what is mean by classful n classless protocols,rip is classful while rip2 is classless,what is the difference between them?
The classful protocols came first (they are the oldest) and they make certain assumptions about IP addressing, about subnetting, and about routing. When these protocols were being developed these assumptions made sense and they made things more simple. As networks grew and became more complex these assumptions did not make as much sense and became restrictive on the development of networks. So the classless protocols do not make those assumptions.
These assumptions include:
- assumption of the definition of class A networks (8 bits for network ID), class B networks (16 bits for network ID), and class C networks (24 bits for network ID).
- assumption that a network could be subnetted but that all subnets of that network must be the same size.
- assumption that all subnets must be contiguous (you can get from any part of a network to any other part of that network without going outside of that network).
- assumption that if you knew some of the subnets of a network that you knew every subnet that existed in that network.
- assumption that you would always summarize at the boundary of a network (if your neighbor router was not in that network then it did not need to know about any subnets in the network, it only needed to know to come through you to get to that network).
- assumption that you could only summarize at network boundaries.
In the early days of networking these assumptions made sense and RIP(v1) and IGRP used them. They made routing easier (you did not need to advertise subnet masks because you already knew what size the subnets would be). You did not ned to worry about whether to summarize or not because it was automatic. You only needed to be aware of subnets of your locally connected networks and could use a summary (or even a default route) to get to all the other networks. But over time these assumptions began to get in the way and were seen as not efficient - why should the subnet of the LAN interface (where there might be 100 devices) be the same size as the serial interface subnet (where there is only one other device)? Different assumptions were made and the classless protocols (like EIGRP and OSPF) were developed.
Classfull routing protocols do not support Variable Length Subnet Mask
(VLSM) whereas classless routing protocols support VLSM. In other words,
classfull routing protocols consider all routing updates to be bound by the
class of the IP address. Classfull routing protocols do not send the mask
information with the routing updates. If they get an update that is not
bound by the class rules, they will apply the class mask (ignore the actual
mask) and then install that route.
On the other hand, classless routing protocols do not worry about the class
of the route but the actual route itself. So, they support and install
specific routes in the routing table.
Here is a document that could be useful:
Hope this helps.