Software Configuration Guide for the Cisco 5900 Embedded Services Routers

Prerequisite Reading

Read the following chapter before selecting a RAR protocol:

• Chapter 5, “Introduction to Radio Aware Routing and MANET”

Note You can use only one RAR protocol per interface.

This chapter contains the following sections:

• PPPoE in a MANET
• VMI in a MANET
• PPPoE and VMI
• Configuring PPPoE for use with VMI
• Showing VMI Neighbors

Link-Quality Metrics

The quality of a radio link has a direct impact on the throughput. The PPPoE protocol has been extended to provide a process by which a router can request report link quality metric information. Cisco’s OSFPv3 and EIGRP implementations are enhanced so that the route cost to a neighbor is dynamically updated based on metrics reported by the radio, thus allowing the best route to be chosen within a given set of radio links.

The routing protocols receive raw radio link data, and compute a composite quality metric for each link. In computing these metrics, the router may consider the following factors:

• Maximum Data Rate—the theoretical maximum data rate of the radio link, in scaled bits per second
• Current Data Rate—the current data rate achieved on the link, in scaled bits per second
• Latency—the transmission delay packets encounter, in milliseconds
• Resources—a percentage (0-100) that can represent the remaining amount of a resource (such as battery power)
• Relative Link Quality—a numeric value (0-100) representing relative quality, with 100 being the highest quality

On the router, metrics can be weighted during the configuration process to emphasize or de-emphasize particular characteristics. For example, if throughput is a particular concern, you can weight the throughput metric so that it is factored more heavily into the composite route cost. Similarly, a metric of no concern can be omitted from the composite calculation.

Link metrics can change rapidly, often by very small degrees, which could result in a flood of meaningless routing updates. In a worst case scenario, the network churns almost continuously as it struggles to react to minor variations in link quality. To alleviate this concern, Cisco provides a tunable dampening mechanism that allows the user to configure threshold values. Any metric change that falls below the threshold is ignored.The quality of a connection to a neighbor varies, based on various characteristics of the interface when OSPFv3 or EIGRP is used as the routing protocol. The routing protocol receives dynamic raw radio link characteristics and computes a composite metric that is used to reduce the effect of frequent routing changes.

A tunable hysteresis mechanism allows you to adjust the threshold to the routing changes that occur when the router receives a signal that a new peer has been discovered, or that an existing peer is unreachable. The tunable metric is weighted and adjusted dynamically to account for the following characteristics:

• Current and Maximum Bandwidth
• Latency
• Resources
• Relative Link Quality (RLQ)

Individual weights can be deconfigured and all weights can be cleared so that the cost returns to the default value for the interface type. Based on the routing changes that occur, cost can be determined by the application of these metrics.

Neighbor Signaling

MANETs are highly dynamic environments. Nodes may move into, or out of, radio range at a fast pace. Each time a node joins or leaves the network, topology must be logically reconstructed by the routers. Routing protocols normally use timer-driven “hello” messages or neighbor time-outs to track topology changes, but MANETs reliance on these mechanisms can result in unacceptably slow convergence.

Neighbor up/down signaling capability provides faster network convergence by using link-status signals generated by the radio. The radio notifies the router each time a link to another neighbor is established or terminated by the creation and termination of PPPoE sessions. In the router, the routing protocols (OSPFv3 or EIGRP) respond immediately to these signals by expediting the formation of a new adjacency (for a new neighbor) or tearing down an existing adjacency (if a neighbor is lost). For example, if a vehicle drives behind a building and loses its connection, the router immediately senses the loss and establishes a new route to the vehicle through neighbors that are not blocked. This high speed network convergence is essential for minimizing dropped voice calls and disruptions to video sessions.

When VMI with PPPoE is used and a partner node has left or a new one has joined, the radio informs the router immediately of the topology change. Upon receiving the signal, the router immediately declares the change and updates the routing tables.

The signaling capability provides the following benefits:

• Reduces routing delays and prevents applications from timing out
• Enables network-based applications and information to be delivered reliably and quickly over directional radio links
• Provides faster convergence and optimal route selection so that delay-sensitive traffic such as voice and video are not disrupted
• Reduces impact on radio equipment by minimizing the need for internal queuing/buffering
• Provides consistent Quality of Service (QoS) for networks with multiple radios

The messaging allows for flexible rerouting when necessary because of the following conditions:

• Noise on the Radio links
• Fading of the Radio links
• Congestion of the Radio links
• Radio link power fade
• Utilization of the Radio

Figure 8-3 illustrates the signaling sequence that occurs when radio links go up and down.

Figure 8-3 Up and Down Signaling Sequence

PPPoE Credit-based Flow Control

Each radio initiates a PPPoE session with its local router as soon as the radio establishes a link to another radio. Once the PPPoE sessions are active for each node, a PPP session is then established end-to-end (router-to-router). This process is duplicated each time a radio establishes a new link.

