IPv6 Packet Header

Humper

Reading about the IPv6 header in Jeff Doyles book, I have a question regarding the DiffServ field and the flow label field.

My understanding of DiffServ (CoS) is that certain applications/services have a higher priority over 'regular' traffic.


Now I have read about flow labels, the router will examine the transport header to determine whether or not the src and dst port should be labelled, I am guessing by a list specified by the administrator. The advantages are that it can be finely grained CoS to ensure the packets take the same path (connection oreinted??) so packets do not arrive out of order.

So my question is, are flow labels just an enhanced version of DiffServ?

Humper

bump!

mikej412

ModemHumper wrote:

So my question is, are flow labels just an enhanced version of DiffServ?

No

RFC 3697 -- IPv6 Flow Label Specification

RFC 1809 is also a good read -- Using the Flow Label Field in IPv6 -- since it gives a behind the scene look at the specification process.

From 3697:
... just the quick logic from the introduction

A flow is a sequence of packets sent from a particular source to a
particular unicast, anycast, or multicast destination that the source
desires to label as a flow.


Traditionally, flow classifiers have been based on the 5-tuple of the
source and destination addresses, ports, and the transport protocol
type. However, some of these fields may be unavailable due to either
fragmentation or encryption, or locating them past a chain of IPv6
option headers may be inefficient.


The usage of the 3-tuple of the Flow Label and the Source and
Destination Address fields enables efficient IPv6 flow
classification, where only IPv6 main header fields in fixed positions
are used.


Plus 1809 said its psuedo-random, and 3697 updates that slightly. There was talk about the Flow Label and Diffserv back in 1998... but I don't see anything that made it into an RFC...... just drafts that didn't make it.

Hum... IPv6 and Diffserv.... maybe RFC 3260 -- New Terminology and Clarifications for Diffserv -- but it updates some other RFC, so there a few more cans of worms that can still be opened.

Humper

Thanks mike! I should of googled myself for the RFC's, I never thought to do it...

It was a really brief description of it in the tcp/ip routing vol1 book, I believe there is more in volume 2, I guess I will find out