On Wed, 2003-06-18 at 14:24, Levi Bard wrote:
> Another point states that KVM traffic will be less than that of
> transferring data back and forth between workstations and servers - is
> there any documentation or statistics supporting that? (My apologies if
> they're in the spreadsheet - I didn't read it (No OOo at my workplace).)
> It has been my impression that the KVM traffic would probably tend to
> exceed routine client/server data transfer, except in outside cases.
Worth double checking.
> An additional disadvantage of the LTSP solution: since it harkens back to
> the mainframe-style network, it brings back the problem of single/few
> point(s) of failure. If for some reason the servers go down, every thin
> client becomes helpless.
This is a disadvantage I need to identify (and clarify as below) though
for offices like mine I don't see it as a deal breaker, or even of great
concern.
* If the servers (PC/Server or Thin Client/Server) are lost due
to power failure then so is everything else, except laptops
* If the server (PC/Server or Thin Client/Server) actually fails
then in my office PCs and Thin Clients alike are functionally
dead because it's a pretty hard and fast rule that documents are
to be kept on the servers.
* If redundant servers haven't yet been explored by the LTSP
folks I'd imagine it's not far off. Meantime, if I had a
critical LTSP system I might use some of the savings to purchase
a spare server and get it completely configured and tested so
that a relatively quick tape restore job could get everyone back
up.
* Hardware is pretty good these days so for my office we
concentrate on power protection, good backups and (implicitly)
the ability to get a replacement server up. The strategy for
this sort of concern would vary widely but the likelihood of an
out and out server failure (assuming reasonably maintained
servers that aren't allowed to get too old) seems pretty remote
- at least based on my experience.
All that said I think it is important to include the issue in the paper
- thanks for pointing it out.
> On the other hand, you could list GNU/Linux's excellent stability and high
> MTBF as an advantage, offsetting the previously mentioned disadvantage.
> Some stats would be nice here as well.
I wouldn't want to gloat. On second thought, maybe I would.
>
> HTH,
> Levi
Thanks
John
This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 17:02:50 EDT