[Beowulf] Redmond is at it, again
landman at scalableinformatics.com
Tue May 25 16:57:04 PDT 2004
On Tue, 2004-05-25 at 17:20, Robert G. Brown wrote:
> > Fascinating... A C# binding for MPI?
> There is a common and obvious thread to M$ choices under nearly all
> circumstances. It can easily be used to interpret the specific
> decisions to make MPI "their own" and eschew TCP/IP, vendor network
> drivers, ANSI standard C or C++ or f-whatever, and especially to
> interface it with their own version of core network services:
> Standards are bad, unless we own them and can alter them on a whim to
> effectively clone, co-opt, and eliminate all competition from any
> products ever developed that are based on those "standards" and
> actually make money.
<disclaimer>Not an apologist for anyone/any corp</disclaimer>
Somehow I have the sense that many folks simply attribute nefarious
intent to various corporate entities, though there are simpler
explanations. Profit motivation drives most companies, apart from those
whose business model involves giving away their value to their consumers
in an uncompensated manner.
Microsoft, like all corporations whose goals are to make money, needs to
review from time to time, its mix of product, its market, etc. Failure
to do so can be (and has been for some companies) fatal.
Heinlein said something apt about this which could be paraphrased, Never
attribute to malicious intent, that which can be explained by
cluelessness. Corporations are not (for the most part) malevolent
entities out to destroy all. They (for the most part) have easy to
discern goals and objectives. In fact they are eminently predictable
(for the most part).
> The key words here being "make money" of course. Microsoft won't ever
> jump on a cluster computing bandwagon unless it is running on wheels
> they own and all of the seats have a highly proprietary "eject" button
> somewhere they can use to get rid of all the rest of the instruments.
Is this in any way different from GE's strategy of being number 1 or
number 2 in every market they are in? All corporations (the ones that
want to make money) seek to maximize profitable revenue. There is
nothing wrong with this as a goal. For the larger folks (not the
smaller fish like us), this means controlling the market.
While controlling the market is great for the shareholders, it isn't so
great for the consumers. Competition in a market is always a good
> A corollary of the above is that "Java is Evil" (an opinion shared by at
> least some people who don't work for M$:-). Of course, so is nearly
heh... more than a few. "If all you have is a hammer, then every
problem looks like a nail."
> everything else used in modern networks for the transport layer and most
> of the layers above transport in the good old ISO/OSI scheme, so is
> http, html, xml, php, perl, python and so forth, and Open Source, Open
> Standard software is The Devil Himself -- to M$.
Oddly enough, it should not be. Open standards relinquish some level of
control. But they provide a common platform for a much larger group of
developers. No one can make a decision, change an ABI, etc. which will
prevent me or my company from releasing a Perl, Python, ... based
application. The same is not true for closed platform languages, of
which the previously mentioned is one, though oddly enough, C# is not.
I can run my C# code on mono on most systems, and on .Net. Go figure.
With this in mind it is very much to Microsoft's advantage to support
Mono. It gets rather hard to accuse someone of embrace and extend with
the tool when the platform runs everywhere and is open (and
> Current systems corporations are thinking "license/lease" and "renewable
> revenue stream" instead of "software sales" as the latter era of the
> computer revolution draws to a slow end. Red Hat has gone there. Sun
They have to. Shareholders (such as the folks in the US with 401k,
403(*), and other retirement bits from TIAA/CREF and elsewhere) have
this preference to get a return on their investments. These
investments, in mutual funds, or stocks, etc. only really come about
(ignoring bubbles) when earnings rise. Earnings rise of course when a)
you sell more, b) you sell more profitably, c) your competitors leave
the market d) your revenue streams become more predictable.
We are the ones who compensate these folks by their success at
maximizing the profitability function of the corporate entity. So we
shouldn't be surprised when they come up with "new" models to "monetize"
Sure, as a model, subscriptions are boring. You sign up, you pay, you
get something, you keep paying, you keep getting things. Over the
lifetime of the thing you use/buy, you keep paying, often as much if not
more than before, and in a regular manner. This makes revenue streams
more predictable, smooths out business, and generally makes Wall street
happier (smooth == nice), which has positive effects upon the stock
price. See the feedback loop there?
> is well along the road. Microsoft is stuck between a fervent wish to go
> there and somehow preserve their high margins and the realities of the
> expectations of their PC customers that they are "buying a copy" of the
> operating system, not the right to use it for a year with automated
> prepaid updates. In the corporate world, however, they've moved a long
> ways there and I feel confident that they'll go the rest of the way
> So the interesting question is "where's the money" in this move. The
> HPC market hasn't proven to be exactly a get-rich-quick proposition for
> any of the various groups that have prospected in it. I suspect (given
HPC has never been a large market. Heck, many of the Sun machines on
the top 500 list in the past have been database servers, or corporate
"mainframes" and not supercomputers in any traditional sense.
