9/0x9)
Message type Attachment Response(7/0x7)
Number of messages sent 6
Number of messages received 0
Number of messages resent 0
Message type Attachment Request(8/0x8)
Number of messages sent 0
Number of messages received 6
Number of messages resent 0
Message type Attachment ACK(9/0x9)
Number of messages sent 0
Number of messages received 6
Number of messages resent 0
Message type Pre Attachment Request(15/0xF)
Number of messages sent 0
Number of messages received 6
Number of messages resent 0
Message type Pre Attachment Response(16/0x10)
Number of messages sent 6
Number of messages received 0
Number of messages resent 0
Message type Pre Attachment ACK(17/0x11)
Number of messages sent 0
Number of messages received 6
Number of messages resent 0
Message function type Keepalive(20/0x14)
Message type Keepalive Request(1/0x1)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Keepalive Response(2/0x2)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Handoff Statistics
Message type Successful Handoff
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Registration Request
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Registration Response
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Registration Ack
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Deregistration Request
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Deregistration Response
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Deregistration Ack
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Undefined Message Function / Message Type
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Mapping of Service Flows to DiffServ Classes
The BWG maps each individual Service flow to a Diffserv Class. The mapping rules are configured on the router. The mapping rules are designated in Table 2-4:
Marking of Packets Corresponding to Service Flows
Each packet is identified and grouped according to the associated service flow. The transport headers corresponding to the packets are then marked with the associated Diffserv Code Point (DSCP) by the BWG based on the above table.
Configuring Service Flows on the BWG
To create service flows on the BWG, perform the following tasks:
Configuration Example
The following are examples of Service Flow configuration commands:
wimax agw service-flow profile isf
direction downlink
cs-type ip-cs
pak-classify-rule isf-classifier-downlink
precedence 1
cs-type ethernet-cs
pak-classify-rule isf-classifier-downlink
precedence 2
qos-info isf-qos-downlink
!
direction uplink
cs-type ip-cs
pak-classify-rule isf-classifier-uplink
precedence 1
cs-type ethernet-cs
pak-classify-rule isf-classifier-uplink
precedence 2
vlan 2 vrf vrf_1
vlan range 3 10 vrf vrf_2
vrf-default vrf_1
qos-info isf-qos-uplink
wimax agw service-flow profile 2sf
direction downlink
cs-type ip-cs
pak-classify-rule dn-secondary-01
qos-info downlink-qos-02
set dscp ef
set precedence immediate
!
direction uplink
cs-type ip-cs
pak-classify-rule up-secondary-01
qos-info uplink-qos-02
!
!
Configuring Service Flow Packet Classification
To configure a service-flow packet classification rule profile on the BWG, perform the following tasks:
Configuration Example
Here is a sample configuration of the Service Flow Packet Classification configuration commands:
wimax agw service-flow pak-classify-rule profile sec1-classifier-uplink
priority 0
ipv4 permit ip any any
ethernet permit any any any
vlan any priority any
!
priority 1
vlan 300 priority 4 7
!
priority 2
ethernet permit 0032.00AE.0023 ffff.ffff.ffff any arp
!
priority 3
ipv4 permit ip 2.2.2.2 /24 192.168.102.0 /24 tos 0 255 100
!
priority 4
ethernet permit any 0032.00AE.0023 ffff.ffff.ffff 8100
vlan permit 900 priority 4
!
priority 5
ipv4 permit ip 2.2.2.2 /24 192.168.102.0 /24 tos 0 255 100
ethernet permit 001C.B046.041B ffff.ffff.0000 0032.00AE.0023 ffff.0000.0000 ipv4
vlan permit 300 priority range 4 7
Note The packet classifiers are viewed collectively for a given user and direction of flow for each packet, and the highest matching priority rule is applied (255 is highest priority). If no classifiers match, the default flow chosen is the ISF in the downlink direction.
QoS Support
QoS support refers to both airlink QoS as well as mapping on the network. The BWG is responsible for sending the QoS parameters to the BS used to create the appropriate service flows.
Certain hosts can be given additional QoS parameters.
A new R6 bearer (service flow) is created that corresponds to the host's IP address. Multiple hosts can use this service flow.
Mapping of the host to the new R6 service flow is created and communicated to the BS/MS through the RR-Request.
BWG Release 1.0 and above offers the following support:
•Support for pre-provisioned QoS through CLI.
•Support for signaling traffic to be marked as separate class.
•Corresponding to every service flow based on the classifiers, a Diffserv Class would be mapped and used by the BS and the BWG.
•Support for all QoS class of service.
Configuring QoS
To configure QoS on the BWG, perform the following tasks:
Configuration Example
Here is a QoS configuration example:
wimax agw service-flow qos-info profile isf-qos-downlink
data-delivery-service real-time-variable-rate
maximum-latency 1
maximum-traffic-burst 2
maximum-traffic-rate-sustained 3
media-flow-type 012041424344
minimum-traffic-rate-reserved 4
policy-transmission-request 5
sdu-size 6
tolerated-jitter 7
traffic-priority 1
unsolicited-interval-grant 8
unsolicited-interval-polling 9
wimax agw service-flow qos-info profile isf-qos-uplink
data-delivery-service unsolicited-grant
maximum-latency 11
maximum-traffic-burst 21
maximum-traffic-rate-sustained 31
minimum-traffic-rate-reserved 41
policy-transmission-request 51
sdu-size 61
tolerated-jitter 71
traffic-priority 3
unsolicited-interval-grant 81
unsolicited-interval-polling 91
!
wimax agw service-flow qos-info profile downlink-qos-02
data-delivery-service real-time-variable-rate
media-flow-type 05abcd
Verifying the Configuration
To verify the QoS values on the BWG, use the show wimax agw subscriber command. Here is sample output for QoS statistics:
Router>sh wimax agw subscriber
MSID 1000.2228.0001
Connection time 000:00:14
Auth policy 0X0(0)
Number of TIDs 1
TID Key 10.1.1.70/2.2.2.2/1000.2228.0001
Peer TID 0X2(2)
FT MS State Change(9), MT Attachment Request(8)
Our TID 0x8001(32769)
QoS information:
Data-delivery-service real-time-variable-rate
Minimum traffic-rate-reserved 4, Maximum latency 1
Unsolicited interval-polling 9, Traffic-priority 1
Maximum traffic-rate-sustained 3, Request/Transmission-
policy 5
Maximum traffic-burst-rate 2
Reduced-resources-code 0
Classifier information:
priority 0 permit ip host 0.0.0.0 host 0.0.0.0
Service Flow information Uplink:
Identifier 4
QoS information:
Data-delivery-service unsolicited-grant
Minimum traffic-rate-reserved 41, Maximum latency 11
Tolerated-jitter 71, SDU-size 61
Unsolicited interval-grant 81, Request/Transmission-policy 51
Reduced-resources-code 0
Classifier information:
priority 0 permit ip host 0.0.0.0 host 0.0.0.0
Table 2-5 and Table 2-6 identify the QoS Classes and Service Parameters for 802.16.
User Group Management
To configure user groups on the BWG, perform the following tasks:
Sample Configuration
The following example illustrates how to configure a user group:
!
wimax agw user group-list wimax
user-group any
aaa accounting method-list agw
sla profile-name gold
dhcp server primary 12.1.1.2
!
user-group domain cisco.com
aaa accounting method-list agw
sla profile-name gold
ip static-allowed
ip route aggregate auto
!
user-group unauthenticated
aaa accounting method-list agw
aaa authentication method-list agw
sla profile-name gold
ip static-allowed
user auto-provisioning
proxy realm cisco.com password ciscoway
Idle Timer Support
An idle timer is configurable on the BWG for a User group. If there is no data traffic for the duration of the timer, the SS/MSS will be de-registered. Idle timeout can be downloaded from the AAA server during the authentication phase.
Here is a sample configuration:
wimax agw user group-list wimax
user-group any
aaa accounting method-list agw
dhcp server primary 11.1.1.93
service-flow pre-defined isf profile isf
timeout idle 30
timeout session 30
!
user-group unauthenticated
aaa accounting method-list agw
dhcp server primary 11.1.1.93
service-flow pre-defined isf profile isf
service-flow pre-defined secondary 1 profile 2sf
!
!
Idle timer support is available for inbound traffic in the ASN.
If an idle timer value is configured in AAA and under an ASN user-group, then AAA is given precedence.
Session Timer Support
A Session or Absolute timer is configurable on the BWG for a User group. When the timer expires, the subscriber is de-registered. Session timeout can be downloaded from the AAA server during the authentication phase.
Mobile Subscriber Station De-Registration
Cisco BWG Release 1.0 and above supports Network Exit as a result of Path Deregistration messaging.
There are two possible ways to deregister a Mobile Subscriber Station:
Mobile Subscriber Station Initiated De-Registration
Step 1 The SS sends DREG-REQ message to the BS, to start de-registration procedure.
Step 2 The BS sends Data Path De-Reg Request to BWG.
Step 3 BWG sends Data Path De-Reg Response to BS with the action code (set to 0x04) to authorize de-registration procedure.
Step 4 BS sends DREG-CMD to SS to de-register the SS.
Step 5 BS sends Data Path De-Reg Ack to BWG to complete the transaction.
Network-Initiated De-Registration
Step 1 The BWG sends out a Data Path De-Reg Request message to the BS indicating the MS to be deleted.
Step 2 The BS sends out a DSD-REQ over the airlink to deregister the specific Service Flows.
Step 3 BS gets DSD-RSP from SS indicating the termination of the service flow.
Step 4 BS sends Data Path De-Reg Response to BWG indicating the termination of service flow.
Step 5 BWG sends Data Path De-Reg Acknowledgement, to terminate the transaction.
AAA Accounting Start-Stop-Interim
The BWG supports both per service flow accounting information, and host based accounting information. Only time based interim accounting updates are supported. The BWG supports per service flow, and generates a unique set of accounting records for each service-flow tuple (Acct-Session-Id + Acct-Multi-Session-Id + PDFID). Each service flow is uniquely identified by a GRE key. A given MS can have more than one service flow.
