Joseph W. Doherty Tue, 08/12/2008 - 04:55
User Badges:
  • Super Bronze, 10000 points or more

Summarization "rolls-up" multiple subnets into a larger network address block that includes the smaller network address block.

For instance, suppose you have router that has four point-to-point links, each with a /30 network address. The router could advertise to another router (one not connected via one of the four p-2-p links) each /30. But if all four /30s belong to the same /28, only one /28 summary address need to be passed to the other router rather than the four /30 addresses.

If the four /30s were not covered by the same /28, it still might be possible to use a larger summary address. Perhaps all four /30s are within the same /25.

When you start using summary addresses, you don't want to have both a "real" subnet /x and a summary /x.

In the above, the /28 address space is totally allocated by the /30s, but for the /25 not all it's address space is used. When using a routing protocol that supports VLSM, it's possible to have discontinous subnets, e.g. another /30 within the /25 address space could be elsewhere in the network, but somthing to avoid since it impacts the primary advanage of summarization, i.e. to avoid passing multiple routes.

One easy way to think of how to use summarization, if you have a hierarchical topology network, each node in such a tree should only need to pass to its parent node (toward the root) one network address that covers all its subordinate nodes.

lamav Tue, 08/12/2008 - 05:02
User Badges:
  • Blue, 1500 points or more


Although I didnt ask the question, I read and really liked your answer.

I rated it, too. :-)



This Discussion