>You wrote:
>> This sounds similar to the Infospheres project -- perhaps it would
>> be worth a trip down for the talk?
>>
>> - Jim
>
>um, actually, this has little or nothing to do with what we do.
Joe Kiniry writes, referring to Infospheres:
>Though our implementation of a generic
> distributed system uses Java, TCP/IP, and the Web, the ideas
> are directly applicable to any distributed (possibly
> object-oriented) system that supports messages and threads.
My understanding of the SHRIMP work is that it is a) distributed, b) a
system, c) uses messages, and d) supports multiple simultaneous
computations (more full-functioned than threads, but still generating the
problems of asynchronous message transfer). Thus it seems to fall under
your class of potential users (technology which can directly apply
Infospheres), as you define them.
While I am intentionally ignoring some key differences, it still strikes me
that there are also some key similarities to what your projects are both
doing. For example, how might I make a globally distributed operating
system (just a distributed, possibly object-oriented system that supports
messages and threads, right?) using Infospheres? The tradeoff seems to be
the SHRIMP folks can assume a local, high-speed, dedicated message passing
infrastructure, while Infospheres takes IP, TCP, and UDP as its messaging
infrastructure, with no assumption over how close the computational
elements are.
- Jim