MESI Caching work start (was Re: caching and userapi)

Sander Vesik sander at haldjas.folklore.ee
Thu Aug 7 10:23:00 PDT 2003


Matthew Dillon <dillon at xxxxxxxxxxxxxxxxxxxx> wrote:
> 
>    Absolutely.  That's the ultimate goal.  I see nothing preventing us
>    from extending it to the buffer cache (actually the VM page cache is
>    what it would be extended to).  We might not be able to extend it all
>    the way to mmap() but we could get close enough that another model
>    could take over for mmap().
> 
>    That's the carrot.  If one can make mmap() fully cache coherent between
>    machines one would be able to run just about any multi-threaded program
>    in a clustered environment without the need for special libraries or 
>    hacks.  Of course, it might not be *efficient*, it would depend on how
>    the program was written, but it would be almost entirely transparent.

aka single system image. Its a good goal.

> 
>    And if we could make the mechanism robust that sort of clustering could
>    be made to work across internet links, in a large scale WAN environment.
>    Supercomputing power cheap!

More interesting (ok, its slightly relative) are various failsafe and failover
modes you would be able to support. Preferably you would be able to shutdown and
upgrade the systems in a cluster one at a time. 

Supercomputing tends to work better when its explicit and not DSM, but even so
having a single system image helps, esp on teh administration side.

> 
>    Not a bad goal, eh?
> 
>                                        -Matt
>                                        Matthew Dillon 
>                                        <dillon at xxxxxxxxxxxxx>

-- 
	Sander

+++ Out of cheese error +++





More information about the Kernel mailing list