But Wait, There's More

Sun, University of Texas Launch Maverick Portion of TeraGrid

Sun Microsystems and the University of Texas have created a new remote supercomputer visualization system that will plug into the TeraGrid, a joint effort by U.S. government-sponsored research centers and universities to create a gridded network of high-performance number-crunching machines.

While clustering RISC/Unix servers to create parallel and massively powerful machines is old hat, and the use of relatively inexpensive Lintel and Lopteron clusters is common, no matter what kind of supercomputer you create, to make human use of that machine means turning terabytes of data into pretty pictures that represent whatever phenomenon you are modeling in the simulations inside the supercomputer. Up until now, scientists who wanted to make such pictures to analyze models had to either move a subset of the data to their workstations or be sitting in the same location as the supercomputer cluster to have it draw those pictures. With the "Maverick" Terascale Visualization System, developed by Sun and the University of Texas at Austin, and funded by the National Science Foundation, researchers have created a separate visualization server that can take the models from supercomputers and do the necessary graphical rendering to illustrate sophisticated computer models and then pump the resulting pictures in real time over a secure Internet link. In other words, researchers no longer have to be in the same facility as the supercomputer to do their work.

In effect, Maverick is the frontal lobe of the TeraGrid project, which, in 2002, was funded with $88 million to create a 20 teraflops supercomputing grid that is spread out across the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, the San Diego Supercomputer Center at the University of California, the Argonne National Laboratory, the Center for Advanced Computing Research at Caltech, and the Pittsburgh Supercomputer Center. Funding for TeraGrid comes from the National Science Foundation, as well as from private partners, including IBM, Sun, and Intel.

Maverick is not a particularly big machine, at least not compared with the supercomputers it can render visualizations for. It is comprised of a single Sun Fire Enterprise 25000 server, using 1.2 GHz dual-core UltraSparc-IV processors, with 512 GB of main memory, and rated at a mere 265 gigaflops. However, that is a large amount of main memory for storing graphics data and the E25K has a very high memory bandwidth, so it can coordinate the movement of data at memory speed to the eight high-powered graphics cards that are lashed together and working in concert to render images. The system runs Sun's Solaris Unix variant, of course, and the graphics cards are running in Linux boxes, attached to the E25K through a high-speed network. The E25K has a special tweaked version of Sun's Grid Engine software that allows researchers to dispatch visualization jobs to this machine over the Internet.

Sun is certainly not the first company to figure out that you should gang up a bunch of 3D workstation video cards to make a visualization box. In July 2003, SGI announced a similar visualization machine, called the Onyx4 UltimateVision system, which can cluster from two to 32 high-end graphics cards in order to create what is in effect one giant and very powerful multipipelined graphics card. But Sun claims to be the first company to allow remote access to such a kind of visualization system.

Sun, Siebel Strengthen Alliance with Solaris/Opteron Support

CRM software maker Siebel Systems and server maker Sun Microsystems have expanded a six-year alliance that will see Siebel port its future release of its software to Sun's Opteron-based servers running the future Solaris 10 operating system.

Scott Anderson, director of the Siebel alliance at Sun, says that about one third of Siebel's customers are already running on Sparc/Solaris platforms, and that those companies tend to be the larger ones that spend big bucks on hardware and software. That is a very large market share, considering that Siebel runs on just about any modern platform, so this deal to port the full suite of Siebel apps to Solaris/Opteron servers is as much about preserving Sun's presence in the Siebel CRM installed base as it is about expanding it down into the SMB customer set, where Windows truly dominates.

While Siebel has strong alliances with IBM and uses its WebSphere middleware and DB2 databases to create its products, it seems likely that Sun will work with Siebel to ensure that Siebel 8 (if it is called that) will run on the Java Enterprise System, Sun's alternative middleware, and various databases (particularly Oracle and maybe Sybase). The agreement between Sun and Siebel includes codevelopment and comarketing provisions, but does not include co-sales efforts. Financial details of the deal were not disclosed.

