NSF, NSR, and GR [translation]

Source: Internet
Author: User

 

Http://www.networkworld.com/community/node/16964.

The modern high-performance router separates the forwarding board from the control panel on the physical structure, each Board has its own processor and memory. The control panel runs the routing protocol, maintains the necessary database during the routing process, and obtains a forwarding table. The forwarding table is handed over to the forwarding board for packet forwarding.

An important advantage of physically separating the forwarding board from the control panel is that if the traffic is heavy and the load of the forwarding board is large, it will not affect the control panel's ability to process new route information. On the contrary, if the control panel loads heavily (processing a large amount of new route information), it will not affect the forwarding Board's ability to continue to forward packets at high speed.

In fact, the control panel can stop working because the forwarding board can work independently after being copied to the forwarding table. This is called NSF (non-stop forwarding ): the forwarding Board continues to "uncertain" forwarding if the control board stops working.

Of course, this is very dangerous. If the network topology changes when the control panel stops working, the forwarding table of the forwarding board becomes invalid, resulting in incorrect packet forwarding. But why do we still need NSF?

The answer is that there are surplus control panels. nsf can switch from the main control panel to the backup control panel without interrupting forwarding. During this switching process, the intermediate publishing may be invalid, the risks during this period are acceptable.

The shorter the switching time, the lower the risk. Therefore, if the backup control panel maintains an active configuration and the current system status, such as an interface, the switchover will become very fast, because you do not need to learn it in advance. This is a second-level switch. Cisco calls this process SSO (stateful switchover), while Juniper calls it GREs (graceful routing engine switchover ).

The third level is NSR (non-stop Routing ). As mentioned so far, when switching the control panel, the adjacent table is always damaged, even if SSO is used. When Active Control Panel B stops working, any of its neighbors router A finds that the session with him has failed, during this period, a will notify a's neighbors that the path through this router B is invalid. Of course, after the backup board is successfully started, the neighbors and neighbors will recreate the path. This process will greatly damage the network.

The goal of NSR is to block or minimize the impact on peering sessions.

The Graceful Restart (GR) protocol extension is used to control the adjacent table during the control panel switch. Each routing protocol has its own detailed gr extension, but they work in almost the same way. When the control panel of a router stops working, his neighbor does not immediately report to his neighbor that the router is down, but waits for a time (called the grace period ). If the control panel of the router is successfully started before this time limit and a peering session is re-established, the temporarily damaged peering session will not affect the network outside the neighbor.

However, GR has two problems:

1. The neighbor router must support the GR protocol extension. The control panel is highly destructive to the PE (provider edge: Provide edge device) router because PE connects many Ce (customer edge: User edge device ), ce routers generally do not support gr.

2. If the router is completely damaged rather than switched, the GR time limit will slow down the network aggregation (reconvergence.

 

The new generation of NSR uses internal processes to keep the backup control panel aware of the routing protocol status and adjacent maintenance activities. Therefore, after the switchover is complete, the backup board can control the existing peer-to-peer session information instead of creating a new one. The switchover is transparent to neighbors because the NSR process is performed internally and there is no need for neighbors to support any Protocol extensions.

 

NSR (internal)> NSR (GR)> SSO> NSF

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.