Dan Creswell writes about the coming onslaught of multicores and multinodes and the prevailing mindset of byte code virtual machines that still closely resemble the UCSD Pascal system from 1980.
For all the attention that multicores are getting from AMD, Intel, Stanford, Berkeley, and elsewhere it is shockingly apparent that on the software side the more things change the more they stay the same. Rather than the other way around.
2 comments:
Patrick, what do you think about http://www.1060research.com/netkernel/index.html in respect to you consistent multiprocess theme?
They say "Applications scale linearly with cores and CPUs. Each call to a service or resource, no matter how fine-grained, is dispatched to and handled by the microkernel. For each request, NetKernel can optimally schedule the work across the available cores."
(I'm not pushing them, just looking into them.)
Bob wonders about NetKernel?
I don't know much about it, short of what Steve Loughran, Jon Udell, and others have written.
It seems to be an interesting idea, but like Jini/Javaspaces, I believe Netkernel still leaves the developer with a dichotomy between programming within the OS process and programming across OS processes. Unlike Erlang where there is essentially no distinction between an Erlang process in the same or in a different OS process.
Erlang's in-OS-process programming model is simpler and eases scaling out without rewriting code that assumed being in the same OS process.
Post a Comment