> the reference to "resource management" and "server applications") that
> it is seeking to plunder the relatively deep pockets and relatively
> non-computer-savvy researchers in the biocluster and medicluster market,
> with their bet backed by the proposition that they can reuse their
> efforts in a pinch in the HA market (those "server applications").
Hmmm. Me-thinks those who think that the biocluster folks have deep
pockets might be mistaken. And some of them are quite computer-saavy.
The module that launched a thousand web sites, CGI.pm was written by
Lincoln Stein, who has since done some other very interesting stuff.
It would be a mistake to mis-characterize these folks, and paint them
with broad strokes. Yes, I am biased, as this is where my companies
play. They (the scientists and end users) have some neat problems, and
they are computationally expensive and getting more so by the day.
> Bioclusters and biogrids as a market probably does total millions of
> dollars per year -- chickenfeed for M$ at this point -- but it also has
> some growth potential and integration potential for the future, when we
There are other very serious problems that simply building clusters and
grids will not overcome. We are starting to run into the edge of them
now, and they will only be exacerbated over time. There are problems of
commercial importance in here.
> reach the point where e.g. individuals are bioassayed in real time in
> their Dr.'s office, their genes are matched against a huge database
> scanning for defects and predispositions and damage and who knows what
> else, and genetically matched and tuned therapies are prescribed, where
> dog pedigrees and likelyhood of developing hip dysplasia are similarly
> determined in real time.
This of course depends upon cheap/fast sequencing, clinical micro-arrays
and other technologies, and the folks at 454 and a few other places may
indeed get us there. At that point you will need some serious
processing horsepower. Even if Moore's law can hold out for a while, it
will not get us where we need to be for this. If you want to know more,
contact me offline.
> Of course, Moore's law and so forth may eat this market before it fully
> matures -- in three or four years at expected rates desktop PCs will
> have (possibly multiple) TB size LOCAL SINGLE disks, multiGB of RAM
> standard, and at least gigabit interfaces standard with 10 Gbps not
likely ... (nVidia motherboards have built in gigabit as do a few others
now, for the consumer market, which they bill as "download from the
internet faster" ... :o )
> uncommon. With 10+ GHz CPU clocks, 64 bit and better data paths, really
There are limits to feature size, clock speed, and thermal dissipation.
With a small enough feature size, the thermodynamics of defect formation
becomes ever more important, and hence cooling and stabilization against
diffusion become even more critical.
We are just going to have to stop computing using fermions and start
using bosons. Talk about natural parallelism... :) How many photons
can you have on a single "wire" ?
> slick memory and device management, and perhaps some onboard
> parallelism, would one NEED a cluster to run such a (relatively simple)
> application? If not, it leaves one back with just the research market,
> and that is, I think, a hopeless case for M$ to make back their not
> inconsiderable development and sales costs in without a hedge. They
> would at best break even and at worst would both lose money and worse,
> lose money EMBARRASSINGLY and publically.
> As this list well knows, real, transparent, "windows-level" scalability
> is an elusive beast in the parallel HPC world, although you can find
> niches where it is possible. And there is no mass market here, per se
This is true.
> -- even if you can find somebody crazy enough to pay out LARGE software
> costs that scale per node, there simply aren't that many customers, and
> most of those customers always have several proven alternatives with
> costs that DON'T scale per node. Those customers will pay for support,
> they'll pay for service, but they don't like paying for software and
> REALLY hate paying for software and getting lousy service and poor
I don't know that people hate paying for software per se, rather they
hate not having the control over their software (e.g. cannot use it they
way they need to due to onerous licensing, binary only releases, etc.),
cannot fix problems with it, even if they know what they are.
> support to go with it. Plenty of people make money on service and
> support, but they don't get (unduly) rich and they work for a living
> earning what they earn.
Support is not easy. It is very hard to build a company out of
support. It is impossible as far as I can tell to get any sort of
external funding (be it equity, SBIR/grant, etc) to do support. It is
not a bad business, just nothing more than a small one. Large
consulting houses are amalgamations of small consulting groups ....
Aside from that, there is nothing wrong with reaping the rewards of your
hard labor, be they research recognition, awards, or remuneration. If
your work/research/product is bad, well, maybe there is something wrong
with it ... :)
What annoys me personally these days are the folks who simply repackage
and resell others open source work without contributing back to the
folks who did all the hard work in the first place. Not plagiarism, but
really a failure to acknowledge and attribute the source of innovation
at minimum, and support it. I see lots of that in the market. Real
innovation is hard. Those who innovate should be recognized for it.
Those who repackage should indicate that is what they do.
Joseph Landman, Ph.D
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web : http://scalableinformatics.com
phone: +1 734 612 4615
More information about the Beowulf