Talk:Routing Information Protocol

From Wikipedia, the free encyclopedia

Contents

[edit] ARPANET

I didn't add the mention of the ARPANET, merely corrected] the erroneous reference that was already there, from "RIP was first developed in 1969 as part of ARPANET" to "The algorithm used in RIP, the Bellman-Ford algorithm, was first deployed in 1969 as part of the ARPANET." Noel (talk) 01:06, 1 October 2005 (UTC)

Could a list of entry level Products which support RIP be added, so others can learn by 'doing'?

[edit] "above the network layer" ???

<SSPecteR> I really didn't liked this phrase: "It runs above the network layer of the Internet protocol suite, using UDP port 520 to carry its data." Isn't it from the network layer? What does it mean as "above"? application layer? Heaven? Please, someone specify it. ♠ Saturday, 2006-06-10, 06:23 (UTC)

It is a layer management protocol for the network layer, and thus a network layer protocol even though it runs over UDP (Port 520).Hcberkowitz 20:01, 23 June 2007 (UTC)
I have to disagree here, it's not implimented in the base stack, can be interchanged with any number of other protocols, and is dependant directly on the transport control layer, thus it's an application layer protocol. (In that it is implimented at that layer.)
Sorry, but this is not true architecturally either from the full OSI (i.e., beyond seven simple layers) or the IETF perspectives. If you are going to cite OSI layering, then please consider the Management Annex (ISO 7498/4) to the basic OSI Reference Model, and the OSI Routeing Framework. Additional ISO work on the user, control, and management planes, under the ATM architectural work, made very clear that there are parallel stacks, not a "base stack". Routing protocols are in a parallel management stack. OSI routing and management documents are quite explicit that it is the nature of the payload, not the means of encapsulation, that determines what protocol is being managed. As you point out, basic ISIS does not run over a network layer protocol, but its total purpose is a set of functions invisible to the end user, or what I assume you mean by "base stack."
Since RIP, in any event, is not used in any OSI stack, how it is characterized by a simplified version of OSI architecture is irrelevant. The authoritative reference is IETF, and the Routing Area of the IETF, with responsibility for all IP-related routing protocols, simply does not use the OSI terminology. I challenge you to find one standards-track RFC that puts a routing protocol into the application layer. SNMP is an application-layer management protocol that does affect routing, but it does not exchange information about routes (e.g., network layer reachability information).Howard C. Berkowitz 14:11, 8 August 2007 (UTC)

It manages a lower layer, so it does have to reach back down, but all routing protocols have to do this and many are not listed as network layer ptotocols. (At least one, IS-IS, doesn't use IP for their own communications at all.)

206.192.18.14 13:29, 8 August 2007 (UTC)
First,

[edit] convergence time

I dont see why one says that convergence time is bad in RIP As long as prefixes are not "withdrawn", RIP works just as well as say BGP. The metric /hops is a good discriminator to choose the correct path. the older versions of RIP had something similar to Origin identifier at a time, which was used to cut out loops —The preceding unsigned comment was added by 220.227.114.98 (talk) 09:24, 11 December 2006 (UTC).

In real networks, prefixes are frequently withdrawn, as routers or links go down. One of the main reasons to have a routing protocol is to facilitate failover, and RIP, depending on timer settings, can take 90-180 seconds to declare a route down and reconverge. With timer tuning, modern OSPF implementations can reconverge sometimes in less than a second, and often in low seconds. Hcberkowitz 20:00, 23 June 2007 (UTC)

[edit] RIP-2B / RIP-2M

my router has these two options but the instruction manual is not that detailed. does anyone know what are the differences? it should be mentionned in the article too. Louis R14 18:16, 5 February 2007 (UTC)

They're probably RIPv2 over broadcast and RIPv2 over multicast. Use multicast if your network is fully RIPv2, broadcast if you still have RIPv1 routers. Jec 00:47, 6 July 2007 (UTC)

[edit] question

is it useful to deactivate RIP function in a single router network? Louis R14 18:18, 5 February 2007 (UTC)

Not always, if the hosts passively listen for RIP updates to learn the address of the default router. This is common with UNIX hosts, especially which run the routeD package in passive mode.Hcberkowitz 20:01, 23 June 2007 (UTC)