Talk:Enhanced Interior Gateway Routing Protocol/Archive 1: Difference between revisions

Revision as of 20:33, 24 September 2010

Latest revision as of 02:56, 10 January 2011

(2 intermediate revisions by one other user not shown)

Line 59:

things - even though it could be very well written, but i wouldn't know it, since I know nothing of its context or discourse - and even call it

"claptrap." Rather, I would just come to the conclusion that the article is more highly specialized than articles, say, on broader subject headings, like computer [[communication networks]], or [[internetworking]],

or [[Routers]], or [[photonics]] for that matter. Incidentally, I am a [[computer]] / [[Information technology|IT]] [[professional]], and I find this article very useful (especially as relates to the discussion on the metrics formula in this Discussion page - which clarifies the formula given on the main page I hope they update the main article with the extra info).[[User:Anon.]]

== Does EIGRP handle IPv6 packet? ==

Line 87:

So to say that EIGRP is a hybrid would be correct. If you have any doubts, complaints, whatnot, contact the author.

----------------------------------------------

I disagree -- jerzy

------------------------------------------------

This is a very week argument: quote from elementary CCNA (Cisco Certified Network Associate) study guide. I fully agree with noel's note in the beginning. Your quote: "And EIGRP has link-state scharacteristics as well - it synchronizes routing tables between neighbors at startup, and then sends specific updates only when the topology changes occur." The last part I would call asynchronous or event driven, nothing to do with maintaining full link-state topology table.

The point of the Bellmandistance-Ford classvector routing algorithm is that router does not need to know full topology but only needs local knowledge of routing summary that it gets from its neighbors. Historically this was significant when routers memory and CPU was limited and savings, real or perceived where important. Dijksra's SPF calculations are CPU intensive and have been "stressful" for large networks for low end-routers that needed to do packet switching in CPU. The link-state vs. distance -vector wars are over, CPUs and memory plentiful but the misconceptions remain. Calling it a hybrid sounds like a compromise remaining from these wars. Chicken may look like a duck from distance but it will not fly or quack :-)