Re: [RRG] Opportunistic Topological Aggregation in the RIB->FIB Calculation?

[Robin Whittle <rw@firstpr.com.au>]

Hi Iljitsch,

You wrote, concerning the limitations of DRAM in a Linux server:

> 100 times the entries * 100 times the churn = 10000 times the
> processing. I'm afraid that your DRAM isn't going to keep up with that
> in a traditional design.
> 
> However, route processing isn't inherently serial. So it should be
> possible to make a route processor that has 100 systems-on-a-chip that
> all process 1% of the routing table. Of course current BGP isn't
> particularly suitable for that, but I don't see a reason why changes
> that allow parallel BGP processing can't be made.

Still, those CPUs need to communicate with each other and maintain a
common RIB, as well as information about each conversation with each
BGP neighbour, via DRAM.

So I am not sure that adding CPUs would be any better than the Linux
server situation of 2 or so fast CPUs with caches with a common
DRAM.  I am not sure how suitable the BGP implementation would be to
100 CPUs each with their own DRAM and some kind of communications
link between them.  That sounds even messier.

  - Robin

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg