The MPLS-OPS Archive

[Eric Osborne <eosborne@cisco.com>]

On Thu, Mar 20, 2003 at 09:46:59AM +0100, Epi Salamanca wrote:
> Hello Eric,
> 
> so if I understood well in  MPLS network RSVP-TE is only used as a way to
> carry (TE) information about unnumbered links in IGP TE extensions and to
> specify unnumbered links in MPLS TE signalling, isn't it?
> 

no...IGP extentions are used to carry information about links in a
network (the unnumbered stuff is relatively recent, coming as part of
the GMPLS work).  IGP extentions mostly carry information about how
much bandwidth is available on a link, although there's some other
stuff there, too.

> So, is not possible to enforce a specific QoS using RSVP, isn't it?
> To do that we should use Diffserv, isn't it?
> 

Yeah, at least in packet-based (as opposed to MPLS-TE over ATM)
networks.  RSVP signalling is purely control plane, there is no
inherent rate-limiting in accordance with an LSP reservation (at least
not on the few implementations I'm familiar with).  This is deliberate
- not only does it avoid requiring a wholesale hardware upgrade in
order to run mpls-te, it also lets you be very flexible with things
like oversubscription policies, letting DS sort things out when
there's a problem.

> 
> I've the following testbed:
> 
> A==[CE_1] ===[PE_1] ====[P]====[PE_2]===[CE_2]==B
> 
> Where [PE_1]==[P]===[PE_2] are MPLS aware, and belongs to the same ospf 
> area.
> I've setup 2 MPLS TE tunnels :
> 
>     o- [PE_1] ---> [PE_2]
>     o- [PE_2] ---> [PE_1]
> 
> and all the traffic that goes from A to B and to B to A is forwardded 
> through  the tunnels. I've allocated a particular amount
> of bandwidth for the tunnels.
> Everything goes well until I introduce an inteferent traffic (UDP 
> traffic), under this situation (the link is congestioned) the good
> traffic (A<-->B) is affected by the interferent traffic, in fact the 
> interferent traffic is better treated by the routers.
> In all the interfaces I'm using fair-queueing. ( I guess that's the 
> reason why the interferent traffic is better treated - low bit rate - 
> low packet size- several flows)
> 
> Under these situation, if I understood well to Eric, I should setup 
> diffserv to guarantee a specific sla, isn't it?
> 

Yes, you should use diffserv.  




eric

> 
> Thanks in advance,
> 
> Cheers,
> Epifanio Salamanca
> 
> 
> 
> Supporting MPLS TE over unnumbered links (i.e., links that do not
>    have IP addresses) involves two components: (a) the ability to carry
>    (TE) information about unnumbered links in IGP TE extensions (ISIS or
>    OSPF), and (b) the ability to specify unnumbered links in MPLS TE
>    signalling.  The former is covered in [GMPLS-ISIS, GMPLS-OSPF].  The
>    focus of this document is on the latter.
> 
> 
> 
> 
> 
> 
> 
> 
> 
> Eric Osborne wrote:
> 
> >On Thu, Mar 20, 2003 at 10:00:43AM +0800, Tommy wrote:
> >  
> >
> >>Hello,
> >>I have some questions which I want to make sure.
> >>
> >>1.How to find congestion in MPLS based on ISIS-TE network?
> >> 
> >>    
> >>
> >
> >look for interface drops, queue backups, etc.
> >
> >  
> >
> >>2.How to realize Qos in MPLS network?
> >>
> >>    
> >>
> >
> >diffserv, same as in IP.
> >
> >  
> >
> >>3.What's the functino of RSVP in MPLS network?
> >>
> >>    
> >>
> >
> >rfc3209.
> >
> >  
> >
> >>4.Does CSPF calculation uesd the information of TLV and sub-TLV of ISIS-TE protocol?
> >>
> >>    
> >>
> >
> >yes.
> >
> >  
> >
> >>5.How to get the sub-TLV in ISIS-TE and their functions?
> >>
> >>    
> >>
> >
> >draft-ietf-isis-traffic-04.txt
> >
> >  
> >
> >>Any help in this regard will be appreciated!
> >>
> >>    
> >>
> >
> >see also
> >http://www.ciscopress.com/catalog/product.asp?product_id={B9972E73-2D87-47E3-BCDA-6A5334233561}
> >
> >
> >
> >
> >eric
> >
> >  
> >
> >>Thanks and Regards.
> >>
> >> Tommy Chen
> >>
> >>-------
> >>The MPLS-OPS Mailing List
> >>Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
> >>Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
> >>    
> >>
> >
> >-------
> >The MPLS-OPS Mailing List
> >Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
> >Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
> >  
> >
> 
> 
> -- 
> +-----------------------------------------------------+
> | Epifanio Salamanca Cuadrado                         |
> | Tel: + 34 93 401 5623                               |
> | email: epi@nmg.upc.es                               |	
> | PGP Key ID :  6A8A05EB                              |
> +-----------------------------------------------------+
> 

-------
The MPLS-OPS Mailing List
Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
Archive: http://www.mplsrc.com/mpls-ops_archive.shtml