The carrying capacity of each radio link may vary due to location changes or environmental conditions, and many radio transmission systems have limited buffering capabilities. To minimize the need for packet queuing in the radio, Cisco has implemented extensions to the PPPoE protocol that enable the router to control traffic buffering in congestion situations. Implementing flow-control on these router-to-radio sessions also allows the use of fair queuing.

The flow control solution utilizes a credit-granting mechanism documented in RFC 5578. When the PPPoE session is established, the radio can request a flow-controlled session. If the router acknowledges the request, all subsequent traffic must be flow-controlled. If a flow control session has been requested and cannot be supported by the router, the session is terminated. Typically, both the radio and the router initially grant credits during session discovery. Once a device exhausts its credits, it must stop sending until additional credits have been granted. Credits can be added incrementally over the course of a session.

High performance radios that require high-speed links use metrics scaling. The radio can express the maximum and current data rates with different scaler values. Credit scaling allows a radio to change the default credit grant (or scaling factor) of 64 bytes to its default value. You can view the maximum and current data rates and the scalar value set by the radio from the output of the show vmi neighbor detail command.

Point-to-Point Protocol over Ethernet

Cross-layer feedback for router-radio integration radio aware routing takes advantage of the functions defined in RFC 5578. RFC 5578 is an Internet Engineering Task Force (IETF) standard that defines Point-to-Point Protocol over Ethernet (PPPoE) extensions for Ethernet-based communications between a router and a device such as a mobile radio that operates in a variable-bandwidth environment and has limited buffering capabilities. These extensions provide a PPPoE session based mechanism for sharing radio network status such as link-quality metrics and establishing flow control between a router and an RFC 5578-capable radio.

An RFC 5578 radio initiates an L2 PPPoE session with its adjacent router on behalf of every router and radio neighbor discovered in the network. These L2 sessions are the means by which radio network status for each neighbor link is reported to the router. The radio establishes correspondence between each PPPoE session and each link to a neighbor.

Continuing with PPPoE Configuration

This chapter provides the following major sections to describe how to configure Point-to-Point Protocol over Ethernet (PPPoE) on a specific interface.

• Configuring PPPoE for use with VMI
• Showing VMI Neighbors

Note See Appendix A, “Command Reference” for detailed command reference.

Creating a Subscriber Profile

Perform this task to configure a subscriber profile for PPPoE service selection.

Note Configuring a subscriber profile for PPPoE service selection is required for VMI to function properly.

SUMMARY STEPS

1. enable

2. configure terminal

3. exit

4. subscriber authorization enable

DETAILED STEPS

Configuring PPPoE Service Selection

Perform this task to associate the subscriber profile with a PPPoE profile. In this configuration, the Broadband Access (BBA) group name must match the subscriber profile name defined in the subscriber profile.

Note In this example, manet_radio serves as the subscriber profile name.

SUMMARY STEPS

1. enable

2. configure terminal

3. bba-group pppoe { group-name | global }

4. virtual-template template-number

5. service profile subscriber-profile-name [ refresh minutes ]

6. end

DETAILED STEPS

Configuring PPPoE on an Ethernet Interface

Perform this task to assign a PPPoE profile to an Ethernet interface.

SUMMARY STEPS

1. enable

2. configure terminal

3. interface [ type slot / port]

4. pppoe enable [ group group-name ]

5. end

DETAILED STEPS

Configuring a Virtual Template Interface

Perform this task to configure a virtual-template interface. The virtual-template interface is required to clone configurations. For each VMI neighbor, a new virtual-access interface will be created dynamically.

SUMMARY STEPS

1. enable

2. configure terminal

3. no virtual-template subinterface

4. policy-map policy-map-name

5. class class-default

6. fair-queue

7. exit

8. interface virtual-template 1

9. ip unnumbered vmi1

10. service-policy output FQ

11. keepalive 60 20

12. end

Detailed Steps

Example Configuration

Mapping Outgoing Packets

Perform this task so that VMI can map outgoing packets to the appropriate PPPoE sessions. VMI will use the next-hop forwarding address from each outgoing packet perform this mapping.

SUMMARY STEPS

1. enable

2. configure terminal

3. interface vmi interface-number

4. ip address ip_addr subnet_mask

5. physical-interface interface-type/slot

6. end

Detailed Steps

Examples

The following examples show the IP address coordination needed between virtual-template configuration and VMI configuration.

VMI in Aggregate Mode for IPv6

The following example shows the configuration of VMI in aggregate mode for IPv6.

VMI in Aggregate Mode for IPv4

The following example shows the configuration of VMI in aggregate mode for IPv4.

