(sorry, wrong source)
On Wed, Nov 26, 2008 at 4:43 PM, Christopher Morrow <[EMAIL PROTECTED]> wrote: > On Wed, Nov 26, 2008 at 12:09 PM, Templin, Fred L > <[EMAIL PROTECTED]> wrote: >> >> >>>-----Original Message----- >>>From: Christian Vogt [mailto:[EMAIL PROTECTED] >>>Sent: Wednesday, November 26, 2008 12:50 AM >>>To: Routing Research Group Mailing List >>>Cc: Noel Chiappa >>>Subject: [rrg] Maybe it's not "either-or": Considering a >> host/network-basedsolution pair >>> >>>[Forked from: "Fundamental objections to a host-based scalable routing >>>solution"] >>> >>> >>>Christian Vogt wrote: >>> >>>> [...] As I have tried to explain in Minneapolis, a hostname-oriented >>>> stack architecture would mitigate these issues [...] >>> >>>I should have been more elaborative for those who couldn't make it to >>>Minneapolis: >>> >>>Assuming that the goals of RRG are to enable multi-homing and to >>>eliminate renumbering in a scalable manner: The argument I brought >>>forth in Minneapolis was NOT that these goals could be fully solved >> with >>>a host-based solution. The suggestion was instead for RRG to consider >> a >>>pair of host-based plus network-based solution. Since the two goals >> are >>>independent of each other, they may well be best addressed with >> separate >>>solutions: It is obvious that renumbering can be eliminated only with >> a >>>network-based solution. And as previous email discussions indicate, >>>multi-homing may best be enabled with a host-based solution. In fact, >> I >>>don't see a convincing technical reason to address both goals with a >>>single solution. Trying to do that would simply make our job harder. >>> >>>Regardless of which solution pair is picked, the solutions in the pair >>>would have to be independent of each other. A mutual dependency >> between >>>host upgrades and network upgrades would impose deployment hurdles, >>>which in my opinion would be insurmountable. But if each solution in >>>the pair provides benefits independently of the other, and if those >>>benefits are complementary, then the solution pair may deployment-wise >>>well be superior to a single one-size-fits-all solution. >> >> That matches what I was saying yesterday: >> >> http://www.ietf.org/mail-archive/web/rrg/current/msg03834.html >> >> but it seems to me that the RRG's attitude has been one of >> "let's take care of the network, and let the hosts take care >> of themselves". Maybe I'm wrong... > > I saw it as: The pain we see is in the network, host solutions don't > (as of yet) provide the control required for large-scale TE issues, > and often the network/host people are different groups with varied > abilities to interact. > > The only host solution presented from the IAD workshop in AMS (in > 2005?) was shim6, which may work fine for some discrete instances, but > isn't going to help large enterprises multihome, nor is it going to > relieve the pressure on the routing platforms. > > RRG ought to, in my mind, step back and decide what solution set (more > than one solution) is going to best serve the Internet in the > long-term (15-20-30 years), and leverage that over to the IETF to > codify so vendors can implement. I don't have a lot of faith that the > current routing paradigm is going to last 10-20 years at current costs > with current growth curves. > > -Chris > _______________________________________________ rrg mailing list [email protected] https://www.irtf.org/mailman/listinfo/rrg
