I tried sending this mail earlier, but do not see it in the mail archive a day later. So, I am sending it again. Apologies if anyone receives this message twice.
I do not support adoption of this draft. There are a number of technical issues with the proposals in this draft, however the biggest reason why this work should not be adopted by the Int-Area WG (or any WG in the IETF) is that it proposes fundamental changes to forwarding in Ethernet, hence it is likely to impact Ethernet Architecture and - if it were to be undertaken at all - should be undertaken by folks in the organization with the most knowledge about Ethernet Architecture - i.e. - IEEE 802. The fact is that MAC NAT is viewed by many (if not most) of the experts, in the topic area, as a hideously bad idea that they have no wish to work on, is not justification to bring this work to an IETF working group. The issue with impact on Ethernet Architecture is neither presumed, nor theoretical. A number of related issues have already been raised on the mailing list. Among these are issues with compatibility with Ethernet measurement and monitoring, dependence in the forwarding plane of a specific IPv(4/6) address associated with a destination MAC address, and a few subtle variations on these issues. Doubtless many others would be discovered over time if this work were to be pursued. Hence a discussion about ways to fix these problems as they come up is based on an assumption that there MUST be some extreme value to be realized in the long term that justifies creation of an arbitrary number of hacks to "fix" each new issue as it is discovered. As has already been said in recent discussion, there is no obvious value in the proposed approach - particularly in view of existing solutions that do not have compatibility issues - that would begin to justify pursuing this proposal. Given this, is there any reasonable justification for trying to identify and address the technical issues with this proposal, until we have objective evidence that the proposal addresses a real problem for which there is not already an existing solution? -- Eric Gray
_______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
