[cello-devel] Re: Socket benchmark?

[Kenny Tilton]

Petter Gustad wrote:
> This is good advice. But I don't have an application. I have an
> interconnect that I wanted to know if it could provide a relevant
> performance increase for Lisp socket based applications. I should have
> emphasized this earlier.

This is interesting. A Cello porter suggested extending Cells (my little 
dataflow hack) to work across a network. This would be useful for I do 
not know what. But online gaming springs to mind, where multiple users 
are effectively sharing a single application. the idea would be to make 
the network disappear by building some networking into Cells. This would 
unleash the developer's creativity by letting them forget about 
funneling everything through a socket (or multiple sockets if that is 
what they do, I haven't a clue); two remote users just seem like two 
instances in a single application space, and the developer is free to 
author as fine a granularity of interaction as they like.

The downside would be the overhead of sending extremely small data 
pulses in individual socket messages. probably a buffering scheme would 
be necessary to bundle up a bunch of dataflow pulses into one socket 
message.

Cells already handle inter-instance dependencies which commence by one 
instance finding another by searching an application namespace, so that 
just needs to be extended to know about URLs or somesuch. Add the 
interconnect mechanism and away we go. I guess GUIDs, too.

How available is your interconnect to others?

-- 
Home? http://tilton-technology.com
Cells? http://www.common-lisp.net/project/cells/
Cello? http://www.common-lisp.net/project/cello/
Why Lisp? http://alu.cliki.net/RtL%20Highlight%20Film
Your Project Here! http://alu.cliki.net/Industry%20Application