- Preface
- Product Overview
- Using the Command Line
- Configuring the Interfaces
- IP Mobility
- Introduction to Radio Aware Routing and MANET
- Understanding and Configuring DLEP
- Configuring R2CP
- Configuring PPPoE
- OSPFv3 Address Families
- Configuring OSPFv3 for a MANET
- Configuring EIGRP in a MANET
- Understanding and Configuring IP Multiplexing
- Zeroization
- Command Reference
- System Message Overview
- Technical Support Reference
Configuring OSPFv3 for a MANET
This chapter provides the following major sections for configuring OSPFv3 in a Mobile Ad-hoc Network (MANET):
OSPFv3 for MANET
Configuring OSPFv3 for a MANET has similar requirements to many traditional OSPFv3 configurations. The primary difference is to configure the network type of OSPFv3 as a MANET. To optimize the use of OSPFv3 with MANETs, Cisco IOS implements extensions to OSPFv3 as defined in IETF RFC 5820. The result is a well-understood routing protocol designed for a constantly changing network topology constrained by limited bandwidth.
This is accomplished in several ways:
- Radio Aware Routing (RAR): Provides tight coupling of OSPFv3 with cooperative radios (fast convergence and re-convergence through neighbor-presence indicators). Determines accurate, real-time, link-metric costs.
- Incremental Hello: Minimizes OSPFv3 packet size.
- Caching Multicast Link-State Advertisements (LSAs): Minimizes OSPFv3 packet transmissions.
- Optimized Flooding (Overlapping Relay): Minimizes the number of flooded LSAs.
- Selective Peering: Reduces OSPFv3 network overhead by limiting redundant full-peering adjacencies.
Cooperative Radios
While non-cooperative radios are supported, OSPFv3 in a MANET operates best when used with cooperative radios, which is a configuration requiring Virtual Multipoint Interfaces (VMIs). See Chapter 5, “Introduction to Radio Aware Routing and MANET” for detailed procedures.
Note This document defines a Cooperative radio as a radio containing the firmware and software required to support RAR-based flows.
Initial Configuration Procedures
Configuring OSPFv3 for a MANET begins with the following tasks:
2. Enabling IPv6 on the Interface
3. Configuring the OSPFv3 Process
4. Configuring the Interface for OSPFv3 MANETs
Enabling IPv6 Routing
This task enables IPv6 packet forwarding and IPv6 routing, both disabled by default.
SUMMARY STEPS
DETAILED STEPS
|
|
|
---|---|---|
|
||
|
||
|
||
|
Exits global configuration mode and returns the router to privileged EXEC mode. |
Enabling IPv6 on the Interface
This task enables IPv6 on an interface—a prerequisite to configuring OSPFv3 on the interface. IPv6 is disabled by default.
SUMMARY STEPS
DETAILED STEPS
Configuring the OSPFv3 Process
This task configures the OSPFv3 process for IPv6 or IPv4.
Note The commands in this task indicate IPv6. If you want to configure the OSPFv3 process for IPv4 instead, see the detailed steps for examples.
SUMMARY STEPS
4. router-id [OSPFv3 router-id in IP address format]
DETAILED STEPS
Configuring the Interface for OSPFv3 MANETs
This configures the OSPFv3 process for IPv6 or IPv4.
Note The commands in this task indicate IPv6. If you want to configure the OSPFv3 process for IPv4 instead, see the detailed steps for examples.
SUMMARY STEPS
4. ospfv3 [ process-id ] area area-id ipv6 [instance instance-id]
DETAILED STEPS
EXAMPLE
Radio Aware Routing in a MANET
This section describes how to configure OSPFv3 in MANETs for Radio Aware Routing (RAR).
Prerequisites
All radios in OSPFv3 for MANET must be cooperative radios attached to a Virtual Multipoint Interface (VMI).
Before performing the tasks in this section, you must configure OSPFv3 for MANETs on a VMI as described in Chapter 5, “Introduction to Radio Aware Routing and MANET”.
Link Metrics
Cooperative radios in the MANET report link-quality metrics, which can include the following information:
- Maximum Data Rate—the theoretical maximum data rate of the radio link, in bytes per second
- Current Data Rate—the current data rate achieved on the link, in bytes 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
Fine-Tuning RAR Configurations
You can fine-tune RAR configurations within a MANET by converting the link metrics to OSPFv3 link costs and configuring a hysteresis threshold. Configuring a hysteresis threshold on the resultant link costs helps minimize the propagation of LSAs responding to link-metric changes.
Metrics can be weighted during the configuration process to emphasize or de-emphasize particular characteristics. For example, if throughput is highly important, the metric for Current Data Rate (CDR) could be weighted more heavily into the composite metric. Similarly, a metric that is of no concern can be omitted.
