May 7, 2004
Earlier this week IBM announced its eServer i5, the first of IBM’s computing systems to be powered by its POWER5 microprocessor. POWER5 delivers 276 million transistors per processor based on IBM's 0.13-micron copper wiring and silicon-on-insulator technologies, and the CPU provides simultaneous multithreading features that effectively allow a single processor core to be seen by operating systems and applications as two logical processors. Advanced "micro-partitioning" technology allows customers to run as many as ten virtual servers per POWER5 processor, and the Virtualization Engine extends the eServer i5’s ability to run multiple operating systems simultaneously thus integrating formerly discrete UNIX solutions alongside existing OS/400 and Linux applications running on the POWER5 CPU, and Windows applications executing on an x86-based integrated eServer xSeries. The eServer i5 570 also offers Reserve Capacity on Demand, On/Off Capacity on Demand, and Memory Capacity on Demand options for customers that wish to purchase additional memory and compute power on a transient or peak load basis. All eServer i5s are pre-loaded and tested with i5/OS, which provides the base software most businesses need, including IBM’s DB2, and WebSphere–Express, a scalable transaction engine for ebusiness applications. The IBM eServer i5 will be available on June 11, 2004. Pricing for eServer i5 520 starts at $9,995 and the eServer i5 520 Express Edition (available through IBM Business Partners) is priced at $11,500. The IBM eServer i5 570 is priced starting at $85,200.
The venerable eServer iSeries or AS/400 is one of the great urban IT legends. Most everyone has heard one variant of the tale whereby a system was found encased in cement, submerged fifty feet deep in the water table, and completely forgotten by IT only to be later unearthed by archaeologists who found the system still humming away with users happily accessing its cuneiform-based applications. OK, so this is more legend than reality, but there is no doubt that the iSeries has a loyal, enthusiastic user base. With this latest incarnation, the eServer i5, there is plenty to cause existing and new users to take note. The eServer i5 with its POWER5 multithreading features is much more than a box that can sit in the corner and run the accounting system forever and ever: it is a very capable high-performance computing solution that offers notable infrastructure simplification and consolidation opportunities. Customers can simplify their IT infrastructures by leveraging underutilized assets and seamlessly accessing IT resources available on the local network while consolidating multiple servers and integrating multiple operating systems environments into a single management facility.
The plethora of capability posited by the i5 in many ways appears to offer mainframe style computing on an SMB budget. A poor man’s mainframe this is not. Rather, the i5 is a properly scaled, high-value-added solution that squarely targets the pain points increasingly felt by SMBs of all stripes. The realization to a small or medium-sized business that it could consolidate several or even scores of i5/OS, UNIX, Linux, and Windows servers onto a single server must be intriguing indeed. In addition, each of these operating environments has full access to system resources, be they CPUs, storage, I/O, or other. The cost reduction in maintenance and ongoing operation could be significant, especially for an organization that finds itself knee deep in cabling and systems. Overall, we believe the eServer i5 sets a new standard in the SMB marketplace, one that is historically respectful (backwards compatible) while offering state of the art operating system support, integrated management facilities, as well as a host of value-added capabilities. While other vendors may offer their customers either/or solutions supporting a base OS along with Linux iterations, the i5 provides either/or/or/or flexibility that is simply unmatched in the marketplace. Who said innovation was dead?
This week Hewlett-Packard made several organizational, services, product, and performance announcements. Several of the announcements were made at the European customer venue, ENSA@work, where 5,400 customers gathered to hear the direction and future of HP and the Adaptive Enterprise. The company announced the completion of the ongoing reorganization of the Enterprise and Services divisions, the availability of the new dual-processor mx2 module featuring two Itanium 2 processors that can plug into existing HP Itanium systems, and a range of Non-Stop Service offerings focusing on availability and business continuity. Additionally, HP announced the HP StorageWorks Reference Information Storage System, a new product focused on easy archival and retrieval of email and other information, as well as expanded ISV support for HP’s Multi-OS server strategy.
These announcements reflect HP keeping pace with its competitors and customers’ requirements and represent the additional puzzle pieces they need to craft a more complete and recognizable enterprise computing strategy. A year after its announcement, Adaptive Enterprise continues to be HP’s strategic mantra and it is important that the company stay the course in developing the strategic concept. However, HP continues to place the burden of recognition of the application of Adaptive Enterprise on its customers. The HP Darwin Reference Architecture that describes the implementation of Adaptive Enterprise is component rich and reflective of HPs’ available products and near-term directions, but is by itself is insufficient to clarify Adaptive Enterprise for customers or the market. It is like trying to order a complete meal from a menu of ingredients. While one can make analogous connections between competing components, products, and overarching direction from IBM and its On Demand architecture, it is not easy to do so with HP’s offerings. Where IBM has crafted an understandable story, HP provides no real tactical framework, no strategic signage, and no contextual directions to the very worthy idea of building a highly responsive, adaptable enterprise computing environment. Maybe the company’s branding of the Darwin Reference Architecture makes particular sense, since Darwin’s theory posited that evolution is not planned, but the result of random events that sometimes work to a species’ benefit.
We are pleased to see the new Services and System Group did not need the extended runway it projected last fall to emerge from its reorganization. We hope that Ann Livermore as General Manager will once again deliver the kind of no-nonsense leadership with which she integrated the HP and Compaq Services Groups. Unfortunately, it was this rehash of the reorganization that was the most interesting plank in this raft of Pythonesque “We’re not dead yet” announcements. As we have hinted before, given Carly’s ongoing and obvious consumer enthusiasms, this consolidation represents a stronger opportunity for the Enterprise Group to spin out as an Itanium-centric systems and server house to compete against Sun and IBM. HP has the technology and the talent, but may be running out of time to establish its strategy and technology as a leader in the business-oriented computing model of the future.
IBM has announced WebSphere Integration Server Express, the latest product in the Express portfolio, which is targeted at the SMB market, most notably the mid-tier enterprise. The Business Integration server contains components of WebSphere Application Server Express, WebSphere MQ Express, Express, and IBM DB2 UDB Express and the WebSphere Business Integration Workbench Modeler. The product comes with integration templates, a Web-based dashboard to monitor business processes and a graphical debugging tool for testing end-to-end integration. The Integration Server Express supports Windows, Linux, and OS/400 and the base model is priced at $5,999 per processor.
IBM has positioned this offering as one it can sell direct, but which can also be sold by business partners into the mid-tier. By combining many parts of the Express portfolio in a single package, IBM is giving its business partners the ability to use what they need for mid-tier integration projects in a single package. Some projects may require using all of the integration tools. Some might only need a fraction. But like other Express offerings, IBM has simplified their deployment and limited the number of knobs and buttons in an effort to minimize complexities and the opportunities for either partners or customers to screw up.
What we find most interesting about IBM — and other large enterprise vendors — is their ongoing zeal to capture the mid-tier portion of the SMB market. In our mind, IBM is perfectly positioned to do so, even though it must overcome the perception that it designs and sells products strictly for the big guys and therefore are not competent play in the mid-tier market. That misperception is largely based on a misunderstanding about the true nature of the SMB IT market opportunity. When we look at mid-tier companies, we see businesses that are creating huge amounts of data each day, and are struggling to keep up with a data production curve that looks a great deal like that of a large enterprise of just a few years ago. Secondly, we see mid-tier companies doing business within a value chain that includes a small number of larger customers. To operate in such an environment, they must have IT deployments that interact with their customers. In such a light, it is clear to see that the mid-tier market opportunity is not a separate market per se, but one that is an extension of the large enterprise market itself. As such, we see no reason that large enterprise IT vendors can not reach down-market. In fact, we would argue that they are perfectly suited to do so, and their customers have an opportunity to deploy products that take the true measure of their place in the value chain.
