To quote a well known phrase, “you may think that, but I could not possible comment” … but I will. IP fundamentally started out as a UNI and extended its scope to attempt to be a universal connectivity protocol. IPv6 was designed around the perception of how the Internet worked over 20 years ago. IPv6 as a UNI is in my view inadequate in that it does not properly support the degree of sophistication needed to specify current and future SLOs. A trivial example is the way that ECN needed to be hacked into it. Adding support for determinism in will be much harder. The one size fits all addressing and the way that we have to work round that is problemantic. The single address size and domain was a good idea at the time, except that most of the time IP traffic sits in another protocol with an entirely different addressing scope and paradigm. However we seem to find it hard to accept that the fundamental internet protocols should be in a perpetual state of renewal or otherwise risk obsolescence. As professionals we should accept that we need to continually test status quo against the unthinkable, but as an institution we seem to find that hard to accept. Best regards Stewart |
