IBM has announced the release of the eServer x360, the company’s first server featuring the IA-32 Intel Xeon processor MP (code name “Foster”) and incorporating IBM’s Enterprise X-Architecture. Powered by IBM’s new XA-32 core chipset, the x360 is a compact, four-way rack server built on a 3U rack-optimized design for space-constrained data center environments. The x360’s Enterprise X-Architecture features include modular scalability, enabling four four-way servers to be deployed as a 16-way machine; and remote I/O technology that allows dozens of PCI/PCI-X adapter slots to be located several yards away from the system. The x360 also provides mainframe-inspired self-healing tools derived from the IBM’s eLiza initiative including failure-resistant Chip-Kill technology and LightPath diagnostics. The x360 supports Microsoft applications and operating systems, Linux, Novell NetWare and other operating systems. The x360 will begin shipping in early December. No pricing details were included in the announcement.
Since Intel is expected to officially launch the new Xeon Processor MP sometime in the first quarter of 2002, and the company has selected IBM’s XA-32 core chipset as one of the Xeon’s validation platforms, IBM’s first-to-market claims for the eServer x360 are likely to be safe for the next few weeks. But does being first to market with Intel’s new Foster chips offer IBM any inherent advantages? Yes and no. On the plus side, major players including IBM, Compaq and HP (all of whom have committed themselves to delivering IA-32-based servers) have been roiling the waters over the opportunities that 32- and 64-bit Intel-based products are supposed to deliver. Given that market noise and Intel’s similar efforts, we expect Foster-based servers to inspire a flood of wet palms among vendors’ PR departments in early 2002. What customers will think is less certain. A built-in market for these products certainly exists among companies already dedicated to Intel and Microsoft, but after that low hanging fruit is gathered, where will vendors go for additional sales? Much of the success of Foster and its next generation big brother “McKinley” will hinge on just how well Intel and Microsoft actually deliver. If the dream of inherently more stable, accessible and scalable platforms for Microsoft productivity applications and operating systems is realized, we expect enterprises will send Intel and Microsoft to the bank, and lead to some erosion in the demand for UNIX-based solutions.
We also believe that Intel’s Xeon processors could usher in a new era and style of competition among hardware vendors. Beyond vendor assurances and the vagaries of benchmark testing results, it is difficult if not impossible accurately differentiate one company’s proprietary chipsets against another. In very real ways, Intel’s IA-32 and IA-64 core chips will offer hardware vendors such as IBM, Compaq and HP a level, industry standard playing field to match and compare the capabilities of their chipset technologies. The key to understanding IBM’s plans for its xSeries eServer products is the Enterprise X-Architecture solution set, which provides a host of flexible, scalable solutions that enhance the x360’s performance. In particular, considering IBM’s experience in mainframe computing, we believe that incorporating mainframe-style self-healing and self-managing processes from the company’s eLiza initiative offers IBM a leg up on the competition. In fact, the heady notion of Intel/Microsoft-based products with mainframe-like stability and flexibility will likely drive other vendors to develop similar solutions. In one sense, the new eServer x360 shows IBM’s server group continuing to execute its go to market strategy of the past year; leveraging and migrating appropriate technologies across its product lines, increasing the capabilities and value of all.
The UCLA Internet project has released the results of its 2001 survey on Internet usage, data acquired from more than 2000 household respondents nationwide. Among its findings are a number of metrics – compared with the 2000 survey – that show increased Internet use and sustained satisfaction levels among Internet users. In 2001, 72.3% of Americans are online, up from 66.9% the year earlier, according to this study. Student rates of Internet usage at school jumped from 55% to 64%. Average weekly hours are up as well to 9.8 hours per week, up from 9.4 hours. Satisfaction with overall Internet experience remains high, with a rating of 4 out of a possible 5. Respondents gave online shopping a 3.7 out of 5 rating. Users were least satisfied with the speed of their online connection, rating it a 3.2 out of 5. The most popular online activities are using email, instant messaging, Web browsing, buying online, finding entertainment information and reading the news. The main reasons for going online are to obtain information quickly, followed by work needs and access to email. Of the respondents that were not online presently, 44.4% said they planned to go online in the next twelve months, up from a 40% rate in the 2000 study. The primary reason for people not being online was “no computer,” which was also the main reason those that were once online and have dropped off have done so. The differences between the activities of new users (less than one year) and more experienced users (five or more years) showed that new users spent much more time in chat rooms, and long term users spent much more time doing professional work. The study also found that Internet use was cutting into television viewing among those online.