Sybase and Sun Demo 1 Trillion Row Data Warehouse

Sybase and Sun Microsystems have been partnering for years to show that the database maker's Sybase IQ 12.5 analytical database, which is distinct from its OLTP-oriented ASE 12.5 database, is scalable enough to take on the biggest data warehousing jobs. Sun and Sybase have demonstrated a data warehouse running on Sparc/Solaris servers and using Sybase IQ 12.5 that scaled to 1 trillion records, the largest data warehouse anyone has benchmarked to date. Two years ago, Sun and Sybase set a record when they benchmarked a similar data warehouse setup with 180 billion records.

Sybase IQ is designed for data warehousing and for rapid ad hoc querying of databases, and it is designed to use a minimal amount of iron (relatively speaking) to support such queries. In the system under test, which was audited by Francois Raab of InfoSizing (which commonly audits the TPC benchmarks, among others), Sun clustered two 24-way Sun Fire 6800 servers using 1.2 GHz UltraSparc-III processors to a 48-way Sun Fire E6900 server using dual-core 1.2 GHz UltraSparc-IV processors. Each server had 96 GB of main memory, the processors ran at 99 percent CPU capacity, and they all had access to data stored on a SAN, comprised of StorEdge 6320 and 6920 disk arrays. Those 1 trillion records amounted to 155 TB of raw input data, which was sucked into Sybase IQ, indexed and summarized until it expanded by a factor of three, and then compressed down into 55 TB of disk storage. Sun tested the configuration running both Solaris 9 and Solaris 10, and claims that, using a conventional relational database, supporting 1 trillion records would have required 1 PB (petabyte), or 1,000 TB, of disk storage.

Sun and Sybase have more than 700 customers running the Sybase IQ software, which is popular among financial services organizations that have to chew through very large amounts of data. AJ Mahajan, senior solutions manager at Sun, says that Sybase IQ has been highly tuned for Solaris servers (this is the platform Sybase uses to create the software, in fact) and that the combination can offer a data warehouse that scales anywhere from two to 10 times larger than anything the competition can field.

Avnet, HP Partner to Push X86, Unix Servers in Mexico

The Mexican units of server maker Hewlett-Packard and IT distributor Avnet announced that they have extended Avnet's HP server reseller agreement in the United States to cover Mexico. Avnet has operated a reseller unit in Mexico since 1985, which now has 55 employees; HP says it is the dominant supplier of X86 and Unix servers in the Mexican market, and is looking to Avnet to help it expand that operation. The Mexican reseller agreement between Avnet and HP also covers storage and systems management software. Avnet will be providing financing and tech support services under the deal as well.

HP Sets Up Blade Server Division, Readies HP-UX Blades

While Sun Microsystems was hogging all of the press while it was visiting Wall Street, Hewlett Packard, which has two of the early pioneers of commercial blade servers (counting HP and Compaq separately), decided to create a new division dedicated solely to blade servers. The division was not just created as a PR stunt (although the timing surely looks like it was), but reflects the reality that blade servers span multiple hardware and software units and need to be presented as a single, integrated product line to customers.

This division, called appropriately enough the BladeSystem division of the Enterprise Systems and Storage group, has Rick Becker as its vice president and general manager. HP expects to sell about $500 million in blade servers in 2005, and that about half of its X86-based servers will be in the blade form factor by 2008, and given the unique opportunities and engineering challenges that blades represent, HP needs a team focused on blade servers.