HDS has announced general availability of the Hitachi Thunder 9585V, a new addition to the company’s Thunder product family. The 9585V boasts new levels of functionality, scalability, and throughput, and is optimized for input/output-intensive applications such as rich media, data mining, high performance computing (HPC), and highly utilized databases. According to HDS, the 9585V’s faster logic processor and enhanced software can deliver over 240,000 I/Os per second, performance the company claims is up to 47% greater performance than competing products. HDS announced the 9585V as part of a new Application Optimized Storage (AOS) solution strategy, which the company says allows enterprise customers to precisely match business application requirements to Hitachi storage system attributes. The first iteration of this strategy combines the new Thunder 9585V arrays with the latest version of HiCommand solutions including Storage Services Manager, Path Provisioning, QoS for Sybase, and Tuning Manager 3.3. According to HDS, the ability of AOS solutions to leverage information as a means to improve business performance offers customers greater gains than data lifecycle management. No pricing information for the Hitachi Thunder 9585V or any AOS solutions was included in the announcement.
While a bit hyperbolic even by IT press release standards, HDS’ new product announcement offers some interesting food for thought concerning the company’s larger tactical and strategic ambitions. On the tactical side, the new Thunder 9585V attempts to fill a hole in the company’s mid-range product line for the market currently led by EMC’s CLARiiON CX700 and IBM’s FAStT 900 arrays. HDS has done well enough with its high-end Lightning solutions, but the mid-market has been a tougher nut for the company to crack. More importantly, that same space has been a boon to the fortunes of both EMC and IBM, so it is no surprise that HDS is interested in improving its position. But the company faces a serious impediment already overcome by its rivals. EMC’s fortunes in the mid-market have improved markedly through its partnership with Dell. IBM has enjoyed historical success in this market, and has done a good job of leveraging existing and emerging customer relationships into FastT sales. While HDS has profited from its OEM partnerships with HP and Sun for the company’s Lightning products, Thunder has followed a lonelier path.
That is where HDS’ new strategic effort comes in. Instead of simply positioning the 9585V as a notable upgrade to the Thunder 9580V, the company instead has come up with AOS, which aims not just to improve hardware sales but to offer an alternative to the Information Lifecycle Management (ILM) solutions that have dominated storage market discussions for the past few months. Why HDS would do this is pretty simple: since the company does not offer the depth or breadth of lower-end and archiving solutions that competitors do, any HDS-centric ILM pitch would ring pretty hollowly. Developing enough new solutions to reach all the necessary ILM-specific niches for a legitimate offering would be prohibitively expensive and time consuming. Instead, HDS is claiming that optimizing applications for specific arrays (and vice versa) and keeping things in line with the appropriate performance measurement and management tools offers better returns than ILM. Is this reasonable? While there is something to be said for and gained from fine tuning hardware and software, AOS smacks of changing the subject to avoid embarrassing questions. In essence, HDS is trying to lead the dance by changing the tune. These new offerings play to the company’s technical strengths, but much of the market and most of the competition is headed along a quite different road. Overall, we doubt that many besides the HDS faithful will be tempted to take the AOS off ramp.
Microsoft announced this week that it has signed an agreement with IronPort Systems to reduce the amount of spam sent to users of Microsoft’s Hotmail and MSN services. IronPort will provide services that screen email senders, requiring them to meet certain criteria and behavior standards, while posting a bond to IronPort. If any of the approved emailers violate the terms of their agreement with IronPort and send unapproved spam to users, they will forfeit part or all of their bond payment. IronPort’s screening process includes certification by Truste, a privacy organization, which determines how the prospective emailer is gathering addresses, and whether they have any existing business relationship with recipients. Microsoft said it hopes that by using the IronPort service, it can add granularity to its Brightmail anti-spam service, so that spam is filtered out while legitimate email is not blocked. The service will be paid for by those seeking bonded status with IronPort; MSN and Hotmail users will not pay for the service.
