[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Some comments on draft-rgaglian-v6ops-v6inixp-00.txt



Greetings,

Hereby my comments and suggestions on draft:

http://www.ietf.org/internet-drafts/draft-rgaglian-v6ops-v6inixp-00.txt

In Section 2: I would strongly suggest to change: "This simplifies separate analysis for IPv4 and IPv6 traffic" this in something like: "This MAY simplify separate analysis for IPv4 and IPv6 traffic".

On our (AMS-IX) equipment it would not simplify that at all.


Section 3. Should want mention that addressing plans must take the universal/local bit in IID [RFC4291] into account. For manually configured addresses the 7th MSB of the IID must be zero. Addressing scheme 1. might violate this.

In this respect I would really suggest that the following scheme is mentioned:

	16 bits ASNs: <prefix>::A5xx:xxxx:yyyy
	32 bits ASNs: <prefix>::A5xx:xxxx:xxxx:yyyy

	The xs are zero padded ASN in decimal notation (BCD encoded).
	The ys are the serial number indicating the interface.

Example of 16 bits ASN:

	IXP LAN prefix: 2001:DB8::/64
	ASN: 64496
	IPv6 Address: 2001:DB8::A506:4496:0001/64

32 bits ASN:

	IXP LAN prefix: 2001:DB8::/64
	ASN: 2703443303
	IPv6 Address: 2001:DB8::A527:0344:3303:0001/64

This scheme is used successfully on several internet exchanges for many years. The "A5" (AS) tag provides a quick hint on the addressing scheme used. It also makes sure that the "u" bit is respected with 32 bits ASNs. Last but not least, the strings sort easily as well, which is not the case for the other examples mentioned. That all makes it IMO a best current practice.

Cheers, Arien