VMI in Aggregate Mode for IPv4 and IPv6

The following example shows the configuration of VMI in aggregate mode for IPv4 and IPv6.

Configuring Multicast Support

This section identifies the recommended modes and tasks for working with multicast:

• Using Aggregate Mode
• Using Bypass Mode
• Enabling Multicast Support on a VMI

Using Aggregate Mode

VMI operates in aggregate mode by default. All of the virtual-access interfaces created by PPPoE sessions are aggregated logically under the configured VMI. Applications above Layer 2 (L2), such as Enhanced Interior Gateway Routing Protocol (EIGRP) and OSPFv3, should be defined only on VMI. Packets sent to VMI are forwarded to the correct virtual-access interface(s). Aggregate mode VMIs operate in Non-Broadcast Multiple Access (NBMA) mode. Multicast traffic is forwarded only to the NBMA neighbors where a listener for that group is present. This is the preferred mode when operating in PIM sparse mode.

Note NBMA multicasting only supports IPv4 and sparse mode.

Perform this task to configure interface vmi1 to operate in NBMA mode and PIM sparse mode:

SUMMARY STEPS

1. enable

2. configure terminal

3. interface vmi interface-number

4. ip address ip_addr subnet_mask

5. ip pim nbma-mode

6. ip pim sparse-mode

7. load-interval number

8. physical-interface interface-type/slot

9. end

Detailed Steps

	Command or Action	Purpose
Step 1	enable   Router> enable	Enables privileged EXEC mode. Enter your password if prompted.
Step 2	configure terminal   Router# configure terminal	Enters global configuration mode.
Step 3	interface vmi number   Router(config)# interface vmi1	Creates a VMI interface.
Step 4	ip address ip_addr isubnet_mask   Router(config-if)# ip address 10.2.2.2 255.255.255.0	Specifies the IP address and subnet mask for the VMI interface. This example sets the IP address to 10.2.2.2 and the subnet mask to 255.255.255.0
Step 5	ip pim nbma-mode	Enables NBMA mode.
Step 6	ip pim sparse-mode	Enables sparse mode. Note You must set this to sparse mode.
Step 7	load interval seconds   Router(config-if)# load-interval 30	Specifies the load interval in seconds. This example sets the load interval to 30 seconds.
Step 8	physical-interface interface-type/slot   Router(config-if)# physical-interface fa0/0	Creates the physical subinterface to be associated with VMI on the router.
Step 9	end   Router(config-if)# end	(Optional) Exits the configuration mode and returns to privileged EXEC mode.

Using Bypass Mode

Using bypass mode is recommended for multicast applications.

In bypass mode, the virtual-access interfaces are directly exposed to applications running above L2. In bypass mode, you must still define a VMI because VMI continues to manage presentation of cross-layer signals, such as, neighbor up, neighbor down, and metrics. However, applications will still be aware of the actual underlying virtual-access interfaces and send packets to them directly.

Using bypass mode can cause databases in the applications to be larger because knowledge of more interfaces are required for normal operation.

If you are running multicast applications that require virtual-access interfaces to be exposed to applications above L2 directly, you can configure VMI to operate in bypass mode. Most multicast applications require that the virtual-access interfaces be exposed directly to routing protocols in order for the multicast Reverse Path Forwarding (RPF) to operate as expected. When you use the bypass mode, you must define a VMI to handle cross-layer signals such as neighbor up, neighbor down, and metrics. Applications will be aware of the actual underlying virtual-access interfaces, and will send packets to them directly. Operating VMI in bypass mode can cause databases in the applications to be larger than normally expected because knowledge of more interfaces is required for normal operation.

Enabling Multicast Support on a VMI

Perform this task to enable bypass mode on a VMI and override the default aggregation that occurs on VMI. This configuration assumes that you have already configured a virtual template and appropriate PPPoE sessions for VMI.

After you enter the enable bypass mode, Cisco recommends that you copy the running configuration to Non-Volatile Random Access Memory (NVRAM) because the default mode of operation for VMI is to logically aggregate the virtual-access interfaces.

SUMMARY STEPS

1. enable

2. configure terminal

3. interface vmi number

4. mode bypass

5. exit

DETAILED STEPS

Examples

Note VMI is required to have IP addresses assigned for VMI to work even though it will be shown as down/down while in bypass mode.

VMI in Bypass Mode for IPv6

The following example shows the configuration of VMI in bypass mode for IPv6.

VMI in Bypass Mode for IPv4

The following example shows the configuration of VMI in bypass mode for IPv4.

Note The IPv4 address configured on VMI will not be advertised or used. Instead, the IPv4 address on the virtual-template will be used.

VMI in Bypass Mode for IPv4 and IPv6

The following example shows the configuration of VMI in bypass mode for IPV4 and IPv6.

Example

The following example shows the details about known VMI neighbors.