Re: [RRG] Properties of mapping solutions

[jnc@mercury.lcs.mit.edu (Noel Chiappa)]

    > From: Eliot Lear <lear@cisco.com>

    > NERD shows that actually pushing the data is not cost prohibitive,
    > using today's numbers

You mean, the computational/communication bandwidth overhead is not cost
prohibitive, I assume? (Actually, now that I think about it, the memory cost
is probably not unreasonable either, particularly if the mapping table can be
kept in DRAM, and recently-used entries cached in faster RAM.)

    > What goes into primary cache memory is a separate question.

Not sure what you mean by "primary cache memory" - is this the same point as
my "cached in faster RAM" above? But you are assuming all ITR's do keep
copies of the entire mapping table, right?


Because there is another axis to 'push', which is 'how much gets pushed';
people often take 'push' to mean 'push the entire table' but of course it's
easy to visualize schemes which e.g. identify highly-used bindings and push
them to everyone, and only use pull on the infrequently used ones.

More complex, yes, but it uses less resources. As I have said elsewhere, it's
like the old "fast, cheap, good" axiom, only this version is "speed,
overhead, complexity: pick any two".

	Noel

--
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