suggestion for DSCP codepoint for "less than best effort" (scavanger class) traffic

[Mikael Abrahamsson <swmike () swm pp se>]

Hi,

in the IETF transport area there are discussions on DSCP interconnection 
between networks. It's quite common that operators clear (bleach) the DSCP 
codepoints when receiving packets from other operators, turning everything 
into BE (best effort) traffic. Historically CS1 has been proposed as 
less-than-best-effort (for instance bittorrent traffic), but my opinion is 
that this is ill suited as I think it's not incrementally deployable. Some 
networks might use CS1 as something that has higher priority and it would 
be hard for them to change this incrementally.

I have proposed a DSCP codepoint that I believe is incrementally 
deployable and that is 000010, which I think is deployable because it maps 
to CS0, PRECEDENCE 0, so any kind of equipment that today takes these bits 
and imposes it into 802.1p, MPLS TC (former EXP) etc, will just map this 
to regular BE. If explicitly configured, one can match on 000010 and put 
this in a different queue, for instance that doesn't have much bandwidth 
guarantees at all.

Since this proposal just comes from my personal experience with equipment 
in the MPLS/L2/IP world, I'd like to hear wider view/opinion on this and 
I'll bring it to the TSVWG at the upcoming IETF in Prague July 20-25.

It would be great if there could be an operational document as well, 
giving recommendations on how to configure the above, and it would be 
great if it could be done with lots of operator input into the matter.


---------- Forwarded message ----------
Date: Thu, 2 Jul 2015 00:32:40 +0000
From: "Black, David" <david.black () emc com>
To: "Ruediger.Geib () telekom de" <Ruediger.Geib () telekom de>,
    "swmike () swm pp se" <swmike () swm pp se>
Cc: "tsvwg () ietf org" <tsvwg () ietf org>
Subject: RE: [tsvwg] draft diffserv-intercon: Handling of a scavenger class /
    CS1

<WG chair hat OFF>

> - work on a scavenger class standard needs to be done in a
>   separate draft.

And here's what that draft might encompass ...

  Beyond that, I'd support a (hopefully short) draft that
        - updates RFC 3662 to change the recommended DSCP for the LE PHB
                from CS1 to 000010,
        - recognizes and allows for continued use of CS1 where deployed, and
        - updates other RFCs (e.g., 4592, 5127) accordingly as needed.

As Brian pointed out, this would need to be a standards track draft with
a downref to RFC 3662 in order to allocate that DSCP.

Writing the draft is straightforward by comparison to figuring out whether
this is what we should do, as indicated by the other messages in this
thread on operator/operational considerations.

This item is on the draft tsvwg agenda for Prague as a discussion topic - I
plan to bring slides and only write the -00 draft after I've survived that
discussion ;-).

</WG chair hat OFF>

Thanks,
--David


> -----Original Message-----
> From: tsvwg [mailto:tsvwg-bounces () ietf org] On Behalf Of
> Ruediger.Geib () telekom de
> Sent: Monday, June 29, 2015 4:51 AM
> To: swmike () swm pp se
> Cc: tsvwg () ietf org
> Subject: Re: [tsvwg] draft diffserv-intercon: Handling of a scavenger class /
> CS1
>
> Hi Mikael,
>
> ok, understood, I'm relaxed.
>
> The relation of Diffserv intercon to scavenger is
>
> - if there is a scavenger class, how should Diffserv intercon handle it
>   (will add text about a 000 xx0 DSCP)? Otherwise Diffserv Intercon
>   says, bleaching may occur for any unexpected DSCP.
>
> - work on a scavenger class standard needs to be done in a
>   separate draft. That's WG consensus and Gorry Fairhurst said during
>   the last meeting:
>   "Scavenger support document been discussed before, if there is
>   renewed interest, then we should take it up again.
>   As an individual contributor, I would be interested.
>   If you are also interested in this topic, please talk to the
>   WG chairs."
>
> - operational issues may be relevant for a bleaching discussion.
>   A bleaching standard is out of scope of Diffserv Intercon (thats
>   WG consensus). I think to recall interest in work on bleaching
>   by others too.
>
> Then I'm happy to meet you in Prague.
>
> Regards,
>
> Ruediger
>
> -----Ursprüngliche Nachricht-----
> Von: Mikael Abrahamsson [mailto:swmike () swm pp se]
> Gesendet: Montag, 29. Juni 2015 09:38
> An: Geib, Rüdiger
> Cc: tsvwg () ietf org
> Betreff: Re: AW: AW: AW: [tsvwg] draft diffserv-intercon: Handling of a
> scavenger class / CS1
>
> On Mon, 29 Jun 2015, Ruediger.Geib () telekom de wrote:
>
> > I repeat that I've talked with my RIPE/NANOG colleagues to get
> > feedback
> > - they came back only with one or two names with whom to discuss QoS
> > interconnection. And I never got any useful response (positive or
> > negative) from those.
>
> This is the problem, that's not how to do this, the way to do this is to
> announce an idea on the nanog mailing list and see who might be interested.
>
> As far as I read your document, it has nothing in it about how to
> incrementally get an Internet-wide scavenger class operationally deployed over
> time. To me it reads more like a document for business VPN Carriers-Carrier
> services style interconnect/interworking. It has no operational advice at all
> in it on how to deal with DDoS, how to configure access equipment, how to set
> up queuing strategies on different kinds of equipment, listing what equipment
> might do what, etc.
>
> I suggested how to actually make this incrementally and widely deployable
> across the world, and that is to aim for a 000xxx style scavenger class.
> This seems to not work for some reason that is unclear to me, your document
> doesn't seem to be operational at all, it doesn't propose something that is
> operationally deployable on the wider Internet, so that's why I think we would
> need another document for this purpose. If that is not a goal we can agree on,
> then there is no need to write one.
>
> So just to sum up: I'm sure your document is fine for what it intends to do,
> it just doesn't answer any of the concerns an IP network engineer/architect
> might have about implementing this on an Internet peering link where there is
> little or no control over what traffic will pass or what this traffic might do
> to the rest of their network. It suggests a way to do something with no
> operational analysis or risk mitigation at all.
>
> The first thing that will pop into any good IP architects head nowadays will
> be "oh, what happens to my network when I get 200 gigabits/s of 64 byte
> packets marked EF from that 5 million strong DDoS-drone network all across the
> world, aimed for my customer base?"
>
> The answer to that is "I'll bleach DSCP it because that's the only way to
> handle it". So in order to get a scavenger class actually deployable, you have
> to come up with something that means they can relax the bleaching a little bit
> by some operators, but their infrastructure would still treat the traffic the
> same. You also have to take into account that a lot of core IP network
> equipment can't do DSCP mapping very well. This kind of functionality makes
> equipment more expensive and thus less likely to exist widely.
>
> --
> Mikael Abrahamsson email: swmike () swm pp se