On Sun, 4 Jun 2000, Horst von Brand wrote:
> David Lang <david.lang@digitalinsight.com> said:
> > as an example of this, I am currently evaluating sparc system that has
> > N+N+N redundancy (with voting), this system will cost ~$80K compared to a
> > single system at <$10K. for some of our firewalls it may well be worth
> > it.
>
> And it has probably non-standard hardware and software to match. Sure, if
> talented hackers get their hands on such machines and the corresponding
> docu, they'll come up with support for it on short notice. But I very much
> doubt all such machines do this in the same (or even similar) way, so you'd
> end up with a heap of different, incompatible stuff. Also, being _very_
> expensive (10 times or so in your case), this won't be remotely mainstream
> enough to warrant inclusion in stock kernels (or userland distributions,
> for that matter). At least not for quite some time to come.
Actually, if this is component-level voting, it could present the OS with
a perfectly normal SPARC box to run on. IIRC, one of NASA's recent
satellite modules does this with memory (radiation in space causes fairly
frequent bit errors, so they have N+N+N voting via an ASIC); the CPU,
however, just sees the memory as being a single, normal block of DRAM.
Even if Linux could be run on it, though, if it's that mission critical,
is a commodity OS really up to the job? I suspect in this case the OS
would become the most significant failure point by far...
James.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:20 EST