Re: Informix vs. Sybase vs. Oracle vs. (gasp) MS SQL Server

In article <862264236.20961_at_dejanews.com>, Brenda Muller <muller.brenda_at_primestar.tci.com> writes:
> > >plans like clustered SMP systems with fiber between the boxes,
> > >or some NUMA-like MPP systems with "memory" connectivity
> > >between boxes (I still don't see how that would work). The
> > Think of the memory bus being extended into an ethernet network
> > between machines. If the memory is 'local' to the processor
> > (think of memory local to machine) then that is a normal
> > memory access. If the memory is 'remote' to the processor
> > (think of memory on a diffent machine) then the memory request
> > is sent across a different bus (think network packet containing
> > the request sent acroos a network with a reply packet sent
> > containing the byte9s) of memory).
>
> Well, that makes sense, but it sounds like a traditional network
> connection between boxes. I saw a presentation from SGI in which
> they were claiming that their boxes actually used memory in place
> of a bus between their servers. THAT's what I don't understand.
>

UMA - uniform memory access, that is, all processors require the same amount of time to access memory from a memory bank. NUMA is non-uniform memory access so a processor requires a disparate amount of time to access memory from another processor. The processors are hooked up via a high speed interconnect. This interconnect is like a mini-network but it's super fast and only for the processors to talk amongst themselves. CC-NUMA is cache-coherency NUMA in that the L2 caches of the respective processors are kept up to date much like the data cache coherency for OPS (SMP does cache coherency as well and the messages on the interconnect (the bus) are what eventually saturate the bus thus limiting scalability).

The above is just computer architecture... take a look at http://csep1.phy.ornl.gov/CSEP/CA/CA.html

<somewhat of a cheap plug here>

The SGI Origin 2000 product line is CC-NUMA architecture. As the box grows (buying more CPU's), you can by *more* bus to increase the bandwidth between processor boards. The architecture is based on a hypercube so for a configuration with 2^N nodes (processors), the degree is N (the number of hops to get the memory in sync).

Take a look at http://www.sgi.com/Products/hardware/servers/techtalk.html

<end of cheap plug>

--
Pablo Sanchez | wk: 415.933.3812| pg: 800.930.5635 -or- pablo_p_at_pager.sgi.com
--------------+-----------------+--------------------------------------------
pablo_at_sgi.com ... when mailing me, place "not spam" in the Subject