Re: The next generation of computing...

Gregg Wonderly wrote:
> Here's an interesting article on some computing research.  These
> performance
> increases make more people content with huge app server like systems it
> seems.
>
> <http://www.asd-network.com/press_detail_B.asp?ID=11674>

I doubt that Gregg. These CPUs/Systems often only get their
throughput by high parallelism. A huge application server, or an
application running in that server, not designed to take advantage of
that is ... 'spinning around the lock'.

If you look to the Jini codebase you see a lot of asynchronous execution
and given the fact we know remote method invocations are a few orders
slower than local invocations often you resort to asynchronous
communication patterns in Jini software systems (at least that is my
experience). So it seems we Jini people have learned it the hard way and
all this knowledge can now be applied now that we really have all these
lanes to do the execution on [1].

Distributed computing to me is not just about communication between
processes on different computers connected through a network cable it is
also about communication between processes on one computer not being
able to share the same address space.

[1] although I don't have the exact details only the rumors, it recently
came to me that a few servers running Seven were suffering from a large
load factor. These servers were replaced by 2 T2000 and while the number
of physical CPUs dropped as well as the clock-speed, the load factor
also dropped considerably and that is exactly what I expected and hoped
for given the way these software systems were built.
--
Mark

--------------------------------------------------------------------------
Getting Started:     http://www.jini.org/wiki/Category:Getting_Started
Community Web Site:  http://jini.org
jini-users Archive:  http://archives.java.sun.com/archives/jini-users.html
Unsubscribing:       email "signoff JINI-USERS"  to [email protected]