[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [RRG] Tunnel fragmentation/reassembly for RRG map-and-encaps architectures
On 8 jan 2008, at 15:35, Eliot Lear wrote:
The control plane memory could be a bottleneck and lack capacity
for 10^9 entries.
Since we have no upper limit on the number of EID prefixes = RIB
entries, it's impossible to say what will not be a bottleneck.
Why do you say that? We have a population limit, if nothing else.
Perhaps things won't scale if birds start having IP addresses...
Last time I check it's still possible to have a child without a
license... Obviously the number of EID prefixes won't be infinite, but
there is no figure you can come up with that satisfies both the
condition that it's not reasonably possible to have more prefixes than
that on the one hand and hardware to handle that number of prefixes
can be made within the forseeable future at reasonable prices.
The problem that we're faced with is rate*size. If we split the
traffic over multiple boxes we limit rate for any particular box, if
we split the routing information over muliple boxes we limit the size.
This way, we should be able to create clusters of smaller boxes than
can provide the necessary performance much more easily than in the
situation where out protocols force every update/packet to go through
a single bottleneck.
to unsubscribe send a message to email@example.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