Spam has become one of the great irritants of Internet life, and those signed up with free email services seem to be sizable targets for spammers. As a result, they seem to cry the loudest at times, even though they seem to forget that old maxim, “You get what you pay for.” That said, even freeloaders deserve relief from the relentless intrusions of spammers offering home loans and a number of other topics that if mentioned by name would trigger spam filters preventing this week’s dispatch from reaching your inbox.
Various strategies have been deployed against spammers, including spam filters, lawsuits, legislation, and the like. Some have proven effective; others have not. In this case, a major ISP is essentially going the route of pre-approving mass emailers before they are allowed on the system. Bonded bulk mailers, like Amazon.com or other e-tailers will in essence get the Good Housekeeping Seal of Approval, allowing them to mail recipients who opt in to such communications. Such a seal will cost the end consumer nothing, but does ensure at least some sort of vetting process has been conducted before the mailer is allowed access to your inbox. Will this stop spam? Of course not. Just as we struggle to dispose of the junk mail delivered by the U.S. Postal Service, we will continue to have to weed through our inboxes for the foreseeable future. At least now, we will know to some extent where to direct our wrath while expunging certified spam from our lives.
McAfee Research, Network Associates’ technology research division, has been recently selected as the industry partner to round out a team of three universities — UC Berkeley, USC Info Sciences Institute, and Penn State — focused on Internet security issues. Together the team will build a large-scale cyber-security test bed for the development of new defenses against computer worms, viruses, and Trojans, along with other cyber threats such as denial of service and routing infrastructure attacks. The project consists of two components: Cyber Defense Technology Experimental Research (DETER) network, which involves the creation of an experimental infrastructure, network, tools and supporting processes; and Evaluation Methods for Internet Security (EMIST), to include the development of scientifically rigorous testing frameworks and methodologies. The project’s goal is to create a network supporting the development and demonstration of next-generation information security technologies for cyber defense to enable national-scale network experimentation in emerging security research and advanced development technologies. The project will be funded by $10.8 million over a three-year period by the National Science Foundation and the Department of Homeland Security.
In the current environment surrounding the Internet — an ongoing flood of security news, advisories, warnings, alerts, and threats — it would be easy to dismiss this partnership and project as marketing noise. However, doing so would fail to acknowledge the growing and future importance of cyber security, and its potential impact on McAfee and other security vendors. Cyber threats have grown exponentially over the last five years. According to Computer Economics Inc., the cost of cyber attacks has grown steadily from $3.3 billion in 1997 to an estimated $12 billion in 2003. In 2003, the SoBig.f virus alone had a worldwide financial impact estimated at $2.5 billion, and the cost of the January 2004 MyDoom virus was said to exceed $4 billion. As a result, protecting systems and networks against malicious threats is becoming more and more critical to businesses, the economy, and national security. This will not be easy. The future protection of cyberspace will be complicated by the continuing growth of the Internet and the increasing sophistication of the threats, hacks, and exploits.
The importance of this project for McAfee is considerable. Under increasing financial pressure and with competitors such as Cisco, ISS, Symantec, and Trend, Network Associates has been quietly intensifying and refining the company focus exclusively on security. Recently, the company sold its Magic service desk solution division, and its Sniffer network monitoring, analysis, and performance division. In addition, new investments have been made in security technologies, such as Entercept/Intershield, with the intention of leapfrogging the competition. We believe that The Cyber Defense project is a strong affirmation and continuation of McAfee’s strategic security focus. While this project is intended to be the breeding ground for solutions to larger security issues, it could also provide McAfee an edge up on competitors. However, how much McAfee will gain commercially from its involvement may well be complicated by academic institutions whose focus is not on the commercial success of McAfee, but rather on research and discovery.