[Beowulf] Redmond is at it, again
unix_no_win at yahoo.com
Thu Jun 3 10:13:47 PDT 2004
--- "Robert G. Brown" <rgb at phy.duke.edu> wrote:
> However, this is really a "what the market will
> bear" issue where my
> opinion isn't important.
This pretty much sums up the whole thing....
> As you note in your
> summary and Joe has
> remarked offline, there are a number of cluster
> markets emerging where
> the administrators are not cluster experts (and
> sometimes not even
> linux/unix experts) and who are perfectly happy to
> pay money on a per
> node or workstation LAN client basis as an
> alternative to hiring an
> experienced administrator or cluster programmer
> (which does come at a
> nontrivial baseline cost of say $40-100K).
I would also like to add that not everybody has access
to undergrad (cheap) labor, there are more and more
clusters being sold to commercial entities where you
have 1 admin in charge of 500+ node clusters (with
little or no help), this requires some vendor love to
> are groups that might
> not be able to set up e.g. a repository, PXE,
> kickstart, yum on their
> own, groups that cannot build/package/assemble PVM,
> MPI etc on their
> own, groups that don't know how to install and
> configure some canned
> toplevel cluster package e.g. blast on their own.
> They want a turnkey
> solution that can run "unmanaged" locally or with
> minimal, fairly poorly
> trained, local management. They will pay for it.
Ahh.. if these were the only requirements to run on a
cluster, hardware management becomes pretty intense
on 500+ nodes. figure 8 fans per node, 2 cpus per
node 4 DIMMS per node, 1 disk per node, if you do the
math having a network of office computers this size
would require at least 4 people to manage, however
since this is sold as a "package" solution, it is
figured that 1 person can take care of it! Aside from
hardware and open source software COTS apps usually
REQUIRE a particular distro.
> Determining whether there are ENOUGH of them to
> support a company
> providing the solution at prices they will pay,
> whether they will remain
> customers or learn enough to do it on their own and
> save the money (high
> prices being a strong incentive to do so), whether a
> company can survive
> as "software only" or if the market is best pursued
> in the Penguin/Scyld
> model of sales of hardware, software, and integrated
Note how many petro-chemical companies are entering
the top500 list with cluster computers.. I think they
have fairly robust financial resources.. Also clusters
are being looked at for financial modeling as well, as
they become more and more accepted into commercial
environments you will see more advanced management and
administration tools, however you will also see costs
rising, if you want to use these tools.
> This particular list, of
> course, likely contains
> relatively poor candidates for customers as most of
> the longtime members
> are strictly DIY kinds of people who are used to
> squeezing dollars til
> they squeak and who will only pay for things that
> provide real
> measurable value compared to their doing things
> themselves or with staff
> resources. Universities in particular tend to have
> both the expertise
> and opportunity cost time available to do a lot of
> things themselves.
This is true, so why just not roll your own distro?
All it takes is time and effort. If given resources
of time and more brains this would probably yeild the
best results, but it does take a fair investment in
labor and time.
> However, you also see new list members who in some
> cases would be ideal
> customers as they are clueless and need help.
Clueless seems a bit harsh, perhaps driven by other
mitigating circumstances would be more appropriate.
I have worked in the higher ed environment, and trust
me, it is COMPLETLEY different than the "real"
corporate world. This argument reminds me a lot of the
typical "slashdot" mentality of "since this is the way
I do it, it must be the only right way!" I would say
pick the model that can best be supported by the
resources you have available.
However I will say that there do seem to be a fair
amount of newbie interest in clusters, however I
suspect they are not the ones that these prices are
being aimed at. (Since most seem to be coming from
higher ed addresses)
> Remember also that the market for integrated
> solutions largely exists
> because of sheer laziness on the part of the
> NON-clueless cluster
> persons around the world.
Lets just sum up by saying: If you don't want to pay
for a distro then use a free distro, if you have time
and energy to maintain a free distro, and the
expertise to support all necessary software on it,
then go for it, otherwise we are stuck with what is
mandated either from management, or hardware/software
vendors. And that is what happens as a technology
matures, seen it happen with M$ software, and with
Linux, just a natural evolutionary process,
Universities and researchers do the "grunt" work and
then it is taken by corporations to refine and make
money.. no suprise there.
Do you Yahoo!?
Friends. Fun. Try the all-new Yahoo! Messenger.
More information about the Beowulf