Liudvikas Bukys

EE Times:

10-Tflops computer built off the shelf

Lawrence Livermore National Lab is putting together a supercomputer that will boast nearly the same performance as the ASCII White system from IBM Corp. that the lab now uses but it promises to be 10 times cheaper. Called Evolocity, the system will be the fastest clustered supercomputer in the world, according to Lawrence Livermore.

…

“This network approach is nice because we can use a standard PCI slot on each processor node, which gives a 4.5-microsecond latency,” he said, as opposed to 90-µs latency for Gigabit Ethernet.

The network uses bus host adapters on each node, supporting a 320-Mbyte transfer speed in one direction and 400-Mbyte bidirectional throughput. Each processing node is a server board from SuperMicro Inc. (San Jose, Calif.), built around the Intel E7500 chip set with two Xeon processors running at 2.4 GHz. The boards are linked by a network assembled by Linux Networx into a clustered system that will have 960 server nodes.

The file system, called Lustre, uses a client/server model. Large, fast RAM-based memory systems support a metadata center, and data is represented across the enterprise in the form of object-storage targets. “Being able to share data across the enterprise is an exciting new capability. It will allow more collaboration among research projects,” Seager said. For example, workstations on the network running visualization programs can directly access data generated by Evolocity.

MSS Initiative Makes Progress
Phil writes
“The MSS Initiative was started by Richard van den Berg and myself to combat sites that are broken (enable Path MTU Discovery AND block ICMP 3,4) which include such big sites as SecurityFocus and CERT (causing those behind PPPoE and other less-than-1500-MTU-protocols to be unable to view the sites). This past week we were priveleged enough to be able to present a paper at the 16th LISA Systems Administration Conference! Check out the paper and slides and be sure, like many members of the audience, to fix the sites you administer!”
[Slashdot]

Slashdot: Members of the Oasis interoperability consortium approved the

Security Assertion Markup Language (SAML)
on Wednesday as an OASIS open standard. The move paves the way for the XML-based framework to enable secure SSO (single sign-on) and other security functions for Web services transactions spanning multiple hosted sites.

Jakob Nielsen’s Alertbox, August 19, 2002::
Let Users Control Font Size:

Sometimes technological progress backfires, and the “better” technology turns out to be worse for users. The Web is no stranger to this problem, and has experienced many innovations that would have been best avoided. Examples include frames, changing the color of browser scrollbars, and scrolling text.

Another example of harmful Web technology comes with the increasing use of style sheets , which let web designers specify the exact size of text down to the pixel. Unfortunately, many designers are using this ability, leading to reduced readability of an increasing number of websites.

…

We can’t wait for Microsoft to ship a good browser, though that has to be the ultimate solution to the font size problem. For now, websites can increase readability by following these guidelines:

• Do not use absolute font sizes in your style sheets. Code font sizes in relative terms, typically using percentages such as 120% for big text and 90% for small text.
• Make your default font size reasonably big (at least 10 point) so that very few users have to resort to manual overrides.
• If your site targets senior citizens, use bigger default font sizes (at least 12 point).
• If possible, avoid text that’s embedded within a graphic , since style sheets and font size buttons don’t have any effect on graphics. If you must use pictures of text, make sure the font size is especially large (at least 12 point) and that you use high-contrast colors.
• Consider adding a button that loads an alternate style sheet with really big font sizes if most of your site’s visitors are senior citizens or low-vision users. Few users know how to find or use the built-in font size feature in current browsers, and adding such a button within your pages will help users easily increase text size. However, because every extra feature takes away from the rest of the page, I don’t recommend such a button for mainstream websites.
• Maximize the color contrast between the text and the background (and do not use busy or watermarked background patterns). Despite the fact that low-contrast text further reduces readability, the Web is plagued by gray text these days.

Aaron Swartz:
“Imminent End of Net Predicted“, a lighthearted commentary on:
Brandon Wiley:
“Curious Yellow: The First Coordinated Worm Design“

When the Spam Hits the Blogs:
Web spammers populate referral logs.
“They’re trying to jump-start a meme.”

Scientists Plan to Shake Hands Via Internet:

Pushing on the pen sends data representing forces through the Internet that can be interpreted by a phantom and therefore felt on the other end,” said Mel Slater, Professor of Computer Science at University College London (UCL)…

UCL will conduct the experiment on Tuesday with colleagues from the Massachusetts Institute of Technology.

File Serving:

It is convenient for students and faculty to have access to their files
no matter what computer they’re using, so many universities provide
some sort of network storage. Requirements for a network storage
system include:

• Scalability. It must be able to handle use by all members
  of the university. (e.g. at MIT, it must be able to handle 15000 users.)

• Reliability and security
  (it has to withstand use by a large number of highly intelligent,
  curious computer science students with lots of time on their hands!)

• Compatibility with Windows, Linux, and Macintosh
• Low Cost. It is likely that only a free software solution will
  be inexpensive enough to be used on every computer on campus.

• Support. It must already be deployed at many universities,
  and books and other training materials must be readily available.

The building block services for a network storage system include:

• File Serving
• Authentication
• Time Synchronization
• User Directory
• Domain Name Service

Oddly enough, it seems there is currently only one choice for the File
Serving service that meets all the above requirements: AFS (in the form
of OpenAFS).
That dictates several other choices:
AFS requires Kerberos as its user authentication service,
and Kerberos requires NTP as its time synchronization service.
Both probably require DNS as their hostname lookup service.

Lots of useful links re AFS, Kerberos, integration with clients, help desk pages at various universities, Active Directory, evaluations/critiques of alternatives, and other related topics.

PGP poised for major comeback [SecurityFocus]

Stanford gives ‘A’ to distributed computing
A group of chemists–including Stanford assistant professor Vijay Pande–said they successfully predicted the folding rate of a protein using calculations worked out on a so-called distributed computing network. Their research, conducted last year, was published this week in the science journal Nature.
ZDNet Oct 23 2002 8:12AM ET [Moreover – Tech latest]