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

Re: [RRG] Renumbering...



Tony,
Thanks, but I'm not going to be accepting that anytime soon.  When true PI
is a requirement, then the 'net cannot possible scale.  Period.  Full stop.

From my perspective PI is absolutely a requirement. And I would also agree with Noel that the problem isn't the Boeings of the world, but rather it's the consumers who decide that they too would like to have redundant PI connectivity. We have an existence proof that we can at least change the equation, although we disagree on how the equation is changed. Which brings us to this:

Map and encap does not 'solve' the problem, it only pushes it from BGP into
the mapping function.  Of course, if BGP is also used as the mapping
function, it's largely the same problem.

By adding an additional layer of indirection LISP hasn't just moved the problem, but changed its nature. The different mapping systems change that nature slightly differently, and so the costs for each are slightly different. It's probably worth more time to characterize the change and then evaluate it against the renumbering challenge. For example, having a 10^9 entries in a table doesn't cost you all THAT much if you are at the edge and only use a small portion of them in a fairly predictable way. OR... don't have the table, establish a strict directory hierarchy and make a query into that periodically. In either case, it's different sorts of state that should be evaluated differently. It's not just more prefixes.

Eliot

Eliot

--
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