Note Per-session accounting is not supported in this release.
For all the accounting records sent by the BWG, the Framed-IP-Address field is set to the mobile's IP address, irrespective of which host behind the mobile the traffic is sent for.
The BWG sends the following messages to the AAA server:
•Accounting Start: The BWG sends this message to the AAA server when a new service flow is created. In case of redundant BWG configuration, a stand-by BWG sends an Accounting Start message only when it becomes active. The trigger for the Accounting Start is the successful creation of the service flows. In case of the initial service flow, the accounting start record is sent only after the IP address is allocated to the users. For the secondary service flow, the accounting record is sent as soon the flow is successfully opened with the BS.
•Accounting Interim Update: The BWG generates an Accounting Update message if periodic accounting update message is configured. The accounting updates are based on a time trigger, and when configured. The minimum permitted value for the timer is 1 minute.
•Accounting Stop: The BWG sends and Accounting Stop message when the service flow is deleted or when the MS completes the deletion.
The attributes sent in the accounting record are listed in Table 2-7:
[b] NAS-ID MUST appear in the Access-Request. NAS-IP-Address may also appear. NAS-ID may be configured on the CLI using the radius-server attribute 32 include-in-access-req command.
[c] If this attribute is present then the home address assigned to the mobile must be as specified by this attributes. If this attribute is absent then the home address is derived from MIP procedures or other means (for example, DHCP).
[d] Both Session-Timeout and Termination-Action MUST be present. Termination-Action MUST be set to "RADIUS-Request"(1). This causes the NAS to re-authenticate when the Session-Timeout expires.
[f] The attribute must be encrypted using the procedures in section 3.5 of RFC2868
[h] If more then one class attribute is found in an Access-Accept message, the NAS shall store all of them and send them back in the accounting request packets.
[i] Must appear in the Access-Request associated with the User Authentication phase of the Double EAP Device, user authentication procedure. Otherwise, the attribute MUST not be present in the Access-Request message.
[k] Attributes must not appear in the Access Accept sent associated with the Device Authentication phase of double EAP.
[m] If the Framed MTU appears in an Access-Request during Access-Authentication then it indicates the MTU on the link between the NAS and the MS. As per RFC3579, the RADIUS shall not send any subsequent packet in this EAP conversation containing EAP-Message attributes whose values, when concantenated, exceed the length specified by the Framed-MTU value.
[n] Either the BS-ID or NAP-ID SHALL be provided. If both are provided the receiver SHALL ignore the NAP-ID attribute. In Release 1.0 and above, NAP_ID is not sent to AAA. NAP-ID is 24 (MSB) bits of 48 bit BSID (when BS will send it in future).
Host Based Accounting
In BWG Release 1.2, the BSG also supports host-based accounting due to the fact that IP address allocation and de-allocation happens on a host level.
Both flow and host-based accounting can be enabled or disabled under the user group. And both host and flow modes can be enabled simultaneously. By default, flow accounting is enabled and host accounting is disabled.
If host accounting is enabled for a host in a user group, then the BWG initiates accounting by sending an Accounting Start request to the RADIUS server every time a host (default host is known as the CPE, DHCP host behind the CPE, Static Host) gets created on BWG. An accounting stop request is sent if the session gets de-registered or the host gets deleted.
The list of attributes that needs to be sent in Accounting Request differs for flow and host accounting. For example, the remote_id tag is included only for host accounting.
Here is a list of the attributes needed for host-based accounting:
•Framed-Protocol
•Called-Station-Id
•Remote Id = MACID of MODEM/CPE
•User-name
•Farmed-Ip-Address = IP address of Host/CPE
•Vrf-Id (if any)
•Calling-Station-Id = MACID of host
•Basesation-ID
Configuring AAA Accounting
To enable the accounting feature on the BWG, perform the following tasks:
Configuration Example
Here is an example of a user group configuration:
wimax agw user group-list wimax
user-group any
aaa accounting method-list agw
aaa authentication method-list agw
!
user-group domain cisco.com
aaa accounting method-list agw
aaa authentication method-list agw
!
user-group unauthenticated
aaa accounting method-list agw
Here is an example of a AAA and RADIS configuration:
aaa new-model
!
aaa accounting update periodic 15
aaa accounting network agw start-stop group radius
aaa authorization network default group radius
aaa authentication dot1x agw group radius
!
radius-server attribute 32 include-in-access-req format %h.%d.%i
radius-server attribute 55 access-request include
radius-server attribute 25 accounting prefer-preauth
radius-server vsa send accounting wimax
radius-server vsa send authentication wimax
radius-server host 172.19.25.8 auth-port 1645 acct-port 1646 key cisco
radius-server host 1.8.91.8 auth-port 1645 acct-port 1646 key cisco
!
You can disable flow accounting by configuring no aaa accounting flow enable. As a result, no aaa accounting flow enable will appear under user-group profile.
Here is a sample configuration when flow accounting is disabled
user-group domain cisco.com
aaa accounting method-list agw
no aaa accounting flow enable
sla profile-name platinum
timeout idle 3000
timeout session 86400
You can enable flow accounting by configuring the aaa accounting flow enable command. Here is a sample configuration when flow accounting is enabled
user-group domain cisco.com
aaa accounting method-list agw
sla profile-name platinum
timeout idle 3000
timeout session 86400
Here is a sample configuration for host-based accounting:
!
user-group domain cisco2.com
aaa accounting method-list agw
sla profile-name gold
ip static-allowed
security subscriber address-filtering ingress
subscriber network-behind
aaa accounting host enable
Verifying the Configuration
Here is an example of the show wimax agw subscriber command, used to verify the accounting configuration:
Router#sh wimax agw subscriber msid 1000.0002.0001
Connection time 000:01:08
Auth policy 0X12(18), Single-EAP, CMAC
Number of TIDs 1
TID Key 10.1.1.82/2.2.2.2/1000.0002.0001
Peer TID 0X4(4)
FT MS State Change(9), MT Attachment Request(8)
Our TID 0x8004(32772)
Subscriber address 2.2.0.8, type IPv4, organization IETF
Subscriber address method Dynamic, source DHCP relay
Subscriber address assigned on flow downlink ID 15
Subscriber address prefix len allocated 32, aggregate 32
Subscriber address traffic sent 0 packets, 0 bytes
Subscriber address traffic received 0 packets, 0 bytes
Subscriber address DHCP XID 2390, server 0.0.0.0, htype 1
Subscriber address DHCP client ID 1000.0002.0001, length 6
Subscriber address DHCP Refresh time 86400 seconds
Number of sessions 1
Session details:
FSM in state Ready(7) on last event Rx Attach Ack(14)
Username eap-md5-u@eap-md5.com
Authentication method EAP
AAA session-id length 7, 0x30313233414243
AAA termination-action 1
Reauthentication attempts from subscriber 0, ASNGW 0
Associated user group **any**
Signalling address local 2.2.2.2, remote 10.1.1.82
Signalling UDP port local 2231, remote 2231
Idle for inbound 00:01:09, outbound 00:01:09
Absolute timeout 1500, remaining 00:23:49
Idle timeout 600 (both), remaining 00:08:50
Ingress Address filtering 0 packets, 0 bytes
Number of flows 1
Flow details ISF(0)
FSM in state SF Ready(4) on last event Up(1)
Transaction ID used 0X8004(32772)
Data ID local 0x8(8), remote 0x1(1)
Data address local 2.2.2.2, remote 10.1.1.82
Data traffic sent 2 packets, 705 bytes
Data traffic received 2 packets, 1208 bytes
Accounting last record sent Interim(3)
Idle for inbound 00:01:09, outbound 00:01:09
Service Flow information Downlink: Identifier 15
Here is sample RADIUS output for a AAA accounting start:
*Aug 11 02:27:21.143: RADIUS(00000006): Send Accounting-Request to
1.8.91.8:1646 id 1646/61, len 165
*Aug 11 02:27:21.143: RADIUS: authenticator C4 F4 3F A3 00 1C 01 66 - 78
DD A4 B4 68 37 F9 5B
*Aug 11 02:27:21.143: RADIUS: Acct-Session-Id [44] 10 "00000006"
*Aug 11 02:27:21.143: RADIUS: Framed-Protocol [7] 6 noval0
[0]
*Aug 11 02:27:21.143: RADIUS: Called-Station-Id [30] 9 "2.2.2.2"
*Aug 11 02:27:21.143: RADIUS: Framed-IP-Address [8] 6 2.2.0.76 *Aug 11 02:27:21.143: RADIUS: Calling-Station-Id [31] 19 "10-00-22-
25-00-01"*Aug 11 02:27:21.143: RADIUS: Acct-Input-Octets [42] 6 1208 *Aug 11 02:27:21.143: RADIUS: Acct-Output-Octets [43] 6 666 *Aug 11 02:27:21.143: RADIUS: Acct-Input-Packets [47] 6 2 *Aug 11 02:27:21.143: RADIUS: Acct-Output-Packets [48] 6 2 *Aug 11 02:27:21.143: RADIUS: Vendor, Wimax [26] 13
*Aug 11 02:27:21.143: RADIUS: GMT-Time-Zone-Offse[3] 7
*Aug 11 02:27:21.143: RADIUS: 00 00 00 00 00
[?????]
*Aug 11 02:27:21.143: RADIUS: Vendor, Wimax [26] 11
*Aug 11 02:27:21.143: RADIUS: Packet-Data-Flow-ID[26] 5
*Aug 11 02:27:21.143: RADIUS: 00 00 00
[???]
