本帖最后由 linzha3 于 2015-4-19 15:57 编辑
Cisco nonstop forwarding (NSF) works with SSO to minimize the amount of time a network is unavailable to its users following a switchover. When a networking device restarts, all routing peers of that device usually detect that the device went down and then came back up. This down-to-up transition results in what is called a "routing flap," which could spread across multiple routing domains. Routing flaps caused by routing restarts create routing instabilities, which are detrimental to the overall network performance. Cisco NSF helps to suppress routing flaps, thus improving network stability.
Cisco NSF allows for the forwarding of data packets to continue along known routes while the routing protocol information is being restored following a switchover. With Cisco NSF, peer networking devices do not experience routing flaps. Data traffic is forwarded through intelligent line cards (dual forwarding processors (FPs) on Cisco 10000 series Internet routers) while the standby RP assumes control from the failed active RP during a switchover. The ability of line cards (and FPs on Cisco 10000 series devices) to remain up through a switchover and to be kept current with the FIB on the active RP is key to Cisco NSF operation.
A key element of Cisco NSF is packet forwarding. In Cisco networking devices, packet forwarding is provided by CEF. CEF maintains the FIB, and uses the FIB information that was current at the time of the switchover to continue forwarding packets during aswitchover. This feature eliminates downtime during the switchover.
Cisco NSF supports the BGP, IS-IS, and OSPF routing protocols. In general, these routing protocols must be SSO-aware to detect a switchover and recover state information (converge) from peer devices. Each protocol depends on CEF to continue forwarding packets during switchover while the routing protocols rebuild the Routing Information Base (RIB) tables.
思科NSF支持与SSO协同工作，用以最小化网络不可用的时间；当网络设备进行重启时，所有的路由节点将检测到设备下线并重新上线。这种下线/上线的转化导致路由抖动，抖动可能在多个路由域之间传递；路由抖动引发路由更新进而导致路由的不稳定性，对网络整体性能造成不利影响，思科NSF通过抑制路由抖动来提高网络的稳定性。
思科NSF允许在引擎切换发生并且在已经形成的路由协议信息的恢复过程中，数据包转发几乎不会中断。对端网络设备不会经历路由抖动。NSF技术的关键点在于，当数据流量经由线卡进行转发，同时在引擎切换发生时，备引擎的FIB（转发信息库）与主引擎保持同步。
思科NSF技术的关键点在于包转发。在思科网络设备中，包转发能力由CEF（思科快速转发）实现。CEF维护FIB表，并且利用在引擎发生切换时FIB表当前的转发信息来继续转发流量。该特性极大减少了引擎切换所需的downtime时间。
思科NSF技术同时支持BGP,IS-IS,OSPF等路由协议。通常这些路由协议必须为SSO-aware状态，意为能够检测到引擎切换以及从冗余引擎的状态信息中完成快速收敛。在引擎切换并且RIB（路由信息库）重建时，每种路由协议都可以依托当下的CEF信息进行不间断包转发。
实验验证：
(dual supervisor SSO) OSPF (dual supervisor SSO) SVI 10.1.1.1/24
SVI 10.1.1.2/24 6509-1 T1/1 --------------------- T1/2 6509-2-loopback0 1.1.1.1/32
|
|
SVI 10.1.1.3/24
3750
测试步骤：
1.两台分别为双引擎的6509交换机直连并启用SSO，直连接口配置IP地址；6509-2启用环回口并启用ospf传递环回地址路由至6509-1和3750。
2.使用3750接口长ping 6509-2 10.1.1.2，6509-1透传测试，进行6509-1 SSO引擎切换。
3.使用3750接口长ping 6509-2 loopback0，6509-1三层转发，进行6509-1 SSO引擎切换。
4.打开6509-1 ospf的NSF特性，使用3750接口长ping 6509-2 loopback0，6509-1三层转发，进行6509-1 SSO引擎切换。
.
..
...
省略测试过程中的日志
实验结果：
1.SSO切换，2层测试，ping丢弃1-2个包。
2.SSO切换，3层测试且ospf未启用NSF，ping丢弃9-10个包。
3.SSO切换，3层测试且ospf启用NSF，ping丢弃1-2个包。
结论：
1.基于SSO的NSF默认启用，在引擎发生切换时只需2-4秒即可恢复二层数据包转发能力。
2.基于ospf的NSF默认关闭，切换发生时需要被三层转发的流量需要等到路由信息完全收敛方可恢复三层数据包转发能力。
3.基于ospf的NSF在打开后，切换发生时三层流量依然可以通过当前的CEF表信息进行转发，与二层数据包转发能力恢复时间同步。
实验结论仅供参考。