From: Dino Farinacci [mailto:[EMAIL PROTECTED] >> I fault the IETF (and RRG) for consistently failing to recognize some
>> exceedingly elementary economics: if you want to stay in business, >> then you don't waste money on needless non-business expenses. >> Therefore, if anybody is going to renumber, it will be an organization >> that has a positive business case to do so, such as an ISP. >Eric, let me ask a question to you, as you represent a large enterprise site. >If you knew you could get more bandwidth out of *all* your SP links by >renumbering, would that be considered that a business expense? If the business demanded more bandwidth and if renumbering were a uniform requirement to obtain greater bandwidth from all ISPs regardless of the current contracts or any "special incentive deals", then it is conceivable that it might be considered an acceptable business expense if the financial benefit were greater than the cost to the end user. This is conceivable but unlikely unless the market significantly changes. >And if you knew you could get more bandwidth out of your links by also >being a good net citizen by not injecting non-aggregatable routes into >the core, would they weight into your decision to spend effort >deploying a Loc/ID split solution? Large businesses generally want to be good citizens if the cause-and-effect are visible to them. Businesses often act altruistically if they understand and value the issue. However, they don't want to be stooges who needlessly pay for the (possibly bad) decisions of others, which is what re-addressing has consistently been to date. They also don't want to subsidize somebody else's business (i.e., pay for actions that only benefit ISPs). >> Let's change topics slightly to talk a bit more about elementary >> economics. End users pay ISPs for their Internet connectivity. It >> is therefore the ISP's responsibility to deliver Internet >> connectivity. If >> they can't, then they will go out of business. ISPs are therefore >But it's not binary. I believe the question is how well do they >provide connectivity. Do they provide good service. Is the pain of >staying with an SP worth the gain. Or if the pain is so high, you >are stuck and can't change. Yes, every business entity has business interests -- the ISP just as well as the end user. It's just that the large end user at this current time holds most of the trump cards regarding this particular issue. >> motivated to eventually solve the Internet scaling problem -- >> whether publicly or privately -- because they know that if they >I don't think that's true. Don't forget they are in business just >like you are. Of course they want the Internet scale so they have >a product to sell, but that is more long-term. I think most >businesses think long-term but act in the short-term. >So the fact that SPs accept PI prefixes and/or long PA prefixes is >a fact that they care about getting customers before scalability. If you are correct, Dino, then this is a significant problem for the RRG: If ISPs don't think that it is worth their while to support a RRG approach, then that approach is not viable. However, selecting approaches that favor ISPs but hurt end users are the worst of all solutions because the large end user won't comply so the approach will fail but maybe somebody would be foolish enough to bare the expense begin deploying that doomed approach anyway. Ouch! There is no substitute to creating designs which do not violate elementary economics. >> effective Internet connectivity, then the end users will use some >> other ISP or, worse case, develop alternative marketplaces that >> exclude them. >> Only a suicidal ISP would dare to suggest that a large end user >> renumber as part of an ISP solution because that would negatively >> differentiate >But they require a site to renumber when the site disconnects from them. I disagree: Any large user's PI space will be accepted by whatever ISP they choose to do business with and the ISP will not ask the large user to renumber. If the IAB, IESG or any other non-governmental entity suddenly says that a large user's PI space has now become a PA, then please realize that they don't have any mechanism to enforce their decree so that it will be ignored. Similarly, if the large user has a PA space and they find an ISP willing to treat it like a PI (and very few wouldn't if it is easily self-aggregatable) then it is effectively a PI regardless of what any non-directly-involved entity would prefer. --Eric -- to unsubscribe send a message to [EMAIL PROTECTED] with the word 'unsubscribe' in a single line as the message text body. archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg
