Hi - > From: "Michael Richardson" <mcr+i...@sandelman.ca> > To: "ietf" <ietf@ietf.org>; "Andrew McGregor" <andrewm...@gmail.com> > Cc: "Christian Huitema" <huit...@microsoft.com>; "SM" <s...@resistor.net> > Sent: Friday, April 26, 2013 5:47 AM > Subject: Re: last call comments for > draft-ietf-6man-stable-privacy-addresses-06 ... > I think that non-contiguous ifindexes are a pain in the ass (based upon > my understanding of enumeration of interfaces in the interface MIB), but > are they essentially forbidden? Having holes would make it easier to > keep things consistent.
This is exactly what RFC 2863 (The Interfaces Group MIB) specifies. Non-contiguous ifIndex values are permitted. This issue was considered at some length in the development of RFC 2863. Section 3.1.5 explains in detail why the semantics of ifIndex were changed (from their original MIB-II definition) to permit "holes", while the semantics of ifNumber were left untouched. Randy