The Internet revolution will occur when it is invisible and not the stuff of news. We are getting much closer to reaching that revolutionary plateau, according to this study, with nearly three-quarters of the population accessing the Internet. It is notable that the single largest reason for not accessing the Internet is the lack of a computer. This indicates clearly that a PC without a connection to the Internet is worth about its weight in sand, and access is where it is at. It should come as no surprise that communicating (email, IM) top the list of Internet activities; the Internet is all about communication, something one-way media like TV or radio can’t provide. Seeking specific news also seems to be of high interest. People are becoming less patient with the idea of waiting for the news cycle to deliver information in a specific area of interest. The news that the percentage of students accessing the Internet is continuing to increase means that moving forward, the penetration rates of this technology will steadily increase as the “Internet generation” moves its way though the demographic cross section by means of the aging process.
While responses to online shopping questions are retarded a bit by security concerns, we suspect that such concerns are going to fade with greater familiarity with the idea that a credit card is more vulnerable in the physical world and that credit card companies have increasingly sophisticated fraud detection systems already up and running. Connection speed issues will obviously linger for some time, as too many upstart connectivity companies are dying a low margin death these days and it will take some time before the lumbering giants of the telephone sector move in to fill these gaps. For the time being, it appears that folks getting a taste of real high-speed access at work are, in many cases, going to have to make do with it for some time into the future. All in all, we see this report as indicating that the Internet has real trenchancy and staying power. Anyone arguing that the dot-bomb phenomena and the NASDAQ meltdown are indications that the Internet is overhyped is missing the boat. The revolution is just beginning, and whole bunches of people are on board.
AT&T has announced that it will be shuttering its i495 service on January 4, 2002. Under the current terms of the plan, users paid $4.95 per month for 150 hours of dial-up Internet access in return for viewing advertisements on the service. AT&T has notified the i495 customers that the service will be shut down and has offered to switch those customers to a $10.95 a month plan that provides only 50 hours a month. AT&T said the service was being shut down because “the bottom dropped out” of the online advertising market and ad sales were not subsidizing the low access fees contained in the plan. AT&T is also attempting to entice i495 plan members to use more expensive options, such as its $16.95 pr month offering that provides 150 hours of dial-up access.
Is anyone really surprised by this move? They shouldn’t be. We’ve been saying for years that the online advertising market is a mirage, and clearly AT&T is in agreement. What we see here is a confluence of revenue stream contraction, with both sides squeezing the middle. Bandwidth is a commodity; margins on it are slim if non-existent: witness recent moves by both Earthlink and AOL to raise basic subscriber rates. On the other side of the equation, we see Web advertising as failing to deliver on its promise to make up revenue for sites that charge no fees or minimal fees. Examples here abound, of course, with Web sites like Salon and Slate charging subscription fees and others, failing to do so, simply going out of business. Further evidence of the advertising myth of profitability can be seen in recent moves surrounding DoubleClick, which lost its CEO earlier this month and has come under increasing pressure to sell or close its Web advertising arm. DoubleClick has announced that it will beat analysts’ expectations next year, and actually turn a profit. To do so, new CEO Kevin Ryan said, the company would grow its email marketing business – no mention of increased ad sales. The company has also unloaded its European ad business. All this bears out something we have been saying for a very long time; online advertising is proving itself to be a largely illusionary revenue stream, one that was clearly inflated in its promise by the mania surrounding all things Internet. Now let’s get back to business.
IBM has announced the availability of its DB2 Universal Database for customers utilizing InfiniBand fabric. InfiniBand is a switch-based serial I/O architecture that improves the performance and reliability of high-speed connections between computers, and between computers and data storage devices. IBM says DB2 fits into the InfiniBand architecture without and modifications, posing no extra costs for customers. According to company, DB2 is the first database product in the market that can be utilized in InfiniBand environments. Hardware and software vendors including SAP, Intel, Network Appliance, InfiniCon, Lane 15 and Qlogic all support DB2 on InfiniBand.
At one level, IBM’s delivery of DB2 for InfiniBand is a product announcement that is unlikely to cause much of a ripple for most customers. InfiniBand solutions are in the earliest stages of availability/adoption. While the technology’s sheer performance capabilities suggests that it will likely make serious inroads in enterprise data centers over the next two to five years, the current demand for InfiniBand-enabled database solutions is low, at best. However, we are intrigued by the product development cycle IBM’s announcement suggests. While IBM is not the biggest database player on the block, it is the only one that produces and delivers InfiniBand hardware solutions. To take advantage of InfiniBand’s potential, database software leaders Oracle and BEA, along with other ISVs, will have to partner with enterprise-level InfiniBand solution providers (such as IBM) to bring their products to market. While partnering is unlikely to negatively impact the quality of eventual solutions, we believe IBM’s ability to provide for both ends of the product development process will likely allow the company to create, deliver and improve InfiniBand solutions more quickly than the competition. That could provide IBM a significant advantage if the adoption of InfiniBand by enterprises is as dramatic as many believe.
The Sageza Group, Inc.
Veterans Blvd, Suite 500
Redwood City, CA 94063-1743
650·366·0700 fax 650·649·2302
Europe (London) 44·020·7900·2819