Link metrics can change rapidly, often by very small degrees, which can result in a flood of meaningless routing updates. In a worst case scenario, the network will churn 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.
A tunable hysteresis mechanism allows users 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 is adjusted dynamically to account for the following characteristics:
Individual weights can be deconfigured and all weights cleared so that the cost is set back 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.
The dynamic cost metric used for interfaces is computed based on the Layer 2 (L2) feedback to Layer 3 (L3), where the metric calculations are as follows:
OC = maximum-data-rate
S1 = ospfv3 6 dynamic weight throughput (Bandwidth component)
S2 = ospfv3 6 dynamic weight resources (Resources component)
S3 = ospfv3 6 dynamic weight latency (Latency component)
S4 = ospfv3 6 dynamic weight L2 factor (L2 factor component)
Note While the commands and output in this section reflect IPv6 configurations, all examples and commands work for IPv4 as well.
Throughput = (current-data-rate)/(maximum-data-rate)
Router-dynamic cost = OC + (S1) + (S2) + (S3) + (S4)
For a dynamic cost to have the same cost as a default cost, all parameters must equal zero.
Each L2 feedback can contribute a cost in the range of 0 to 65535. To tune down this cost range, use the optional weight keyword in conjunction with the throughput, resources, latency, or L2-factor keyword. Each of these weights has a default value of 100 percent and can be configured in a range from 0 to 100. When 0 is configured for a specific weight, that weight does not contribute to the OSPFv3 cost.
Because cost components can change rapidly, you may need to dampen the amount of changes in order to reduce network-wide churn. Use the optional hysteresis keyword with the threshold threshold-value keyword and argument to set a cost change threshold. Any cost change below this threshold is ignored.
You can use the hysteresis keyword to specify a hysteresis value based on the percentage of change of the currently stored value in the routing table for the peer.
Each time the router receives a new PADQ packet from the radio for a peer, a new cost will be calculated for it. The hysteresis keyword specifies the amount of change required before saving the new value.
The hysteresis percent calculated is performed as follows:
If the absolute value of (new_cost - saved_cost) is greater than (hysteresis_percent*saved_cost), then the new_cost will be saved.
Because cost components can change rapidly, it might be necessary to dampen the volume of changes to reduce network-wide churn. The recommended values for S2, S3, and S4 are based on network simulations that may reduce the rate of network changes. The recommended value for S1 is zero to eliminate this variable from the route cost calculation.
While each network might have unique characteristics that require different settings to optimize actual network performance, these are recommended values intended as a starting point for optimizing a OSPFv3 network. Table 10-1 lists the recommended value settings for OSPFv3 cost metrics.
|
|
|
|
---|---|---|---|
The overall link cost is computed using the following formula:
EXAMPLE
To illustrate these settings, the following example shows how OSPFv3 cost metrics can be defined for a VMI interface with one type of radio:
ospfv3 6 cost dynamic weight throughput 0
ospfv3 6 cost dynamic weight resources 29
EXAMPLE
The following is an IPv6 example of configuration:
Selective Peering for Efficiency
Use selective peering to minimize network costs by minimizing each node’s redundant adjacencies. For each OSPFv3 MANET node, you can restrict full-peering rights to the adjacent neighbors that enhance reachability while remaining cost-effective. For each neighbor adjacency calculated to cause excessive link costs, you can use selective peering to keep that neighbor in a 2-way state. This reduces the need for control-plane bandwidth by reducing database exchanges and routing updates.
Note Selective peering does not reduce dataplane connectivity. User traffic will flow over 2-way links when provided with the best path through the network.
Determining Peering Criteria
Upon discovery of each new neighbor within an OSPFv3 MANET node, selective peering determines whether the forming of an adjacency is cost-effective:
- Yes—Form the full-peering adjacency if the neighbor is not in the OSPFv3 link-state database or reachable via the Shortest Path Tree (SPT).
- No—Instead of forming a full-peering adjacency, maintain a 2-way state when the neighbor is in the OSPFv3 link-state database, reachable, and configured with a redundant-path threshold.
Because dynamic topologies can cause a neighbor path redunancy level to fall below the configured threshold, selective peering can change a neighbor 2-way state to full peering.
Link Costs
Selective peering includes link cost as a factor when determining adjacency formation. Ideally, only the links having the lowest costs are granted full-peering adjacency. You can configure OSPFv3 link costs manually, and with cooperative radio interfaces, link costs are obtained directly from the radios through the VMI.