*Aug 11 02:27:21.143: RADIUS: Acct-Session-Time [46] 6 1630 *Aug 11 02:27:21.143: RADIUS: Acct-Status-Type [40] 6 start
[3]
*Aug 11 02:27:21.143: RADIUS: NAS-Port-Type [61] 6 802.16e Wimax
[27]
*Aug 11 02:27:21.143: RADIUS: NAS-Port-Id [87] 11 "WiMAX-AGW"
*Aug 11 02:27:21.143: RADIUS: Service-Type [6] 6 Framed
[2]
*Aug 11 02:27:21.143: RADIUS: NAS-IP-Address [4] 6 2.2.2.2 *Aug 11 02:27:21.143: RADIUS: Acct-Delay-Time [41] 6 0 *Aug 11 02:27:21.175: RADIUS/ENCODE(00000007):Orig. component type = AGW
*Aug 11 02:27:21.175: RADIUS/ENCODE: NAS PORT sending disabled
*Aug 11 02:27:21.175: RADIUS(00000007): Config NAS IP: 0.0.0.0
*Aug 11 02:27:21.175: RADIUS(00000007): sending
*Aug 11 02:27:21.175: RADIUS/ENCODE: Best Local IP-Address 2.2.2.2 for
Radius-Server 1.8.91.8
Here is sample RADIUS output for a AAA accounting stop:
*Feb 18 15:30:29.011: RADIUS(00000006): Send Accounting-Request to
172.19.25.8:1646 id 1646/24, len 252
*Feb 18 15:30:29.011: RADIUS: authenticator 6D FC 9B 49 59 28 56 41 - 3F 2E A5
3C 7B 7A 3A B1
*Feb 18 15:30:29.011: RADIUS: Acct-Session-Id [44] 10 "00000008"
*Feb 18 15:30:29.011: RADIUS: Framed-Protocol [7] 6 noval0
[0]
*Feb 18 15:30:29.011: RADIUS: Called-Station-Id [30] 9 "2.2.2.2"
*Feb 18 15:30:29.011: RADIUS: Framed-IP-Address [8] 6 2.2.0.2 *Feb 18 15:30:29.011: RADIUS: Calling-Station-Id [31] 19 "06-76-22-24-22-22"
*Feb 18 15:30:29.011: RADIUS: Vendor, Wimax [26] 10
*Feb 18 15:30:29.011: RADIUS: AAA-Session-ID [4] 4
*Feb 18 15:30:29.011: RADIUS: 00 00
[??]
*Feb 18 15:30:29.011: RADIUS: User-Name [1] 23 "eap-md5-u@eap-
md5.com"
*Feb 18 15:30:29.011: RADIUS: Acct-Input-Octets [42] 6 0 *Feb 18 15:30:29.011: RADIUS: Acct-Output-Octets [43] 6 0 *Feb 18 15:30:29.011: RADIUS: Acct-Input-Packets [47] 6 0 *Feb 18 15:30:29.011: RADIUS: Acct-Output-Packets [48] 6 0 *Feb 18 15:30:29.011: RADIUS: Multilink-Session-ID[50] 10 "30313233"
*Feb 18 15:30:29.011: RADIUS: Class [25] 21
*Feb 18 15:30:29.011: RADIUS: 63 6C 61 73 73 2D 77 69 6D 61 78 2D 63 68 61 6E [class-wimax-chan]
*Feb 18 15:30:29.011: RADIUS: 67 65 64
[ged]
*Feb 18 15:30:29.011: RADIUS: Vendor, Wimax [26] 13
*Feb 18 15:30:29.011: RADIUS: GMT-Time-Zone-Offse[3] 7
*Feb 18 15:30:29.011: RADIUS: 00 00 00 00 00
[?????]
*Feb 18 15:30:29.011: RADIUS: Vendor, Wimax [26] 17
*Feb 18 15:30:29.011: RADIUS: BaseStation-ID [46] 11
*Feb 18 15:30:29.011: RADIUS: 00 0A 01 01 46 00 00 00 00 [????F????]
*Feb 18 15:30:29.011: RADIUS: Vendor, Wimax [26] 11
*Feb 18 15:30:29.011: RADIUS: Packet-Data-Flow-ID[26] 5
*Feb 18 15:30:29.011: RADIUS: 00 05 01
[???]
*Feb 18 15:30:29.011: RADIUS: Acct-Session-Time [46] 6 25
*Feb 18 15:30:29.011: RADIUS: Acct-Terminate-Cause[49] 6 none
[0]
*Feb 18 15:30:29.011: RADIUS: Acct-Status-Type [40] 6 Stop
[2]
*Feb 18 15:30:29.011: RADIUS: NAS-Port-Type [61] 6 802.16e Wimax [27]
*Feb 18 15:30:29.011: RADIUS: NAS-Port-Id [87] 11 "WiMAX-AGW"
*Feb 18 15:30:29.011: RADIUS: Service-Type [6] 6 Framed
[2]
*Feb 18 15:30:29.011: RADIUS: NAS-IP-Address [4] 6 172.19.24.88 *Feb 18 15:30:29.011: RADIUS: Acct-Delay-Time [41] 6 0 *Feb 18 15:30:29.019: RADIUS: Received from id 1646/23 172.19.25.8:1646, Accounting-response, len 20
*Feb 18 15:30:29.019: RADIUS: authenticator 4D 1A 1B 4D C5 0E 39 FD - 36 6B 90 FF 96 21 66 64
*Feb 18 15:30:29.019: RADIUS: Received from id 1646/24 172.19.25.8:1646, Accounting-response, len 20
*Feb 18 15:30:29.019: RADIUS: authenticator EB 25 42 F1 48 2C BF 13 - 43 B0 0A 3A 7A 04 F4 1F
WiMAX Specific VSAs
The following VSAs are specific to WiMax:
•Wimax Capability—Indicates the WiMAX release, accounting capabilities indication, Hotlining capabilities, and Idle Mode Notification capabilities of the BWG to the AAA in an Access Request.
•GMT Time Zone Offset—The current offset in seconds of the local time at the NAS with respect to GMT time.
•Packet Data Flow-Id (PDFID)—The value of this attribute matches all records from the same packet data flow. PDFID is assigned by the CSN, and remains constant through all handover scenarios. In Release 1.0 and above, the BWG generates the PDFID for a flow in the session.
•Base Station ID—Uniquely identifies a NAP and a base station within that NAP. The BWG forwards the R6 BS ID in this attribute.
•AAA Session ID—A unique per realm identifier assigned to the WiMAX session by the home network during network entry. The value is included in all subsequent AAA packets for that session.
Handoffs
Multiple forms of handoff are supported for WiMax, both inter-base station and inter-BWG. BWG Release 1.0 and above only supports inter-base station handoff.
Inter-BS handoff includes both uncontrolled and controlled handoffs. In the controlled case, the target BS has obtained the session information from the serving BS through R8 interface before the handoff actually occurs. Uncontrolled handoffs occur where information exchange between base stations is not possible prior to the target BS triggering a handoff at the BWG. Uncontrolled handover is treated identically to Initial Network Entry, with the addition that the BWG ensures that paths registered with the serving base station are deregistered. The only addition being that for Release 1.0 and above, one attempt is made to send the deregistration message to the serving BS, and the handoff takes place regardless of the completion of the deregistration handshake between the ASN and SBS
Handoffs do not trigger interim accounting updates.
Uncontrolled Handoff
An uncontrolled handover is signaled from the BS to the BWG using a Path Registration Request message. This message contains information for each service flow that is already established with the source BS. It also contains the DP-IDs used for downlink flows.
Note There is no need to re-authenticate the device or the subscriber, as the session is maintained at the same BWG.
Note In uncontrolled handoff, the target BS will trigger a MS network entry in which the MS will get authenticated.
The BWG initiates the deregistration of the path to the old BS. This deregistration will be scheduled by the BWG. It does not necessarily occur directly after successful completion of handoff to the new BS.
There is no requirement to buffer bearer path data during handoff. Downlink data received at the BWG during the handover procedure is discarded.
Any traffic that is "in-flight" through the old path is lost because the device has already moved to the service area of the target BS before to the handoff trigger is received at the BWG.
It is possible that the device may move to a new BS while the handoff procedures between the target BS and the BWG are completed. Because the handover is uncontrolled, the handoff to the current target BS is completed (including R6 message retransmissions, if necessary) before the new handoff event is processed.
The handover exchange comprises three messages (applicable only for controlled handoff):
•Path Registration Request—sent from the Target BS to the BWG—which contains the following:
–Registration Type
–SF INFO(s) with SFID, Reservation Action (set to Create), Direction, QoS parameters, Data Path Info and GRE Key (for downlink flows)
–BS INFO with BSID
•Path Registration Response—sent from the BWG to the Target BSq—which contains the following:
–Registration Type
–SF INFO(s) with SFID, Reservation Action (set to Success), Direction, Data Path Info & GRE Key (for uplink flows)
–BS INFO with BSID
•Path Registration Acknowledgement—sent from the Target BS to the BWG—which contains the following:
–Registration Type
If the BWG cannot accept the handover, it sends the response with "reject cause code TLV".
If the BWG accepts the handover for only a subset of the desired Service Flows, the handover is rejected.
Handoff will not be rejected if secondary flow is missing, but if primary flow is missing it will be rejected.
The Deregistration Request and ACK sent to SBS will have the registration type as "Handover" while Deregistration response from SBS will have "Network exit". This is an expected behavior. On receiving this, the BWG does not send the ACK with "reject cause code TLV".
Controlled Handoff
A controlled handoff occurs when the current and target BSs are able to communicate information and exchange details about service flows, classifiers, and other details, prior to the target BS triggering the handoff at the BWG. This means that the target BS has all relevant information about the mobile device prior to sending the BWG handoff trigger. This trigger occurs when the mobile device has already been connected to the target BS using 802.16e procedures. You can tell a controlled handoff occurred at the BWG by the receipt of a Path Registration Request message from the BS without a previous authentication exchange (which would be observed for a Network Entry event).
The following flow sequence illustrates the events that occur during a controlled handoff:
Step 1 The Target Base Station sends a Path Registration Request to the BWG containing the service flow information received from the Serving Base Station.
Step 2 The BWG responds with a Path Registration Response accepting registration of the data path with the Target base Station.
Step 3 The Target Base Station responds with a Path Registration Acknowledgement.
