Hi Margaret,
> Your document and your arguments have convinced me that we > should publish a standard definition of the minimal requirements > for an IPv6 node, an "IPv6 Node Requirements" document (or perhaps > two documents, one for hosts and one for routers?). This should > be a standards-track document, not an informational one, and I > think that your document would serve as an excellent starting > point for this work. Point of clarification, I personally prefer splitting the host requirements from the router requirements. A host is a node which does not forward packets on behalf of another node, while a router does. In creating the current document, this seperation really helped in scoping it (perhaps it didn't scope it enough, I'm sure a few of you are saying ...) > It is important that the minimal host requirements of IPv6 be > applicable to low-end systems, such as cell phones, and that > should be reflected in our general IPv6 node requirements effort. Without a doubt. > However, I don't think that we want to have a fragmented set > of informational host requirements documents with different > requirements for different IPv6 application spaces (cellular hosts > vs network appliances vs. home gateways vs. car infotainment > equipment, etc.). If I'm missing some reason why cellular > hosts are special, that explains why they would need an > informational requirements document (when other applications > would not), please explain. We don't want a fragmented market, that is why we've thought that the work should be done here. I think the reason why we have been saying that they are special is because there is urgent need for this kind of document and there is nothing else out there. I don't think that anyone is going to argue on these 2 facts: 1) if we put out alot of IPv6 enabled cellular hosts, those hosts better behave well. 2) if we put out a lot of IPv6 enabled cellular hosts, IPv6 should work well over the cellular link. I guess we are just differing on how best to achieve ensuring the above. Oh, one additionally point, perhaps it is not as important since it is an implementation issue, but the initial implementations of IPv6 most likely will be in firmware, making tweaking extremely difficult, which puts more pressure on making sure we get it right. Does this make any sense? John -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
