Re: *Real* Distributed Computing


<examachine@xxxxxxxxx> wrote in message
news:1123671895.331449.245810@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> Michael N. Christoff wrote:
>
> I think there are still lots of unsolved problems. The GRID community
> seems to be trying straightforward applications of distributed
> computing, but
> I think you are really interested in parallel distributed computing,
> which is something above distributing tasks of an embarrasingly
> parallel
> program, like SETI.
>
> The true challenge in distributed computing arises when the data sizes
> exceed terabytes, and the kind of processing required is on the order
> of petaflops, and the data partitioning is not trivial I guess :)
>
> One would like to see more unifying approaches to distributed computing
> rather than the lots of small hacks we use today. When I open up a
> parallel computing textbook, it feels like a bag of tricks.
>
> Cheers,
>

Hi Eray.

Please check out comp.distributed if you're interested in continuing this
discussion. I don't think comp.theory is a good newsgroup for these topics
now that I've posted here. I've responded to some other people in
comp.distributed, and my responses seem to address many of the points you've
brought up.

Thanks,

-mike




.