Step 4 The BWG sends a Path Deregistration Request to the Serving Base Station.
Step 5 The Serving Base Station responds with a Path Deregistration Response.
Step 6 The BWG acknowledges the response with a Path Deregistration Acknowledgement.
Step 7 The Target Base Station sends a Context Report to the BWG.
Step 8 The BWG acknowledges with a Context Acknowledgement.
Step 9 The target BS sends a CMAC Key Count Update message, and the BWG responds with a CMAC Key Count Ack message.
Verifying the Configuration
To view the handoff statistics for the BWG, use the show wimax agw statistics section handoff command.
Here is a sample configuration:
Router#show wimax agw statistics section handoff
Message type Successful Handoff
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Registration Request
Number of messages sent 0
Number of messages received 2
Number of messages resent 0
Message type Handoff Registration Response
Number of messages sent 2
Number of messages received 0
Number of messages resent 0
Message type Handoff Registration Ack
Number of messages sent 0
Number of messages received 2
Number of messages resent 0
Message type Handoff Deregistration Request
Number of messages sent 2
Number of messages received 0
Number of messages resent 0
Message type Handoff Deregistration Response
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Handoff Deregistration Ack
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Security Context Exchange
In order for a BS to secure the airlink, it requires keying material from the BWG. A handoff cannot be successful from the perspective of the BS and the device until the data path registration has completed, and the BS receives the keying material. The BS is responsible to initiate both procedures. The BWG treats a context exchange with the BS as an entirely separate event from handover.
A context exchange can occur at any time. The AK transfer protocol is used to transfer the keying material to the BS. This material comprises the AK, AKID, AK Lifetime, AK sequence number and EIK.
If the PMK has expired, then a new PMK must be created.
The security context exchange comprises two messages.
•Context Request—sent from the target BS to the BWG—which contains the following:
–Context Purpose Identifier
–BS Info
–Target BS ID
•Context Report—sent from the BWG to the target BS—which contains the following:
–MS Info
–AK Context
–AKID
–AK lifetime
–AK SN
–CMAC Key count
–Target BS Info
–Target BS ID
Keepalive Support for R6 Interface
The keepalive mechanism is based on the periodic transmission of "keepalive" messages between the BS and the BWG.
Transmission of the keepalive messaging is configurable at the BWG.
Note When a Base Station BWG receives an R6 Keepalive Request message, it must send an R6 Keepalive reply.
For each R6 instance, BWG maintains the following configurable parameters:
•Tk: Keepalive timer
•N: Number of consecutive keepalive failures. Initialized to 0
•M: Permitted maximum number of consecutive keepalive failures.
The supported keepalive functionality in this release is as follows:
Step 1 The base station or BWG sends a Keepalive_Request and starts timer Tk.
Step 2 On receipt of a Keepalive_Reply, the value of N is reset to 0.
Step 3 When Tk expires, the node sends the next Keepalive_Request. N is incremented if a Keepalive_Reply for the last Keepalive_Request message was not received prior to expiry of timer Tk.
Step 4 When N equals M, N is reset to 0 and any R6 sessions established with the remote node are terminated locally (with no attempt to send any further cleanup message to remote end) and any data related to subscribers, belonging to the specific R6 interface is cleaned up.
The following two TLVs will be used in this release:
•BS ID TLV to identify the BS which is either a recipient or initiator of the request.
•BWG ID to identify the BWG which is either the recipient or initiator of the request.
Keepalive Mechanism Failover Handling
While BS and BWG are in operation, suppose one or both of these units experienced a failover condition (rebooted), without any intervention from the operator, BS and BWG may get out of synchronization with respect to the registered subscribers and their states.
Solution
The Mechanism uses Keepalive messaging/Reset-Time TLV to recover from such condition immediately after such a failover.
BWG Reboots:
If a Keepalive_Request is the first message the rebooted BWG receives from the BS assuming keepalive are enabled on the latter, the BWG responds by a Keepalive_Reply with Reset-Time TLV. It indicates the reset time of the BWG and signals the BS to begin the process of deregistering all subscribers that are associated with this BWG.
If a Pre-attachment_Request is the first message the rebooted BWG receives from the BS as a result of a network entry, the former sends a Keepalive_Request with Reset-Time TLV to the BS which should take the corrective approach of clearing all subscribers.
BS Reboots:
If the BWG receives a Keepalive_Request with Reset-Time TLV from a BS which reset time is different than the BWG boot time, the former clears all the subscribers associated with this BS.
Configuring Keepalive
To configure the keepalive value on the BWG, perform the following task:
Configuration Example
Here is a sample configuration of the Keepalive configuration commands:
wimax agw base-station group default
reference-point r6 keepalive timeout 30
reference-point r6 response retransmit 10
reference-point r6 response timeout 10
Here is a configuration example of the reference-point r6 path purge-timeout command:
Router(config)#wimax agw base-station group default
Router(config-wimax-agw-bs)#reference-point r6 ?
keepalive Enable AGW-BS keepalive feature
path WiMAX AGW BS R6 reference point base station path
response WiMAX AGW BS R6 reference point response configuration commands
Router(config-wimax-agw-bs)#reference-point r6 path ?
purge-timeout WiMAX AGW BS R6 reference point path purge timeout
Router(config-wimax-agw-bs)#reference-point r6 path purge-timeout ?
<1-4320> WiMAX AGW BS R6 reference point path purge timeout in minutes
Router(config-wimax-agw-bs)#reference-point r6 path purge-timeout 30
Verifying the Configuration
To verify various BWG system parameters, perform the following tasks:
Configuration Examples
Here is a sample configuration that identifies the BWG keepalive statistics:
Router#show wimax agw statistics | section Keepalive
Message function type Keepalive(20/0x14)
Message type Keepalive Request(1/0x1)
Number of messages sent 21
Number of messages received 0
Number of messages resent 0
Message type Keepalive Response(2/0x2)
Number of messages sent 0
Number of messages received 21
Number of messages resent 0
Here is a sample configuration that identifies generic BWG statistics:
Router#show wimax agw
Access network gateway version 1.0, service is enabled
Signaling UDP port 2231
Maximum Number of base station 500 allowed
Maximum Number of subscriber 20000 allowed
Current number of signalling paths 1
Current number of data paths 1
Current number of subscribers 3
Current number of sessions 3
Current number of flows 6
Current number of hosts 0
Traffic Sent 6 packets, 1998 bytes
Traffic Rcvd 7 packets, 4228 bytes
Here is a sample configuration that identifies BWG base station statistics:
Router#show wimax agw path 10.1.1.70
Path type Sig-UDP
State current Ready, old Idle
Number of sessions connected 3
Number of old sessions connected 0
Address local 2.2.2.2(AF_INET), remote 10.1.1.70(AF_INET)
UDP port local 2231(0x8B7), remote 2231(0x8B7)
Identification, Our 0x02020202
Keepalive timer expires in 00:00:25, timeout 30 secs
Keepalive consecutive failures max allowed 5, current 0
Keepalive Request received valid 0, invalid 0
Keepalive Response received valid 11, invalid 0
Keepalive Request sent success 11, fail 0
Keepalive Response sent success 0, fail 0
Traffic sent 29 packets, 3175 bytes
Traffic received 28 packets, 2658 bytes
Path type Data-GRE
Number of flows connected 6
Address local 2.2.2.2(AF_INET), remote 10.1.1.70(AF_INET)
Traffic sent 6 packets, 2166 bytes
Traffic received 7 packets, 4522 bytes
Here is a sample configuration that identifies BWG subscriber statistics:
Router#show wimax agw subscriber brief
MSID Address Age Flows Hosts Pkts-Tx Pkts-Rx
1000.2223.0001 2.2.0.75 000.22.08 2 0 2 3
1111.1113.1111 2.2.0.74 000.22.05 2 0 2 2
1000.2225.0001 2.2.0.76 000.21.56 2 0 2 2
Session Redundancy
Note BWG Session redundancy is not supported on the Cisco 7301 Router platform.
The BWG Session Redundancy architecture provides user session failover capability in a 1:1 redundancy model, with a standby present for every active BWG. The active BWG sends state information to the standby BWG for state synchronization on a as needed basis. When an active BWG failure occurs, the standby BWG has state information needed to provide service to all existing sessions. It then takes over as the active BWG and begins servicing user sessions, thus providing session redundancy. When the previously active BWG comes back online, it takes over as standby for the now active BWG, and obtains state information for all existing sessions from it.
The BWG is hosted on the SAMI blade, and only card to card redundancy will be supported. In other words, failure of a single processor unit on SAMI will result in the entire card being switched over.
Note In Cisco BWG Release 1.1, session redundancy is still supported. However, the configuration on both standby and active should be identical regarding classifier information and CS-Type information for each service flow direction. The new fields introduced for the host data structure are synced from the active to the standby.
BWG Session Redundancy and High Availability Infrastructure
The BWG Session Redundancy is based on the Cisco IOS Hot Standby Routing Protocol (HSRP), the Cisco IOS Check-point Facility (CF) and Redundancy Framework (RF), and Stream Control Transmission Protocol (SCTP) to provide inter-device redundancy and high availability. The Figure 2-1 shows the system view of the BWG SR with relation to the IOS HA infrastructure.
Figure 2-1 Session Redundancy on the BWG
Subscriber Management
Subscriber information includes session and flows associated with a subscriber context, and is created, updated, or eventually deleted.
Subscriber information includes the following details:
•Authentication info (method, keying info, etc.)
•Addressing info (MS MAC, assigned DHCP address, etc.)
•VRF name
•Username
•Session info (signaling address, and associated timers, etc.
•Flow info per session (and associated QoS info per flow)
DHCP and AAA
The BWG supports DHCP relay mode and keeps track of client IP addresses allocated by DHCP servers (and the associated server IP addresses) so that it can relay future DHCP messages from clients to the servers. The client IP address and DHCP server IP address are saved in the subscriber context and are synced to the standby. Once the standby becomes active, it continues to relay DHCP messages from a client to the right server (there can be multiple servers configured: primary/secondary).
