The MPLS WG Archive

[Khaled Elsayed <khaled@ieee.org>]

Lee,

In addition to what Vishal listed, you can refer to IEEE JSAC
vol. 14 no 5 (1996) and IEEE JSAC vol. 16 no 7 (1998) for some
good articles.

Khaled

Vishal Sharma wrote:
> 
> Lee,
> 
> A partial response to your question below.
> 
> There are situations where some form of "limited" wavelength conversion
> might be sufficient to give good performance
> and full wavelength conversion may not be needed or be worthwhile.
> (Strictly speaking, there is still wavelength conversion in this scheme,
> just not as much of it as you would need for full wavelength conversion,
> and this could be a significant cost and/or implementation advantage.)
> 
> There are some research papers on this subject. Some relevant references are:
> 
> Yates, J.; Lacey, J.; Everitt, D.; Summerfield, M.
>     Limited-range wavelength translation in all-optical networks.
>     IN:  Proceedings of IEEE INFOCOM '96. Conference on
>     Computer Communications, San Francisco, CA, USA, 24-28 March 1996. p. 954-61 vol.3.
> 
> Sharma, V.; Varvarigos, E.A.
>       Limited wavelength translation in all-optical WDM mesh networks.
>     IN: Proceedings IEEE
>     INFOCOM'98, San Francisco, CA, USA, 29 March-2 April 1998. p. 893-901 vol.2.
> 
> Ramaswami, R.; Sasaki, G.
>       Multiwavelength optical networks with limited wavelength conversion.
>     IEEE/ACM Transactions on Networking, Dec. 1998, vol.6, (no.6):744-54.
> 
> Venugopal, K.R.; Shivakumar, M.; Kumar, P.S.
>      A heuristic for placement of limited range wavelength converters in
>    all-optical networks.
>    IN:  Proceedings of INFOCOM'99: Conference on Computer
>    Communications, New York, NY, USA, 21-25 March 1999,  p. 908-15 vol.2.
> 
> Tripathi, T.; Sivarajan, K.N.
>      Computing approximate blocking probabilities in wavelength routed
>    all-optical networks with limited-range wavelength conversion.
>    IN:  Proceedings of INFOCOM'99: Conference on Computer
>    Communications, New York, NY, USA, 21-25 March 1999. p. 329-36 vol.1.
> 
> 
> Regards,
> 
> -Vishal
> 
> On Tuesday, March 21, 2000 6:21 PM, lee.thomas@wcom.com [SMTP:lee.thomas@wcom.com] wrote:
> >
> >
> > Peter Ashwood-Smith wrote:
> > >   You may waste enormous amounts of bandwidth if you cannot do
> > > wavelength conversion. Convert the question to the ATM domain and
> > > imagine how an ATM network would operate if you had to use the
> > > same VCC from end to end. Basically, if you do not have the ability
> > > to change "labels" then you need one label for EVERY unique path
> > > through the network. Given that the number of paths through a network
> > > grows what .. N factorial? You run out of labels very quickly. This is
> > > especially true with DWDM equipement which may only support 50-100
> > "labels".
> > > This means that a tiny  network would exhaust all the labels/paths
> > > AND you have many links that you can only use a small percentage of
> > > the available "labels" ie bandwidth. So its a very unsatisfactory
> > > solution. Also, to do a good job of it, you need a global view of
> > > each path and label it will use.
> > >
> >
> > I am not questioning the usefulness of wavelength conversion. It
> > is definitely a good mechansim. (Although I have seen some
> > research work that claims that wavelenght conversion may not be
> > worthwhile in some scenarios).
> >
> > Khaled,
> >
> > Could you provide links to some of this research or describe some scenarios
> > where wavelength conversion may not be worthwhile?
> >
> > Thanks,
> >
> > Lee
> >