ietf
[Top] [All Lists]

Re: QOS [was Re: Why IPv6 is a must?]

2001-10-18 00:30:02
----- Original Message -----
From: "Brian E Carpenter" <brian(_at_)hursley(_dot_)ibm(_dot_)com>
To: <Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu>
Cc: <ietf(_at_)ietf(_dot_)org>
Sent: Thursday, October 18, 2001 1:58 AM
Subject: QOS [was Re: Why IPv6 is a must?]


Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu wrote:

... The QoS field in the header suffers from the same basic
issues as source-routing of packets - they try to modify the global
handling
of packets with insufficient knowledge of global conditions.

Your text mainly refers to IntServ about which I make no comment. But the
diffserv
header field (formerly known as TOS in IPv4, known as Traffic Class in
IPv6)
is explicitly *not* global - it is meaningful per domain, and only makes
sense in a domain that has been appropriately configured. See RFC 2474,
2475
and 3086 for more.


The QoS field in IPv8 (formerly known as TOS in IPv4) is divided into two
4-bit fields. This expands the addressing of the existing IPv4 Internet by a
factor of 16, with no change to the existing infrastructure. Those same 4
bits
then carry over into IPv8 and IPv16 Addressing. The 2,048 address blocks
freely allocated to IPv8 as shown below, are actually each much larger than
the existing IPv4 address space, which needs to be replaced because of the
poor management of the resource and the unfair allocation policies.

Jim Fleming
http://www.in-addr.info
3:219 INFO
http://www.ntia.doc.gov/ntiahome/domainname/130dftmail/unir.txt



<Prev in Thread] Current Thread [Next in Thread>