IOS AAA is not HA-aware at the moment, so the sync of AAA-related information is part of the session replication.
Dynamic Synchronization
In order for the standby to take over processing from the active in case of a failure, information regarding all sessions and flows on the active are dynamically synchronized to the standby at well defined synchronization points. Separate TLVs are used to synchronize session, flow, and path related information. Dynamic syncing happens for new session/flow events after the standby is at hot-standby state, and after bulk-sync is complete.
The following list identifies current synchronization points:
•During initial network entry, session and flow information is synched to standby only after the Initial Service Flow (ISF) is created.
•After the ISF is up, each new flow created on the active is separately synched to the standby.
•Any updates to the service flow will cause the flow to be synched to the standby.
•Every time an address allocation happens, the flow will be synched to the standby.
•Any changes to the path on the active are synchronized to the standby
•During handoff, flow information is synchronized to the standby only after the handoff is complete. Cloned flows are not synched. New flows created on the active as a result of handoff are synchronized to standby by a FLOW UPDATE message that carries modified parameters as result of handoff.
•Flow synchronization after the transmission of an interim accounting request from the active. This causes FLOW UPDATE messages to be sent from active to standby, and the necessary message carries accounting counters that are sent to AAA as a part of interim accounting update.
Configuring Session Redundancy
The following configuration tasks are required before you can configure session redundancy:
•Configure HSRP on the interface.
•Configure redundancy inter-device
•Configure SCTP for RF Check pointing
•Configure Network Time Protocol (NTP) server
•Configure AAA on the active BWG and standby BWG
To configure session redundancy on the BWG, perform the following tasks:
Configuration Example
This configuration is for AAA only.
On the active BWG
---------------------------------------------------------------
!
interface Loopback192
ip address 192.168.0.70 255.255.255.255
!
!
aaa group server radius car-sg
server 1.8.70.99 auth-port 1812 acct-port 1813
!
aaa authentication dot1x car_auth_list group car-sg
aaa accounting network car_acct_list start-stop group car-sg
!
!
ip radius source-interface Loopback192
radius-server host 1.8.70.99 auth-port 1812 acct-port 1813
radius-server key r6AAA
radius-server vsa send accounting wimax
radius-server vsa send authentication wimax
!
On the Standby BWG
---------------------------------------------------------------
!
interface Loopback192
ip address 192.168.0.70 255.255.255.255
!
!
aaa new-model
!
!
aaa group server radius car-sg
server 1.8.70.99 auth-port 1812 acct-port 1813
!
aaa authentication dot1x car_auth_list group car-sg
aaa accounting network car_acct_list start-stop group car-sg
!
!
ip radius source-interface Loopback192
radius-server host 1.8.70.99 auth-port 1812 acct-port 1813
radius-server key r6AAA
radius-server vsa send accounting wimax
radius-server vsa send authentication wimax
Sample Configuration of BWG: Active
interface GigabitEthernet0/0.70
description to AAA/DHCP
encapsulation dot1Q 70
ip address 1.8.70.147 255.255.255.0
standby 70 ip 1.8.70.70
standby 70 follow P7_REDUNDANCY
Note Please reload the BWG if it suffers a time-zone change.
This configuration example includes information about DHCP:
interface Loopback102
ip address 102.0.0.1 255.255.255.0
!
user-group domain eaptls.com2
aaa accounting method-list AAA-ACC1
aaa authentication method-list AAA-AUTHN1
dhcp gateway address 102.0.0.1
dhcp server primary 27.0.0.8
service-flow pre-defined isf profile sf3
service-flow pre-defined secondary 1 profile sf4
vrf VRF_2
Authentication
A subscriber is authenticated using EAP on the active before the sessions/flows are recreated on the standby. The associated MSK, AK context and other credentials need to be transferred to the standby, along with the session stateful data. If geographic redundancy is deployed, this data must be protected. If the standby becomes active after a switchover, and if the same subscriber is re-authenticated on the new active, it follows the same authentication procedure as on the previous active.
Accounting
The accounting start, stop and interim update are only sent from the active. The standby never sends accounting records until it becomes active.
As part of the session/flow recreation on the standby, the IOS AAA database is populated with accounting records for each session/flow. For example, the "class" attribute and accounting session ID are synched from the active to the standby, and are saved to the related accounting record. This ensures that once the standby becomes active, it can send accounting records with the right info.
Synchronization of accounting counters is a function of the AAA interim accounting update feature. If the AAA interim accounting update feature is enabled, then the active BWG sends accounting records to AAA server. And the same event is used as a trigger to initiate a FLOW UPDATE event (which carries accounting counters the same as were sent to AAA server). Conversely, if this feature is disabled on the active, since there is not going to be an accounting update to AAA, there is an absence of triggers to send the accounting update synchronization message to standby. By itself, the BWG SR feature does not implement triggers to synchronize accounting counters.
The accounting session ID is a key attribute used in accounting events (start, stop, interim) and is used to collaborate records on the AAA server. It is a 4-byte unassigned integer, and is assigned uniquely within a ASNWG and increased sequentially until it rolls over. To ensure the new active can continue to generate unique accounting session IDs upon switchover, the new accounting session ID starts from the latest accounting session ID on the prior active.
Subscriber IP Address
Currently, the subscriber IP address is assigned by the DHCP server and a host route is inserted for it. When the standby recreates the subscriber session, the same host route is also inserted on the standby. The standby will not relay any DHCP messages between the DHCP client and the server until it becomes active.
QoS
For BWG Release 1.0 and above, after a flow is created and the QoS parameters for the flow are sent to the BS, the BWG active synchronizes all the QoS parameters to the standby. Out of all of the parameters, the DSCP code for a flow synched to the standby is used to mark the packets once it becomes active.
Statistics and Counters
Statistics and counters are not synched to the standby. Instead, the standby rebuilds them as it processes stateful data from the active to create, modify, and delete sessions/flows. For example, the number of sessions/flows on the standby is updated as the standby processes session/flow creation and deletion. The number of received R6 messages on the standby is accumulated from the moment it becomes active and starts to receive R6 messages.
BWG Load Balancing
When a load balancer is running, it uses the loading information of all the BWGs to select one of them, and forwards an incoming NetEntry message from the BS (with regard to a SS/MS) to the selected BWG.
When Dynamic Feedback Protocol (DFP) is configured, an active BWG periodically sends its loading information to the load balancer. A standby BWG does not send feedback and does not have accurate loading information because it does not process R6 messages and handle user traffic. Once a standby becomes active, it gradually builds accurate load as it processes R6 messages and handles user traffic. So there is an adjusting period before it can send back feedback about its current accurate load.
Data Path and GRE
The data path for a flow is recreated on the standby. The GRE keys for both the upstream and downstream of a flow are synched to the standby. Upon switchover, the new active ensures that any new GRE key allocated locally must not collide with any in-use GRE keys allocated by the previous active.
Version Control
Upgrading from the immediate lower software version to a higher version is supported. Downgrading of software version is not supported. For example, if a redundant pair of BWGs runs on version A, and the next immediate software version is version B, then upgrading from version A to B is supported (but not from B to A). This requires that the higher version understands the stateful data synched from a lower version.
Limitations
The following limitations exist in the Session Redundancy feature on the BWG:
•Synchronization of Accounting Counters
This is configurable, and depends upon the AAA interim accounting update feature to be enabled. If the AAA interim accounting feature is disabled, then the default behavior of BWG SR is to not synchronize accounting data/payload counters. This may lead to under charging. For example, if a switchover occurs between two consecutive interim updates, the counts accumulated on the active after the previous interim update are lost, since the new interim update sends only the counts that are accumulated on the new active. Additionally, a STOP could be lost right before the switchover.
Note Signaling counters are not synched.
Note For a standalone system, current AAA/Radius counters work accurately. However, when Session Redundancy is enabled, the flow (or session in terms of radius) age can be incorrect if relied upon the current counters. An additional attribute will now be sent from the ASN-GW called "session_elapsed_time" which will reflect the number of seconds since the particular flow started.
•Missing of a Session on the Standby
Although SCTP provides reliable transport, the stateful data used to replicate a session can be lost due to congestion or max retrials. In this case, the session is not recreated on the standby. In case there is a switchover, this session is lost.
•Stale Session on the Standby
For the same reason as above, if a stateful data for a session deletion is lost, then the session is not deleted from the standby while it's gone on the active.
–If there is no switchover before the next session creation of the same subscriber, it's expected that the next synching of creation of a new session for the same subscriber will clean up the stale session.
–If there is a switchover, then the stale session hangs until cleanup by manual intervention or by features like idle/session timeout.
•Mid-call Abort
If a call setup is in progression but before reaching the first synching point and if there is a switchover, the call setup is aborted, and the subscriber has to retry the call.
Switchover
When switchover occurs, a trap is generated and sent to the NMS system to indicate that the active unit has failed, and the standby has taken over as active. The following behavior is expected:
•Any new GRE key allocated locally must not collide with any in-use GRE keys allocated by the previous active.
•Any new accounting session ID allocated locally must not collide with any in-use accounting session ID allocated by the previous active.
•DHCP relay for a new session is forwarded to a configured DHCP server.
•Some of the statistics will have a fresh start.
•DFP load is rebuilt and sent to the load balancer.
The following list identifies events which will cause a switchover:
•Router reload/crash because of reasons like software crash, CPU hogging, etc.
•HSRP tracks interfaces based on the configuration. On detecting interface flap <on-off transition> HSRP will enforce reload current active which will cause switchover of activity
•Manual intervention to switch activity from current active router to redundant hot standby. This can be accomplished using the following commands:
–redundancy switch-activity force
This command is used to switch activity from current active to current hot standby. Issuing the command causes the current router to reload, the current hot standby to become active, and when the current active router comes back up it assumes the role of hot standby.
–reload
This is normal router reload command that causes switchover. The current active router will reload, and the current hot standby will become active.
