Can anybody tell how RIP is working on cisco router. Q. how RIP is sending and receiving routing information contained packet . is it through IOS or directely it is sending and receiving from hardware through hardware abstraction layer(means without intervension of IOS).

Regards, Arjun Prasad

Reply to
Arjun Prasad
Loading thread data ...

IOS is the operating system name for a very wide range of devices with different capabilities and different implementations. It isn't even just one CPU type: Cisco uses a range of completely different CPU families on different devices.

The mechanisms for sending out routing information for one device might be completely different than on a different device -- or can differ in important ways on the same device between different software revisions.

I am fairly sure that on some devices, the routing information must go through the CPU path mediated by IOS -- there are IOS devices that just aren't sophisticated enough to have distributed computing of any kind.

I -believe-, based upon my recollection of reading a few years ago about some of the device architectures, that on -some- systems, routing information is sent out at the card level, using copies of the information that is cached on the card, with the CPU path and IOS responsible (I think) for processing routing updates and signalling the cards that new versions of the data are available.

Thus, as far as I know, the answer to your question is "It is done in different ways on different devices and different software versions."

Reply to
Walter Roberson

While Walter is correct, my understanding is that in most lower-end routers (38xx, 28xx, 18xx, 8xx), the routing tables are built using CEF in the control plane, regardless of the protocol used. CEF is performed in software, but it can be done in hardware in some cases.

The control plane functions like this: traditionally a router exchanges routes (prefixes) with it's neighbors and builds what is called a RIB table. CEF then takes this information and adds an entry for each prefix into the FIB, which includes the next-hop, and outgoing interface. Then, a CEF adjacency table is built that has a pre-built header for all future packets to be forwarded. As packets travel into the router, the existing Layer 2 headers are stripped, and these new pre-made headers are copied onto the packets at very high speeds. This helps get forwarding rates to near wire-rate speeds.

Now, in the case of high-end devices, these sometimes use what's known as dCEF, or distributed CEF. This is essentially the same process, which takes place in the main routing engine, but once those pre-made headers are created, they are sent down to each line card and cached. This keeps the header write process on the local card, and further improves forwarding rates. When a new prefix is added, or a change occurs, the central CPU again recalculates the new header and updates the line cards.

Hope this helps.

Reply to
response3 Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.