c0t0d0s0.org

Then consider the fact that at IDC's breakfast briefing at ISC in Dresden this week, we were told HPC currently accounts for 19% of ALL worldwide server sales and 25% of (server) processors. With "niches" like that...

(Josh Simons in his blog)

Interesting times ahead of us: The thirth Version of the GNU General Public Licence will be released soon. I will be quite interesing, to observe the impact of this new licence to the OSS world. Will Linus use V3 ? Will Sun use it ? Will IBM do an marketing stunt and open source AIX unter GPLv3. Nobody knows.

theShepler summarizes the features of the initial implementation of NFSv4.1's pNFS for Solaris. Very interesting read.

Jim Laurent worte an nice comparisiion of the EAL 4+ evaluation targets between RHEL 5 and Solaris 10 with Trusted Extensions. It look like, that IBM/Red Hat leaved out most of the difficult parts. Perhaps out of technical reasons, perhaps to be a few month faster for a marketing stunt at the price of having an meaningless certification, as it excludes many standard parts of a modern Linux implementation (no desktop, no swapable disks, no networked trusted printing, but it read your self).

I had an email in my inbox from a reader, who asked me for the math behind the 300 switches. So i cite the email of Andreas Bechtolsheim to Jonathan Schwartz (Jonathan published it his blog, you will find more information about our Constellation announcement in his blog)

Begin forwarded message:
From: Andreas Bechtolsheim
Date: June 28, 2007 6:58:59 AM PDT
To: Jonathan Schwartz
Cc: John Fowler
Subject: 3,456

We implement a 5-stage fabric, and with a 24-port switching element
the maximum number of ports is n*n/2*n/2, or 24*12*12 =3456.

Other Infiniband switches in the market today are 3-stage fabrics
and they have n*n/2 or 24*12 = 288 ports.

Now you can build a 5-stage 3456 port switch with 12 288-port switches
and 288 24-port leaf switches but you end up with 300 boxes occupying
about 456U of rack space or 12 racks, and 6912 cables.
We use one double rack with 1152 cables, so it is 1/6th the space,
1/6th the cables and 1/6th the weight.

On Jun 28, 2007, at 6:36 AM, Jonathan Schwartz wrote:

so - why 3,456 ports?

But this is not the most interesting point Jonathan makes in his blog. He answers the question, why we do such systems:

The academic supercomputing community (there's that word again) sets the pace for enterprise computing across the world – which has grabbed on to HPC for an array of real world challenges, from virus, disease, and drug discovery, to customer purchase pattern analytics, capital markets trading, energy discovery, dynamic resource management - you name it, it's one of the fastest growing segments in the marketplace. Proving that what starts in academia, ends up on main street. Industry looks to academia and research institutions to understand the innovations that enable breakthrough scale and performance (just ask Linus - who, come to think of it, still hasn't responded to my dinner invite... I hope it's not my cooking.)

It´s not only a simple marketing stunt. The problem of HPC in really large scales ist to an part, that these systems need large amounts of labour to build them. Getting from

here	to there

transforms HPC from an hand-crafted environment to an more commoditized environment, so new sets of customers and problems gets in a economical reasonable reach for HPC systems, thus opening new markets for such environments, thus moving HPC to a more standard tool in the architectural toolbox.

Ich ärgere mich zwar auch über diese äusserst freundlichen und zuvorkommenden Luftsicherheitshilfskräfte (Achtung! Ironie! Aber: Die in Hamburg sind freundlicher als in München, die Dresdener freundlicher als die Hamburger), die sich da jedesmal zwischen mich und meinen Abflug stellen, aber sowas ist mir auch noch nicht passiert.

Ich spreche ja normalerweise nicht wildfremde Frauen aufm Flughafen an, insbesondere wenn diese durch das Verwenden eines Kopfhörers zu verstehen geben, das sie nicht angesprochen werden möchten (ja, bin ein Schüchtener einer ). Aber heute musste ich das einfach. Ein Rätsel ist geklärt.

Dave Miller announced, that he ported Linux as a guest to the LDOM environment.Some parts are still missing, but hey ... it´s the first version ...

Sun: We Can Build a Faster Supercomputer Than IBM
Constellation Petascale Computing

We should start to choose our internal development code names in a way, that plans the leakage of them. (Nobody called the Sun Enterprise 10000 with it´s official name, some called them E10K, many used Starfire). In either way, the codenames still leak to the press in an amazing speed: The Register presents the codename of our new IB switch. Okay, this one was obvious.

With Sun Cluster (at least since 3.0) we have a really excellent product to build high availability clusters. Yesterday the first bits of this product went open source at opensolaris.org.
Or as Nicholas writes:

We're committing to open-source the full Solaris Cluster product over the next 18 months, including the agents, core, Geographic Edition, test suites, and docs. We expect to open nearly three million lines of code! (Two million lines of product code and one million lines of test code).

An excellent article about our new super computer and the leading brain behind this system - "Billionaire Thinks in Trillions for His Computer Designs"

A number of Silicon Valley technologists are, however, betting on Mr. Bechtolsheim. “He’s a perfectionist,” said Eric Schmidt, Google’s chief executive, who worked with Mr. Bechtolsheim beginning in 1983 at Sun. “He works 18 hours a day and he’s very disciplined. Every computer he has built has been the fastest of its generation.”