BWG Load Balancing
The purpose of load balancing is scalability of the BWG without distributing intelligence across base stations. This scalability is provided by load balancing across a set of BWGs, while representing the cluster as a single BWG from the perspective of the BS. Thus, the base station will have a single point of contact. And any new BWG that is added to the system will not impact the base station provisioning.
Note Server Load Balancing and Session Redundancy are only available on Cisco 7600 SAMI platform. They are not available on the Cisco 7301 router platform.
BWG Load Balancing is based on the IOS Server Load Balancing (SLB) feature. BSs are configured with the virtual IP address of the SLB as the BWG ID. The BWG selection flow for load balancing is illustrated below. Both dispatch mode and directed mode are supported. DFP is supported for the SLB to discover the load on the real BWGs.
Note In Release 1.0 and above, the SLB sticky feature and BWG handover call flows are not supported.
The session created on SLB, when the initial context request is processed, is maintained for a configurable time period to handle the re-transmission. During this period, if a re-transmission of context request is detected for a given MS/SS, SLB directs the re-transmission request to the same real BWG, that was selected for the first context request.
DFP is supported for the SLB to discover the load on the real BWGs. Each real BWG has a limit on the number of maximum sessions it supports. Real BWG calculated load on itself is based on the existing number of sessions versus the maximum sessions that it can support, memory usage, and bandwidth usage and reports the load to SLB. SLB directs the initial context request to one of the real BWGs based on either the round robin or least connections method. An BWG will not accept any more sessions if its load as calculated by DFP is 100%. Thus, this mechanism also supports CAC.
Figure 2-2 Server Load Balancing on the BWG
BWG Selection
•During Initial Network Entry phase, the BS sends the NetEntry MS Pre-Attachment Request corresponding to the SS/MSS to the BWG configured as the default.
•This BWG sends the NetEntry MS Pre-Attachment Response to the BS. The response may contain the IP address of an alternate Authenticator ID that can handle subsequent transactions corresponding to the SS/MSS. The BS, on receiving the NetEntry MS Pre-Attachment Response, sends the NetEntry MS State Change Ack to complete the transaction.
•All subsequent transactions corresponding to the SS/MSS occur between the BS and the BWG specified in the NetEntry MS Pre-Attachment Response Message.
Modes of Operation
There are two operation modes on the BWG:
•Dispatched Mode— In this mode packets are sent to the real server without any change to the original packet. A loopback is configured in the real server with an IP equal to virtual IP, and it replies back with virtual IP address as the source address.
•Directed Mode—The packet's destination IP address is rewritten to choose the BWG's IP address, and no loopback with the virtual ip address is configured on the BWG.
In both modes the selected BWG sends the pre-attachment response.
Configuring Load Balancing
This section lists configuration details regarding server load balancing. These configuration details are mainly for Directed Mode unless otherwise specified.
Load Balancing Configuration Task List
This section lists the tasks used to configure load balancing. Required and optional tasks
are indicated.
1. On the Cisco IOS SLB, complete the following tasks:
a. Configuring a Server Farm and Real Server, (Required)
b. Configuring a Virtual Server, (Required)
c. Configuring DFP Support, (Optional, but recommended)
2. On the real BWG, complete the following tasks:
a. Configuring a Loopback Interface for SLB, (Required if using dispatched mode)
b. Configuring DFP Support on the BWG, (Optional, but recommended)
Configuring Cisco IOS SLB for Load Balancing
This section describes how to configure a Server Farm and a Real Server. To configure a Cisco IOS SLB server farm, use the following commands, beginning in global configuration mode:
Sample Configuration
ip slb serverfarm ASNGW-SR-SF
nat server
probe PINGPROBE
!
real 11.11.11.50
weight 0
inservice
!
real 11.11.11.70
weight 0
inservice
Configuring Real BWG
Configuring the BWG for Load Balancing
To configure load balancing on the BWG, complete the tasks in the following sections:
•Configuring a Loopback Interface for SLB,
•Configuring the BWG as a DFP Agent, (Optional, but recommended)
Configuring a Loopback Interface for SLB
To enable load balancing, a loopback interface must be configured with the same IP address as thevirtual server on the Cisco IOS SLB on each BWG in a farm.
To create a loopback interface, use the following commands, beginning in global configuration mode:
Configuring the BWG as a DFP Agent
To define the port number to be used by the DFP manager (the Cisco IOS SLB in this instance) to connect to the DFP agent; enter the following commands in order, beginning in global configuration mode:
Sample Configuration
ip dfp agent agw
port 5555
inservice
To configure load balancing on the BWG, perform the following task:
BWG Configuration Example
Please note that following sample configuration applies only to SAMI platform.
SLB Related Configuration of Supervisor Card
7606-R6-sup720#show running-configuration | section slb
ip dfp agent slb
port 5555
ip slb probe PINGPROBE ping
interval 3
faildetect 3
ip slb serverfarm ASNGW-SR-SF
nat server
probe PINGPROBE
!
real 11.11.11.50
weight 0
inservice
!
real 11.11.11.70
weight 0
inservice
ip slb vserver V-ASNGW-SR
virtual 50.70.80.100 udp 2231 service asn r6
serverfarm ASNGW-SR-SF
idle asn r6 request 90
inservice
ip slb dfp
agent 11.11.11.50 5555 10 0 5
agent 11.11.11.70 7777 10 0 5
7606-R6-sup720#
Sample configuration of real BWG for above configuration of Supervisor card.
asngw-real-s4p5# show running-configuration | section dfp
ip dfp agent agw
port 5555
inservice
asngw-real-s4p5#
asngw-real-s4p7#sh runn | section dfp
ip dfp agent agw
port 7777
inservice
asngw-real-s4p7#
Verifying the Configuration
To verify that load balancing is enabled on the BWG, perform the following tasks:
Configuration Example
Here is a sample configuration for SLB show commands on the BWG:
7606-R6-sup720#show ip slb sessions asn r6
vserver MSID Base Station real state
----------------------------------------------------------------------------
7606-R6-sup720#show ip slb sessions asn r6
vserver MSID Base Station real state
----------------------------------------------------------------------------
V-ASNGW-SR 0000AAAAC38ECCCC 50.35.50.1 11.11.11.50 ASNR6_ESTAB
V-ASNGW-SR 0000AAAAC392CCCC 50.35.50.1 11.11.11.50 ASNR6_ESTAB
V-ASNGW-SR 0000AAAAC396CCCC 50.35.50.1 11.11.11.50 ASNR6_ESTAB
V-ASNGW-SR 0000AAAAC39ACCCC 50.35.50.1 11.11.11.50 ASNR6_ESTAB
V-ASNGW-SR 0000AAAAC39ECCCC 50.35.50.1 11.11.11.50 ASNR6_ESTAB
< S N I P P E D >
7606-R6-sup720#show ip slb vserver detail
V-ASNGW-SR, state = OPERATIONAL, v_index = 7, interface(s) = <any>
virtual = 50.70.80.100/32:2231, UDP, service = ASNR6, advertise = TRUE
server farm = ASNGW-SR-SF, delay = 10, idle = 3600
asnr6: request idle = 90, Parse error pkt drops= 56,
Number of reject responses = 0
sticky: <none>
sticky: group id = 0
synguard counter = 0, synguard period = 0
conns = 101, total conns = 509069, syns = 0, syn drops = 0
standby group = None
7606-R6-sup720#show ip slb reals
real farm name weight state conns
-------------------------------------------------------------------
11.11.11.50 ASNGW-SR-SF 92 OPERATIONAL 83
11.11.11.70 ASNGW-SR-SF 92 OPERATIONAL 18
7606-R6-sup720#show ip slb serv
7606-R6-sup720#show ip slb serverfarms
server farm predictor nat reals bind id interface(s)
--------------------------------------------------------------------------
ASNGW-SR-SF ROUNDROBIN S 2 0 <any>
7606-R6-sup720#show ip slb sessions asn r6 de
7606-R6-sup720#show ip slb sessions asn r6 detail
V-ASNGW-SR, client = 50.35.50.1:2231, virtual = 50.70.80.100:2231
state = ASNR6_ESTAB, real = 11.11.11.50
Key = 0000AAAAC38ECCCC, retry = 1
V-ASNGW-SR, client = 50.35.50.1:2231, virtual = 50.70.80.100:2231
state = ASNR6_ESTAB, real = 11.11.11.50
Key = 0000AAAAC392CCCC, retry = 1
V-ASNGW-SR, client = 50.35.50.1:2231, virtual = 50.70.80.100:2231
state = ASNR6_ESTAB, real = 11.11.11.50
Key = 0000AAAAC396CCCC, retry = 1
V-ASNGW-SR, client = 50.35.50.1:2231, virtual = 50.70.80.100:2231
state = ASNR6_ESTAB, real = 11.11.11.50
Key = 0000AAAAC39ACCCC, retry = 1
V-ASNGW-SR, client = 50.35.50.1:2231, virtual = 50.70.80.100:2231
state = ASNR6_ESTAB, real = 11.11.11.50
Key = 0000AAAAC39ECCCC, retry = 1
V-ASNGW-SR, client = 50.35.50.1:2231, virtual = 50.70.80.100:2231
state = ASNR6_ESTAB, real = 11.11.11.50
< S N I P P E D >
7606-R6-sup720#
Configuring a Virtual Server
To configure a Cisco IOS SLB virtual server, use the following commands, beginning in global configuration mode:
Sample Configuration
Router-SLB(config)# ip slb vserver V-ASNGW-SR
Router-SLB(config-slb-vserver)# virtual 50.70.80.100 udp 2231 service asn r6
Router-SLB(config-slb-vserver)#serverfarm ASNGW-SR-SF
Router-SLB(config-slb-vserver)# idle asn r6 request 90
Router-SLB(config-slb-vserver)#inservice
Configuring DFP Support
You can define Cisco IOS SLB as a DFP manager, as a DFP agent for another DFP manager (such as Distributed Director), or as both at the same time. Depending on your network configuration, you might enter the commands for configuring Cisco IOS SLB as a DFP manager and the commands for configuring Cisco IOS SLB as a DFP agent on the same device or on different devices.
