[governance] Feld on network management

Karl Auerbach karl at cavebear.com
Fri Jan 11 20:41:49 EST 2008


Dan Krimm wrote:

> If anyone has to do it (and for Feld that's still a *big* "if"), the proper
> way is to empower users to do it.

Path control from the edges is somewhat akin to using a limp rope to 
push a heavy spherical stone up a steep mountain.  Hard, but, in theory, 
not impossible.  And the idea of clueless user fingers reaching into the 
delicate innards of carrier networks is something that providers, 
rightfully, have reason to fear.

Not trying to proselytize but merely to illustrate one method (of many): 
I have a product, via my company, InterWorking Labs, that is designed 
to, well, to "empower users", at least to a limited extent and with 
reasonably fair minded providers along the data path.

What we do is give the user the means to shape outgoing traffic (and 
very, very indirectly shape incoming traffic) according to one of 
several different typical usage "profiles".  (We call it "Speedbump" 
because the main idea is to slow non-critical things down a bit so that 
congestion of the critical flows can be avoided.)

One of the hard things is that there are piles of queues in typical 
consumer gear.  So one of the things we try to do is to slightly 
underrun the actual outgoing bandwidth so that the queues build up where 
the user can control 'em rather than in some, usually provider provided, 
opaque box.

Way back when, as a result of my work in the late 1990's creating IP/TV, 
I proposed a protocol to help measure and ascertain path conditions so 
that the binding of clients to servers, path selection, and path 
management could be done intelligently.  See 
http://www.cavebear.com/cbblog-archives/000151.html and 
http://www.cavebear.com/archive/fpcp/fpcp-sept-19-2000.html

That, incomplete, protocol, however, wasn't the kind of thing that could 
be done from the edges - it needed hooks in the routers along the path. 
  (I was in the Cisco's CTO's Advanced Internet Architectures group at 
that time, so the notion of getting it into a large number of internet 
routers wasn't as wild as it might otherwise have been.)

		--karl--
____________________________________________________________
You received this message as a subscriber on the list:
     governance at lists.cpsr.org
To be removed from the list, send any message to:
     governance-unsubscribe at lists.cpsr.org

For all list information and functions, see:
     http://lists.cpsr.org/lists/info/governance



More information about the Governance mailing list