Steinar, -----Original Message----- From: sth...@nethelp.no [mailto:sth...@nethelp.no] Sent: Tuesday, August 24, 2010 1:13 PM To: adur...@juniper.net Cc: Hemant Singh (shemant); ipv6@ietf.org; dtha...@wollive.windowsmedia.com.akadns.net Subject: Re: 6man discussion on /127 document @ IETF78
>If the complexity can be *hidden* such that the operator configures >the /127 link "as usual", this might be okay. If the operator needs >to perform extra configuration to get the /127 link working, probably >not. Great thought. Indeed the off-link model functionality I and Dave Thaler speak of for getting around the anycast issue with use of a /127 is "hidden". The functionality is an internal implementation of the router. The external user only sees the fact that when the user configured a /127 on the router interface, the router does not jump to invoking anycast data forwarding mode. The SP's from Japan I spoke with at the IETF 78 told me they want to future-proof their network. They are working fine right now with /127 on their Cisco and Juniper routers. But they say, what if router vendors in future change their code to start anycast processing with a /127 configured. Also they were burnt recently when Linux changed its /127 implementation to kick of anycast processing - apparently some small router company was using this Linux code. Hemant -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