In addition to announcing the formation of the BladeSystem unit, which rebrands the existing two-way ProLiant BL20p and BL30p blades and four-way ProLiant BL 40p blades as BladeSystems, Becker raised the curtain a little on the hardware plans for his new division. Specifically, he said that HP would roll out variants of these blades based on the Opteron processors by the end of 2004 or in the first quarter of 2005, called the BL25p, BL35p, and the BL45p, respectively. He confirmed that, as HP has promised in the past, it would deliver an Itanium-based blade that is capable of supporting HP-UX (and very likely Linux and Windows as well) sometime on the first half of 2005. Why it is taking so long to get a two-way Itanium blade server to market, now that the "Deerfield" variants of the Itanium 2 chip have been on the market for a year now, is a bit perplexing. The issue might have more to do with HP-UX or the Itanium application ecosystem than it does with hardware engineering.

In addition to announcing the BladeSystem division, Becker said that HP was rolling out the Essentials Automation Controller Pack (which is a mix of code from acquisitions and existing OpenView systems management software) that can provision blade server hardware, their operating systems, virtual LANs, and other networking features. HP also rolled out the Essentials Patch and Vulnerability Pack, which is integrated into the new System Insight Manager 4.2 (which now equally supports Windows and Linux platforms). The Patch and Vulnerability Pack is based on OpenView Radia (which is the new name for the patch programs that HP got through its acquisition of Novadigm). HP also rolled out Integrated Lights Out Manager 1.6.2, and Becker said that HP would eventually release a future program called Essentials Power Governing that will allow companies to automatically clock down the X86 processors in their machines when utilization is low, which can save them 20 percent on their electricity and cooling bills.

He also said that HP was working on something called the Essentials Virtual Machine Management Pack, which would integrate with SIM 4.2 and which would provide a single management front-end for virtual machines based on VMware, Microsoft, or HP (specifically, HP-UX vPar) server virtualization technologies. This software pack will be available in the first quarter of 2005, and it may be the thing that is holding back Itanium-based HP-UX blades. The company is furthermore finishing up the code on the Essentials Intelligent Networking Pack, which will deal with network bottlenecks by determining the least-clogged path between any two devices on the network and figuring out alternate paths in realtime when network congestion hits.

Finally, HP also rolled out a new BladeSystem installation startup service that has HP bring in a technician and train an administrator on how to use SIM 4.2 and set it up on a half dozen blade servers for a $1,600 fee. Additional "statement of work" services for more complex software installations or additional blade server installations is available at an unspecified hourly rate.

Government Study Finds IT Offshoring on the Rise, but More Data Needed

A new government study indicates that offshoring is growing but there is not enough data to be certain of the total impact offshoring on the economy. The non-partisan Government Accountability Office analyzed the Department of Commerce's trade data and found that offshoring of IT jobs is indeed growing. From 1997 to 2002, the GAO found that offshoring of business, professional, and technical services grew from a $21.2 billion business to $37.5 billion, a 77 percent increase. What's more, the study found that outsourcing jobs to a subsidiary, what the GAO refers to as "affiliated trade," accounted for 71 percent of all IT outsourcing and is growing faster than unaffiliated trade, which grew at a 67 percent rate over the period. Among the different BPT services, computer processing, accounting and bookkeeping, and research and development jobs have grown the fastest since 1997, although, curiously, the amount of computer processing work being sent offshore actually declined from about $1.5 billion of work to $1 billion, from 2000 to 2002, while R&D grew steadily to $1 billion.

The GAO also looked at data from the Bureau of Labor Statistics and found that, from March 2001 through June 2004, in industries commonly associated with offshoring, like computer system design and accounting and booking, Americans lost jobs at a higher rate than industries not associated with the phenomenon. The average annual rate of decline over this period was 5.7 percent of all jobs in computer systems design and related industries, and 7.9 percent for accounting and bookkeeping, while during this period total nonfarm employment actually increased by two-tenths of a percent. The study also found that total projected U.S. employment for the period from 2002 to 2012 is 2.4 million jobs lower than what was projected. However, the GAO says it cannot, based upon available data, come to any concrete conclusions about how offshoring is affecting the overall economy, and specifically cited the collapse of the dot-com bubble as one of the key events clouding the employment picture.