>So are you saying that what Radia wrote is outdated and that MPLS is indeed
>significantly faster than straight IP forwarding? Bill St. Arnaud and
>Howard Berkowitz would emphatically disagree with that, so could you point
>me to some evidence supporting this contention that MPLS is indeed much
>faster?
>
>Not trying to flame, just trying to learn.
You have to assess how important raw forwarding performance is, given
some of the bandwidths in use or nearly in use. If you can use GMPLS
to control 10 or 40 Gbps lambdas, or multiple lambdas on a fiber,
you're not doing per-packet forwarding.
One of the reasons that ATM and its cell tax are less important is
that the queueing you might see at OC-3, which is alleviated by small
cells, is lost in the noise at OC-192.
Packet forwarding rates are one aspect of a network, but not the
be-all end-all. Filtering is going to be done at label edge routers,
and filtering and traffic classification aren't within the scope of
MPLS.
>
>
>""KY"" wrote in message
>[EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
>> ""NRF"" wrote in message
>> [EMAIL PROTECTED]">news:[EMAIL PROTECTED]...
>>
>> > And even the idea of higher throughput has been questioned by the
mother
>> of
>> > all networking, Radia Perlman:
>> > " Originally [MPLS] was designed to make it possible to build fast
>> routers,
>> > but then, using techniques such as [trie searches, parallelism, K-ary
>> > searches] people built routers fast enough on native IP packets. So
now
>> > MPLS is thought to be mostly a technique for classifying the type of
>> packet
>> > for quality of service or for assigning routes for traffic
>engineering..."
>> > (Interconnections, 2nd Ed., p. 347-348). And I think we would all
agree
>> > that anything Ms. Perlman says must be given serious weight.
Very reasonable observations on her part.
> >
>>
>> Her book was published on 01/2000, I would imagine the actual context
must
>> be written 6 month earlier than that date, so her comments on MPLS was
>> almost two years old, we all know in our network world two years means
>what.
>> Just read all those RFCs/Drafts since late 1999.
>> I believe MPLS will play a key role in the optical world, such as DWDM.
>>
> > KY
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=6167&t=6151
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