To configure Cisco IOS SLB as a DFP manager, and to identify a DFP agent with which Cisco IOS SLB can initiate connections, use the following commands, beginning in global configuration mode:
Sample Configuration
Router-SLB(config) # ip slb dfp
Router-SLB(config-slb-dfp)# agent 11.11.11.50 5555 10 0 5
Router-SLB(config-slb-dfp)# agent 11.11.11.70 7777 10 0 5
Configuring the BWG
This section describes various other configuration tasks that you need to perform to make the BWG function properly. It includes the following topics:
Configuring SNMP on the BWG
This section provides information on how to configure Simple Network Management Protocol (SNMP) on the BWG. It contains the following configuration tasks:
•Configuring SNMP Access in Routers
•Configuring SNMP-Server Trap-Source
Configuring SNMP Access in Routers
To set up the community access string to permit access to the Simple Network Management Protocol (SNMP), perform the following tasks:
Configuring SNMP-Server Host
To specify the recipient of an Simple Network Management Protocol notification operation, perform the following tasks:
This command is disabled by default. No notifications are sent. If you enter this command with no keywords, the default is to send all trap types to the host. No informs will be sent to this host.
If no version keyword is present, the default is version 1. The no snmp-server host command with no keywords will disable traps, but not informs, to the host. In order to disable informs, use the no snmp-server host informs command.
Configuring SNMP-Server Trap-Source
To specify the interface (and the corresponding IP address) that an Simple Network Management Protocol trap should originate from, perform the following task:
Configuring SNMP Traps
To enable the router to send Simple Network Management Protocol traps or informs (SNMP notifications), perform the following task:
Configuration Examples
The following example enables the router to send all traps to the host specified by the name "myhost.cisco.com", using the community string defined as "public":
snmp-server enable traps
snmp-server host myhost.cisco.com public
The following example enables the router to send Frame Relay and environmental monitor traps to the host "myhost.cisco.com" using the community string "public":
snmp-server enable traps frame-relay
snmp-server enable traps envmon temperature
snmp-server host myhost.cisco.com public
The following example will not send traps to any host. The BGP traps are enabled for all hosts, but the only traps enabled to be sent to a host are ISDN traps (which are not enabled in this example).
snmp-server enable traps bgp
snmp-server host bob public isdn
router(config)# [no] logging snmp-authfail
Note Using the logging snmp-authfail command enables all SNMP authentication failure logging messages. The no version of this command will disable the logging of authentication failure messages.
Note If you do not use the SNMP management tools of the router to monitor PPP sessions, you can prevent the virtual-access subinterfaces from being registered with the SNMP functionality of the router and using memory by using the no virtual-template snmp command. For example:
router(config)# [no] virtual-template snmp
SNMP Configuration Examples on the BWG
Logging
=========
!
logging snmp-authfail
logging queue-limit 100
logging buffered 1000000
enable password lab
!
Virtual Template
=============
!
no virtual-template snmp
!
SNMP Traps
==========
snmp-server community private RW
snmp-server trap-source GigabitEthernet0/2
snmp-server enable traps snmp authentication linkdown linkup coldstart warmstart
snmp-server enable traps vrrp
snmp-server enable traps ds1
snmp-server enable traps tty
snmp-server enable traps eigrp
snmp-server enable traps casa
snmp-server enable traps cnpd
snmp-server enable traps pw vc
snmp-server enable traps syslog
snmp-server enable traps isdn call-information
snmp-server enable traps isdn layer2
snmp-server enable traps isdn chan-not-avail
snmp-server enable traps isdn ietf
snmp-server enable traps ds3
snmp-server enable traps atm subif
snmp-server enable traps channel
snmp-server enable traps ima
snmp-server enable traps srp
snmp-server enable traps flash insertion removal
snmp-server enable traps hsrp
snmp-server enable traps config
snmp-server enable traps entity
snmp-server enable traps fru-ctrl
snmp-server enable traps cpu threshold
snmp-server enable traps config-copy
snmp-server enable traps envmon
snmp-server enable traps aaa_server
snmp-server enable traps agw
snmp-server enable traps bgp
snmp-server enable traps ospf state-change
snmp-server enable traps ospf errors
snmp-server enable traps ospf retransmit
snmp-server enable traps ospf lsa
snmp-server enable traps ospf cisco-specific state-change nssa-trans-change
snmp-server enable traps ospf cisco-specific state-change shamlink interface-old
snmp-server enable traps ospf cisco-specific state-change shamlink neighbor
snmp-server enable traps ospf cisco-specific errors
snmp-server enable traps ospf cisco-specific retransmit
snmp-server enable traps ospf cisco-specific lsa
snmp-server enable traps pim neighbor-change rp-mapping-change invalid-pim-message
snmp-server enable traps ipmulticast
snmp-server enable traps mvpn
snmp-server enable traps msdp
snmp-server enable traps rsvp
snmp-server enable traps frame-relay
snmp-server enable traps frame-relay subif
snmp-server enable traps ipsla
snmp-server enable traps stun
snmp-server enable traps dlsw
snmp-server enable traps bstun
snmp-server enable traps pppoe
snmp-server enable traps l2tun session
snmp-server enable traps l2tun pseudowire status
snmp-server enable traps ipmobile
snmp-server enable traps frame-relay multilink bundle-mismatch
snmp-server enable traps dsp card-status
snmp-server enable traps dsp oper-state
snmp-server enable traps event-manager
snmp-server enable traps alarms informational
snmp-server host 171.71.129.34 public
Mobile to Mobile Traffic Steering
In releases prior to BWG IOS Release 12.4(15)XL3, switching of peer-to-peer traffic only occured locally on the BWG. This feature enables the BWG to force all upstream traffic to the configured next hop address.
You can configure this feature through CLI as well as from AAA, and the name of the attribute is "uplink_nexthop" (a CISCO AVpair attribute). This is per-subscriber provisioning in AAA. Unlike many other AAA-based features, the BWG can dynamically make full use of the uplink_next hop from AAA during re-authentication as well. The "uplink_nexthop" attribute downloaded from AAA takes precedence over the configured "uplink nexthop" in the CLI.
Configuring Mobile to Mobile Traffic Steering
To configure the BWG to force all upstream traffic to the configured next hop address, perform the following tasks:
Here is a sample configuration:
wimax agw user group-list wimax
user-group domain cisco.com
sla profile-name silver
ip redirect traffic all 10.1.1.5
ip static-allowed
!
CAR/AAA Configuration
CAR can be configured multiple ways to enable it to send attributes; one of the ways is to write an outgoing TCL script. This script is executed after basic authentication is performed. The CAR configuration is out of scope of this document, however a sample outgoing script is provided here for convenience.
Sample Outgoing Script on CAR
proc tAddBlaTS {request response environ} {
if {[string compare [$environ get Response-Type] Access-Accept] == 0} {
$response trace 5 "Tcl tAddAttrSLA entry point"
$response put Class "class-wimax-changed"
$response put WiMAX-AAA-Session-ID "A0123"
$response put Termination-Action "RADIUS-Request"
$response put Cisco-AVPair "user-realm=cisco.com"
$response put Cisco-AVPair "sla-profile-name= platinum gold silver" APPEND
$response put Cisco-AVPair "cpe-settings=7" APPEND
$response put Cisco-AVPair "uplink_nexthop=10.10.10.24" APPEND'
$response put Cisco-AVPair "vlan-id=12" APPEND
$response put Cisco-AVPair "static-ip-allowed=1" APPEND
$response put Framed-Route "5.5.0.0/16"
MIB Support
The BWG supports a Management Information Base (MIB) that describes objects that enable users and network management to remotely monitor the BWG using SNMP commands. The BWG supports two separate MIBs:
One contains global system information and parameters, base-station information, subscriber, flow, traffic and trap notification information.
The second contains information about the R6 signaling protocol information used between the base-station and the BWG. This includes overall gateway R6 information, and information per base-station.
The BWG MIB variables are not synchronized across a fail-over. Many MIB variables can be recreated on the standby from the synchronized state data. The NMS attempts to handle such a situation, and any inconsistencies in MIB data that result from this approach. The existing RF/CF MIB is also available.
Note The R6 MIB is not supported in the initial release of the BWG on the Cisco 7301 router platform.
Verifying MIB Support
To display various MIB parameters, perform the following tasks:
Configuration Examples
Here is sample output for the show wimax agw command:
router# show wimax agw
Access network gateway version 0.1, service is enabled
AGW listening on UDP control port 2231
Maximum Number of base station 500 allowed
Maximum Number of subscriber 20000 allowed
Number of signalling paths created 0
Number of brearer paths created 0
Number of subscribers connected 0
Number of sessions created 0
Number of flows created 0
Traffic Sent 0 packets, 0 bytes
Traffic Rcvd 0 packets, 0 bytes
Number of framed routes
Number of subscribers using the framed routes
Current number of user auto-provisioned sessions
The traffic is split for IP CS and ETH CS.
Here is sample output for the show wimax agw user-group command:
router# show wimax agw user-group
AGW User-Group-List
There are 3 user-groups configured in list wimax
User group domain name any
User-Group overwritten Counter 0
Service mode operational
Sessions 0 associated
IP-GRE Traffic Sent 0 packets, 0 bytes
IP-GRE Traffic Received 0 packets, 0 bytes
Eth-GRE Traffic Sent 0 packets, 0 bytes
Eth-GRE Traffic Received 0 packets, 0 bytes
Ingress Address filtering 0 packets, 0 bytes
Traffic Received redirected 0 packets, 0 bytes
User group domain name cisco
Service mode operational
Sessions 0 associated
Traffic Sent 0 packets, 0 bytes
Traffic Received 0 packets, 0 bytes
Ingress Address filtering 0 packets, 0 bytes
User group domain name unauthenticated
Service mode operational
Sessions 0 associated
Traffic Sent 0 packets, 0 bytes
Traffic Received 0 packets, 0 bytes
Ingress Address filtering 0 packets, 0 bytes
router#show wimax agw user-group brief ?