Working with selective peering involves the following tasks:
Enabling Selective Peering
This task explains how to enable OSPFv3 selective peering for IPv6 or IPv4.
Note The commands in this task indicate IPv6. If you want to configure the OSPFv3 process for IPv4 instead, see the detailed steps for examples.
SUMMARY STEPS
3. router ospfv3 [ process-id ]
4. address-family ipv6 unicast
5. manet peering selective [redundancy <level>] [per-interface]
DETAILED STEPS
Preventing Full Peering over Poor Links
You can prevent full peering over poor links by performing this optional task, which configures the following:
- Configure OSPFv3 to wait for link metrics before considering a neighbor for OSPFv3 peering. (A cooperative radio may not advertise link metrics to the router before being discovered as a new OSPFv3 neighbor.)
- Configure OSPFv3 with a minimum metric threshold. If the radio-reported link metric is above this threshold, the neighbor will be held in 2-way state.
Note The commands in this task indicate IPv6. If you want to configure the OSPFv3 process for IPv4 instead, see the detailed steps for examples.
SUMMARY STEPS
4. ospfv3 [ process-id ] manet peering link-metrics [<threshold>]
DETAILED STEPS
Fine-Tuning Selective Peering
This section describes how to optimize dynamic path costs by means of fine-tuning selective peering. Given a scenario without fine-tuning, each one-hop neighbor is awarded full-peering capabilities upon discovery, regardless of link cost:
- Selective-peering redundancy level is greater than zero
- Link metrics are good (as determined by the configuration settings established in the “Preventing Full Peering over Poor Links” section)
As each additional neighbor is discovered, dynamic path costs are measurable immediately. To minimize path costs dynamically, you can configure the higher-cost links to remain in 2-way states until other peering opportunities become available.
Higher Costs without the Fine-Tuning
Consider the topology shown in Figure 10-1.
Given the example shown in Figure 10-1, we have a static snapshot of a dynamic topology, beginning from this point:
- The redundancy level is set to 1 (the default value)—Router A attempts to maintain two paths for each one-hop neighbor.
- From the perspective of Router A, established neighbor sessions exist only with Routers B and C. Router D will join later.
- Router A has a full-peering relationship established with each of these known routers (B and C).
- The link cost for each of these neighbor sessions has a value of 50.
- At this point, only Router B has a link up to Router D—its peering relationship is full, and the link cost has a value of 30.
Change is then introduced between Router A and Router D:
1. Router D comes into radio range of Router A with a link cost of 70.
2. Router A establishes a full-peering relationship with this new neighbor. (The number of paths from Router A to Router D is currently 1 (through Router B).
The conclusion in this scenario (assigning full-peering capabilities between Routers A and D) is allowed given the original condition specified—the selective-peering redundancy level being greater than zero.
Improved Cost-Effectiveness through Fine-Tuning
To prevent the kind of scenario described in the “Higher Costs without the Fine-Tuning” section, you can fine-tune selective peering so that Routers A and D remain in a 2-way state until the link cost improves or an additional router comes into range—one with better link costs available to both routers (A and D).
Cost Thresholds for Redundant Paths
Setting a redundant-path cost threshold requires each redundant path to cost less than the existing, best path cost by a minimum value. For example, if the best link cost is 80, and you set the threshold value to 20, the new link cost must be less than 60 (80 minus 20).
Note The incremental improvement can be an absolute value or percentage.
Given the topology from Figure 10-1, if you set the redundant-path cost threshold to 20, you can prevent full peering between Routers A and D. This changes the outcome of our scenario, then, as follows:
1. Router D comes into radio range of Router A with a link cost of 70.
2. Selective peering compares link costs:
- 80—Existing link cost between Routers A and D; the sum of link costs via Router B (50 + 30)
- 70—The additional link cost between Routers A and D, if full peering is granted
3. The additional link cost (70) is incrementally better than the existing link cost (80) by a value of 10.
4. The incremental improvement (10) does not meet the minimum threshold (20); therefore, Routers A and D remain in the 2-way state.
Note The commands in this task indicate IPv6. If you want to configure the OSPFv3 process for IPv4 instead, see the detailed steps for examples.
SUMMARY STEPS
4. ospfv3 [ process-id ] manet peering cost {threshold <0-65535> | percent <0-100>}
DETAILED STEPS
Verifying OSPFv3 MANET Configuration and Operation
You can use any combination of the commands listed in this section to check the operation status of OSPFv3 MANET for IPv6 or IPv4. See Appendix A, “Command Reference” for detailed command reference.
Note You must be in privileged EXEC mode to enter the command listed in this section.
EXAMPLES
The examples in this section show how you can use the show ospfv3 command to display general information about the OSPFv3 router process.