I think there's a legitimate case to be made for best effort services to be
representable through the SLS. The main scenario that I can think of relates
to closed user groups. For instance, if as a campus network administrator I
am being provided access to two other sites over a provider's shared
infrastructure (with traffic to and from all other IP addresses being
discarded), and I want to add a third site to the list of sites accessible
from mine, then I would have to express my preferences through semantics
very similar to those in the SLS draft.
Whether this use case -- closed user groups -- is outside the framework of
*this* effort of SLS for QoS, is something I would like to hear more about.
raju rajan
----- Original Message -----
From: Panos Trimintzios <p.trimintzios@eim.surrey.ac.uk>
To: <sls@ist-tequila.org>
Sent: Thursday, November 02, 2000 5:56 AM
Subject: Re: [tequila/sls] predefined services / SLS ?
Dear Jerry hi,
On Wed, 1 Nov 2000, Jaroslaw Sydir wrote:
-Michael,
-
-I'm not sure that I understand your question. If the traffic is truely
-best effort then it requires no guarantees and you don't care about it
-in your traffic engineering. It gets the lowest priority in the network
-and is dropped in favor of the guaranteed traffic.
Even if the traffic is "truely best effort" as you mention above you DO
care about it in the traffic engineering of your network, since you do not
want this type of service to be starved. This means that wnen you engineer
your AS you should always try to keep some of its resources "free" to be
used by BE traffic.
-If you want to provide a "slightly better than best effort" service,
What do you mean "slightly better than best effort"? Couldn't this type of
service offering handled by a qualitative (or another)SLS? So if you name
this as the "lowest of Olympic Services" or "slightly better than best
effort" does it make much difference? I don't know in general but when it
comes to engineering again it doesn't.
cheers,
PanOS
-then this implies some level of performance guarantees for some amount
-of this traffic. In this case you would describe it in terms of
-individual e2e SLS just like the other types of services.
-
-Jerry
-
-
-Michael Smirnov wrote:
->
-> On Wed Nov 1 00:51:41 2000 Jaroslaw Sydir wrote:
->
-> > Jean-Pierre and Pierrick,
-> >
-> > A traffic matrix (network wide SLS) can be specified as a set of
individual e2e
-> > SLS/services (represending the trunks that make up the traffic matrix).
-> <...>
->
-> How you plan to specify best effort traffic?
->
-> cheers
->
-> Michael
-
=======================================================
Panos Trimintzios
Research Fellow, Networks Research Group
Centre for Communication Systems Research (CCSR)
Univ. of Surrey, Guildford, Surrey GU2 7XH, U.K.
Office: U37 / BA Building
Tel: +44 (0)1483 876005 Fax: +44 (0)1483 876011
Email: <p.trimintzios@eim.surrey.ac.uk>
WWW: <http://www.ee.surrey.ac.uk/CCSR/Networks>
=======================================================
This archive was generated by hypermail 2b29 : Thu Nov 02 2000 - 19:04:44 CET