[Info-vax] Message from HP.

[johnwallace4 at yahoo.co.uk]

On Monday, 9 December 2013 17:20:47 UTC, John Reagan  wrote:
> On Monday, December 9, 2013 11:59:00 AM UTC-5, Gérard Calliet (pia-sofer) wrote:
> 
> > Le 09/12/2013 17:19, John Reagan a �crit :
> 
> > 
> 
> > > Well, the ServerNet switch is proprietary [think: Star Coupler].
> 
> > 
> 
> > It is just what I said : there will not be ServerNet on NSK port on x86
> 
> > 
> 
> > 
> 
> ServerNet has nothing to do with the "nonstop" aspect of NSK no more that CI adapters have anything to do with OpenVMS Clustering other than providing a high speed data path with some redundancy.
> 
> 
> 
> NSK on x86 will use a different (and off-the-shelf) interconnect with better performance.

Try this for size, in the absence of the Hal Massey presentation:

http://ftp.hp.com/pub/nonstop/ccc/nov0608.pdf

"Before this move to blades, Integrity NonStop
servers were based on what HP calls the NonStop
Advanced Architecture (NSAA). In a DMR (dual
modular redundant) NSAA configuration, two
processors on two different physical modules carry
out each computation.  The results are then
compared; if the results aren’t identical, a
comparison error between the two boards is noted,
and the system will try to determine the source of
the error. If it can do so, it will disable the defective
hardware and continue processing. Otherwise, it
will disable both halves of the logical processor and
transfer the job transparently to another logical
processor using a NonStop software FT technique
called “process-pair takeover.”  Logical
Synchronization Units (LSU) handle the
comparisons by looking at all the I/O traffic across
the ServerNet interconnect. The LSUs compare
output packets to ensure that an error or corruption
hasn’t occurred, and they convert I/O packets to
and from ServerNet’s format.

NonStop has used conceptually similar techniques
since abandoning its own processors in favor of
those from MIPS in 1991."

(from an Illuminata paper in 2008).