Name Sessions Pkts-Tx Bytes-Tx Pkts-Rx Bytes-Rx VRF
any 0 0 0 0 0
cisco 0 0 0 0 0
unauthenticated 0 0 0 0 0
router#show wimax agw user-group any ?
brief Brief output
| Output modifiers
<cr>
router#show wimax agw user-group any
User group domain name any
--------------------------------------
Service mode operational
Sessions 0 associated
Traffic Sent 0 packets, 0 bytes
Traffic Recevied 0 packets, 0 bytes
Ingress Address filtering 0 packets, 0 bytes
router#show wimax agw user-group any brief
Name Sessions Pkts-Tx Bytes-Tx Pkts-Rx Bytes-Rx VRF
any 0 0 0 0 0
router#show wimax agw user-group name ?
WORD Enter User-group Name
router#show wimax agw user-group name cisco ?
brief Brief output
| Output modifiers
<cr>
router#show wimax agw user-group name cisco
User group domain name cisco
--------------------------------------
Service mode operational
Sessions 0 associated
Traffic Sent 0 packets, 0 bytes
Traffic Recevied 0 packets, 0 bytes
Ingress Address filtering 0 packets, 0 bytes
router#show wimax agw user-group name cisco brief ?
| Output modifiers
<cr>
router#show wimax agw user-group name cisco brief
Name Sessions Pkts-Tx Bytes-Tx Pkts-Rx Bytes-Rx VRF
cisco 0 0 0 0 0
router#show wimax agw user-group unauthenticated ?
brief Brief output
| Output modifiers
<cr>
router#show wimax agw user-group unauthenticated
User group domain name unauthenticated
--------------------------------------
Service mode operational
Sessions 0 associated
Traffic Sent 0 packets, 0 bytes
Traffic Recevied 0 packets, 0 bytes
Ingress Address filtering 0 packets, 0 bytes
asn#sh wimax agw user-group unauthenticated b
asn#sh wimax agw user-group unauthenticated brief ?
| Output modifiers
<cr>
router#show wimax agw user-group unauthenticated brief
Name Sessions Pkts-Tx Bytes-Tx Pkts-Rx Bytes-Rx VRF
unauthenticated 0 0 0 0 0
Here is sample output for the show wimax agw statistics command:
router# show wimax agw statistics
AGW Statistics
Message function type Undefined(0/0x0)
Message function type Data Path(3/0x3)
Message type Deregistration Request(4/0x4)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Deregistration Response(5/0x5)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Deregistration Ack(6/0x6)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Registration Request(12/0xC)
Number of messages sent 0
Number of messages received 0
Number of messages resent 0
Message type Registration Response(13/0xD)
Number of messages sent 0
Number of messages received 0
...
Here is sample output for the show wimax agw statistics dhcp-relay command:
router# show wimax agw statistics dhcp-relay
AGW DHCP Statistics
Tx to DHCP server Discover 0, Request 0
Tx to DHCP server Release 0, Decline 0
Tx to DHCP server Inform 0
Rx from DHCP server Offer 0, Ack 0
Rx from DHCP server Nak 0, Unknown 0
This output gives the statistics of DHCP messages that are relayed through the BWG. If the BWG happened to initiate any DHCP messages, these counters are not incremented.
Here is sample output for the show wimax agw statistics internal command:
Router# show wimax agw statistics internal
Last clearing of "show wimax agw statistics internal" counters 3d23h
Signalling plane related statistics
Signal packets processed messages 483
Signal packets has pending messages 0
Signal packets requeued messages 0
Signal packets dropped too many pending messages 0
Signal packets dropped service disabled 0
Signal packets dropped service not ready 0
Signal packets dropped no encapsulation interface 0
Signal packets dropped CAC denied request 0
Signal packets disposed by agw 0
Data plane related statistics
Data packets not ours encapsulation 0
Data packets not ours encapsulation address 0
Data packets not ours service disabled 0
Data packets not ours invalid protocol type 0
Data packets dropped invalid ip len 0
Data packets dropped absent key data 0
Data packets dropped flow not found 21
Data packets dropped flow path not found 0
Data packets dropped flow path invalid src address 0
Data packets dropped session not found 0
Data packets dropped subscriber not found 0
Data packets dropped checksum error 0
Data packets dropped ingress filtering 0
Data packets dropped sequence number mismatch 0
Data packets dropped invalid redirect address 0
Data packets dropped throttling of punts from cef to process 0
Data packets dropped gateway learning from upstream data packets 0
Data packets dropped non-ARP and non-DHCP L2 multicast/broadcast data packets 0
Data packets punted fragmented 0
Data packets punted from cef path to process path 0
Other related statistics
Number of N/w behind MS in usrgrp enabled 0
Total subscriber created 41
Total subscriber deleted 45
Total session created 41
Total session deleted 45
Total flow created 4
Total flow deleted 7
Total host created 0
Total host deleted 0
Total signalling path created 3
Total signalling path deleted 3
Total data path created 2
Total data path deleted 2
Number of hosts rejected 0
Number of packets dropped due to Static IP Host not allowed 0
Number of static hosts aged out 0
Total sessions rejected due to unapproved BS 0
Configuration related statistics
Service flow profile not found 0
QoS profile not found 0
Classifier profile not found 0
Sla profile not found 0
Handoff related statistics
Total handoffs succeeded 0
Total handoff failed 0
Total cmac key update succeeded 0
Total cmac key update failed 0
Total security key exchange succeeded 0
Total security key exchange failed 0
Miscellaneous statistics
Maximum Subscriber exceeded 0
Maximum BS exceeded 0
MIB Enhancements for BWG Release 1.1
The BWG Management Information Base (MIB) is updated to add Ethernet CS related counters of packets/bytes sent/received. The following MIB objects have been updated in BWG Release 1.1:
•Includes the new EthCS related pkt/bytes sent/receive variables.
•Includes a description of packets/bytes sent/receive counters that were modified.
For example, "The total number of Data Packets received" has been modified to "The total number of IP Data Packets received"
ASN GW Global Statistics
•The total number of IP Data Packets received
•The total number of IP Data Packets Sent
•The total number of IP Data bytes received
•The total number of IP Data bytes sent
•The total number of Ethernet CS Data Packets received
•The total number of Ethernet CS Data Packets Sent
•The total number of Ethernet CS Data bytes received
•The total number of Ethernet CS Data bytes sent
•Total number of hosts rejected
•Total number of static hosts aged out
The MIB also includes new objects for statistics related to ARP (total number of ARP requests received, total number of ARP replies sent, total number of ARP packets dropped), rejected host, and aged out static hosts.
GRE Keying
A GRE key is allocated per service flow in each direction. The GRE keys are exchanged using the RR-Request used to create the Data Path bearer. The GRE keys that correspond to the BWG are allocated by the BWG and sent to the BS during creation of the service flow using the RR-Request. Similarly, the GRE key values corresponding to the BS are allocated by the BS and sent to the BWG using the RR-Response message.
During Inter-BS mobility, new keys are allocated by the Base-station. The BWG keeps the same GRE keys.
The BWG allocates the GRE keys such that the values are not assigned immediately upon release.
VRF Support
A user-group can be configured with virtual route forwarding (VRF) support. This allows you to create an internal VRF entity to connect all traffic to/from the specific user-group.
QoS Support
QoS Support refers to both airlink QoS as well as mapping on the network. The ASNGW is responsible for sending down the qos parameters to the Base-station for creating appropriate service flows.
•Certain hosts can be given additional quality of service parameters
•A new R6 bearer (service flow) may be created corresponding to the hosts IP address. Multiple host can use this service flow
•The mapping of the host to the new R6 service flow is created, and communicated to the BS/MS via the RR-Request.
DSCP Marking Per Service Flow
Each service flow is mapped uniquely to a Diffserv Code Point (DSCP). This DSCP value is used to mark the outer IP header for downstream packets by the BWG, and by the BS for upstream packets.
The inner IP header for upstream and downstream packets is set by the BWG as per the mapping for the service flow, unless explicitly disabled by a CLI.
ACLs
ACLs are supported, and can be configured at a per-user group basis. This applies to all users that connect to the same user-group.
Source IP Address Validation
For all uplink packets, the allocated IP address for the corresponding MS or service flow is validated. If a mismatch is found, those packets are discarded.
To configure this feature, use the security subscriber address-filtering ingress command in gateway user group submode.
Support of Split Control and Data End Points for BS
The BS may have different end point IP addresses for the control and the data plane. Depending on the availability of the Data Path End Point ID TLV (sent in path registration response message from the BS for the flow), the BWG can create the GRE path taking the ipv4 from the available TLV.
If the specified TLV is not present, the control plane end point address is used as the remote data end point to create GRE path.
Bearer Accounting
Bearer volume counts are maintained for all service flows. These include the input and output packets and octet counts.
Restrictions
The following restriction apply in Cisco BWG Release 1.0 and above:
•To avoid issues with high CPU usage, we recommend the following configurations:
–To reduce the CPU usage during bootup, disable logging to the console terminal by configuring the no logging console global configuration command.
–To ensure that the HSRP interface does not declare itself active until it is ready to process a peers Hello packets, configure the delay period before the initialization of HSRP groups with the standby delay minimum 100 reload 100 interface configuration command under the HRSP interface.
–To minimize issues with high CPU usage for additional reasons, such as periods of high PPP PDP processing (creating and deleting), disable the notification of interface data link status changes on all virtual template interfaces of the BWG using the no logging event link-status interface configuration command.
!
interface Virtual-Template1
description ASNGW-VT
ip unnumbered Loopback0
encapsulation agw
no logging event link-status
access-point-list wimax
end
Features Not Supported
The following features are not supported in this release:
•Configuration synchronization: BWG relies on manual configurations to ensure identical feature configuration on the active and standby.
•In-Service-Software-Upgrade (ISSU): Upgrading from the immediate lower software version to a higher version is supported.