>I thought somebody was going to talk about masking instabilities. But then >that begs the question - in a typical enterprise network (therefore a small >one of 100 routes or less), if you are suffering from routing instabilities, >isn't your time better spent to try to figure out why your routes are so >unstable and then remedying it rather than engaging in summarization in >order to mask the instability. But the summarization localizes the instability and speeds the time to find the links that are flapping. If I cut myself, and the wound stops bleeding, I am still likely to put a dressing on it to keep it clean. A large percent of the time, if the wound was clean to begin with, it won't get infected whether I put the dressing on or not. There are certainly differences in local practice -- American medical personnel wipe an injection site with alcohol first, while Europeans generally don't -- and, for routine injections (i.e., not, say arterial or IV), there's no noticeable difference in infection rate. Summarization is what the IETF calls "best current practice," because experience shows that it tends to avoid problems. One of the most important reasons is that it forces, if you will, infection controls onto your addressing plan. Hierarchical networks are much easier to expand if your network suddenly grows, which it may in a world of mergers and acquisitions. I always wear my seat belt. Oh...and as long as we're mentioning grails, I'm off to the IETF in London. Perhaps I should request that I be addressed as Arthur, King of the Britons, and that I am on a quest. > >Like I said previously, I completely agree that summarization is indeed very >useful in large networks like NSP/ISP's or large enterprises (1000+ routes), >for many reasons (better lookup performance, masking truly becomes useful >because you can't be expected to fix all your flaky links in a huge network, >etc.). But I would like to understand if summarization can be useful in a >typical enterprise network ( wrote in message >[EMAIL PROTECTED]">news:[EMAIL PROTECTED]... >> Performance gains are only a small part of the >> picture... what is more important is enforcing a >> proper hierarchical addressing scheme that conceals >> routing instabilities from the network as a whole, and >> lessens the amount of routing update traffic >> propagated across the entire network. >> >> It's gotten to the point >> > that Cisco-trained >> > personnel treat summarization like the holy grail, >> > and they go around trying >> > to use summarization techniques wherever they can. >> >> A network always benefits from the consistent >> application of design goals. Summarization scales >> well because of the architecture which flows from a >> properly addressed network. I can't think of anyone >> outside of an SP network concerned with global routing >> table bloat that ever equates the benefits of >> summarization in terms of increased routing table >> lookup efficiency. The benefit is that flapping >> routes and their attendant update traffic are confined >> to a small manageable area. Not only does this >> preserve bw but it greatly aids in network management >> by narrowing the scope of the network that you need to >> troubleshoot. >> >> So, when I weigh >> > the cons of suboptimal routing as well as the >> > possibility of >> > misconfiguration, I find it difficult to see why the >> > typical enterprise >> > would ever really want to do summarization, as the >> > gains are miniscule at >> > best. >> >> If the network architects can't properly summarize, >> there are bound to be bigger problems than what that >> particular misconfiguration will bring. We are not >> talking rocket science here, it is simple binary math. >> >> Best regards, >> >> Geoff Zinderdine >> CCNP MCP2K CCA > > MTS Communications Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=14629&t=14629 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
