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

Re: [RRG] thoughts on the design space 3: caching



Jari,

It is not at all clear if it is a problem. In fact, I'll argue on some points it could be a feature.

Sure. I don't want to make an argument that the delay/drop characteristics of caching are horrible. In fact, while I know this is an issue, I do not have any information about whether it is of significant practical consequence. This why I'm asking you to find a test for this in EXPLISP.

I'm also not convinced that the mapping delay matters on real networks provided that the granularity of the mapping replies is sufficiently large. In this case, the mapping service is only contacted once for the first packet towards the corresponding EID prefix. In most networks, most of the traffic is sent towards a small number of prefixes, see e.g.

http://inl.info.ucl.ac.be/publications/implications-topological-properties-i
http://inl.info.ucl.ac.be/publications/implications-interdomain-traffic-char

Thus, the ratio ( number of packets waiting for mapping reply) / (total number of packets) will be very low

We did simulations based on Netflow traces on a 1Gbps link connecting 10.000 users and assuming that the mapping replies are for BGP prefixes, the amount of mapping request is reasonable. See e.g.

http://inl.info.ucl.ac.be/publications/cost-caching-locatorid-mappings

The only case where you would detect problems with mapping are when a single host is opening a single TCP connection. In this case, the SYN packet will be delayed by the mapping. Note that using the DNS also causes such a delay...

If EXPLISP needs to perform an evaluation, I would suggest to perform the evaluation based on packet traces or netflow traces as we did and not only on lab experiments with limited traffic.


Olivier

--
http://inl.info.ucl.ac.be , UCLouvain, Belgium

--
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