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

Re: research questions



> At 4:16 PM -0400 10/11/00, Dhiman Barman wrote:
> >Hi,
> >   May I know what are some of the open problems with Linux-diffserv
> >which one can look at as potential research areas.
> >
> >db
> 
> There are a number of bit-fiddling questions which get debated on 
> this newsgroup.  But one of the real-world problems is how to control 
> QoS.  If everybody has the 'priority button' what's to keep them from 
> all pushing it? Leaving us back at the best-effort position we're at 
> now.

IMHO, the SLAs that are established between the edge and the core (or
enterprise and ISP) will be able to resolve this issue by MF
classification, policing and remarking. Therefore even if everybody pushes
the 'priority button' they will be given only the capacity that they have
paid for and negotiated in their SLAs. Such a situation has been accounted
for in the RFCs.

  There's language on bandwidth brokers in the IETF working group 
> papers, but not much in way of implementation.  As a suggestion....

Bandwidth Broker specs and implementations were assigned to the
Internet2's Qbone community and is available from their website. There is
a freebsd implementation by Andreas as well.

There has been attempt for Intserv over diffserv using COPS-ODRA by
Roberto Mameli and Stefano Salsano from CoRiTeL, my understanding from
Roberto's presentation at Adelaide IETF meeting was that they had used
linux, I don't know the status (if Roberto is still on the list, he can
give us an update). The IDs were

	   COPS Usage for Outsourcing Diffserv Resource Allocation 
           draft-salsano-issll-cops-odra-00.txt 

           Integrated services over DiffServ network using COPS-ODRA 
           draft-mameli-issll-is-ds-cops-00.txt 

I have been concentrating on extending the Linux QoS architecture for an
Active and Programmable network. A few vendors are headed in this
direction and standards are in there incubation stage, a linux
implementation attempt would keep us ahead. I will send out an
announcement once I have something decent going.

All the best.

Cheers
Muneyb.
_________________________________________________________
Muneyb Minhazuddin - Telecommunications Research Engineer    
CSIRO Telecommunications and Industrial Physics                
Sydney, NSW, Australia.                                    
                                                             
Phone no. : 61 2 9372 4113
FAX       : 61 2 9372 4490          
e-mail    : mminhazu@tip.csiro.au  
Home Page : http://www-networks.tip.csiro.au/~mminhazu
---------------------------------------------------------