Intel exits, but InfiniBand intact

The news is grim–or is it? Semiconductor giant Intel, which had planned to develop chips to enable InfiniBand, decided to apply its development budget elsewhere. Is this the beginning of the end for InfiniBand, stillborn before it ever reaches commercial deployment?

Hardly. The announcement is more about Intel’s resources than InfiniBand’s promise.

InfiniBand is an interconnect fabric that can connect disk arrays, storage area networks, local area networks, servers, and server clusters. It differs from technologies built on PCI in the same way that ATM differs from Ethernet: InfiniBand uses a point-to-point, channel-based connection model instead of a shared bus. Its 10Gbps full-duplex (20Gbps aggregate) theoretical throughput is well-suited for high-end enterprise usage where a lot of storage, network, and CPU resources need to work together.

The InfiniBand market is very active, as evidenced by the 180 members of the InfiniBand Trade Association. Mellanox Technologies and IBM already have chips out in the hands of component manufacturers. Volume production is expected next year. InfiniBand will be deployed initially on high-end database hardware, processing huge amounts of stored data.

Like everyone else in the technology sector, Intel is facing tough times. It has to make decisions on what products are key to its business for the future, given that it doesn’t have infinite ability to fund every avenue. Presumably, that means that whatever chips most closely support its bread-and-butter CPUs are the most strategic for the company.

While dropping its development plans, Intel says it still supports the technology, and says it will introduce a server chip set with hooks to InfiniBand. It also remains invested in InfiniBand startups like Mellanox, Banderacom, and Lane15 Software.

Instead of InfiniBand, Intel is expected to devote a lot of developmental resources to PCI Express, an upcoming I/O bus specification formerly known as 3GIO. (Don’t confuse PCI Express with PCI-X, the 133MHz, 64-bit, 8.5Gbps version of the PCI bus, which competes with InfiniBand as high-performance alternatives for servers and workstations. See this white paper for a high-level overview of PCI Express.) PCI Express will co-exist with and ultimately replace the PCI bus found in every Intel-based PC since the introduction of the Pentium. It offers higher bandwidth–2.5Gbps, versus PCI’s 1.06Gbps, with the promise of scaling higher by adding additional serial “lanes”–and advanced features like packet prioritization and power management.

Given the mass-market size of PCI Express and its synergy with Intel CPUs, it makes sense for Intel to apply its development resources here. Look for the first PCI Express machines to make an appearance in 2004.

I’m not worried about InfiniBand’s immediate future. Throughout history, the leaders in new waves of technology have been different players from the leaders in the existing dominant technologies. None of the railroad giants became airline powerhouses. No newspaper company built a radio or TV empire. And while IBM invented the ISA bus, Intel’s PCI bus took over.

It remains to be seen who, if anyone, will emerge as the giant of InfiniBand. There’s room for several–adapter makers and the server vendors who employ the adapters to enhance their hardware’s performance. With HP, Dell, IBM, Sun, and Microsoft among the InfiniBand Trade Association’s founders and supporters, it’s clear the technology has broad support.

Just as the PCI bus had to fight it out with the VESA bus, InfiniBand faces competitors like PCI-X 2.0, but InfiniBand’s immediate future looks bright.

Skip a generation

We all know how grinding the software update schedule can be. Many application vendors are on an annual release calendar so they can refresh their revenue streams. But not all those annual releases include features compelling enough to make them worth the headache of deploying them within your organization.

The problem is much worse with client operating systems, though they tend to get upgrades on a slightly longer calendar. Windows ME was less stable than Windows 98. Windows XP has virtually no penetration within businesses yet, so it’s hard to say how it stacks up against Windows 2000. However, because XP uses a radically different code base than 98 or ME, ensuring it works with all your organization’s applications may be tricky, despite the extensive work Microsoft put in to promote good compatibility.

Server operating systems are the worst. Stability is a key criterion for server operation, and anything as major as a new operating system is bound to adversely affect that stability (unless you have a quirky server to start with, in which case you may want to grasp at any straw that promises help). Windows 2000 was a major upgrade to Windows NT in that it added Active Directory. Yet, according to IDC analysts, almost 60 percent of installed Windows operating systems at the end of last year were still running on Windows NT.

These organizations have the right idea. If it ain’t broke, don’t fix it.

All vendors want to entice you with compelling new features, some of which are actually worth paying for. Directory services, for instance, can be extremely valuable because they can simplify ongoing management tasks across the enterprise, especially when third-party application vendors take advantage of them for authentication.

I propose, therefore, that you plan to skip a generation in your software purchases, beginning with your next client and server operating systems.

If the bulk of your clients are running Windows 98, move to XP with new hardware purchases. On the other hand, if you’re running Windows 2000 Professional, pass on XP and stick with 2000. There’s virtue in having a single standard client platform. If you have some Windows 98 mixed in with your Windows 2000 clients, things get a little trickier. I’d pick whichever platform predominates and apply my rule to it.

The same rule applies on the server side. If you skipped NetWare 5 after biting the bullet and moving to Novell Directory Services in NetWare 4, consider upgrading to NetWare 6. If you’re still running NT and have passed on 2000, consider Windows .Net Server, due later this year. It will enable attractive policy management features in Windows XP clients, which, chances are, you will roll out someday. Even though you can’t start piloting .Net Server yet, you can use a copy of Windows 2000 as a stand-inon your test subnet to gain familiarity with Active Directory; any plan you devise to move your NT domains to Active Directory will still apply to .Net Server.

If you’re already standardized on 2000, ignore .Net Server; the incremental gain won’t be worth the acquisition expense or implementation time. Any time you plan a major server operating system upgrade across your enterprise, it behooves you to consider all your choices. That’s particularly true right now, as Microsoft’s deadline for customers to sign up for its Software Assurance licensing plan looms on July 31. Most customers haven’t agreed to let Microsoft feed them regular updates in exchange for regular payments, and I don’t blame them. I want to pay for software on my own timetable. Opting out lets you do that; you pay more each time, but only when you choose to update.

Microsoft’s licensing demand comes at a bad time for its customers, thanks to the economy, but a good time for its competitors, whose products are stronger than ever. NetWare 6 is still an excellent file and print server, and the new version adds some very attractive Internet and Web access features. Linux is growing in popularity, and though it still lacks the number of server-based applications Windows can boast, Linux offers at least one of just about every kind of enterprise management application available.

Microsoft will no longer sell NT after 2003 and drop support at the end of 2004. That doesn’t mean you need to clean it out of your server room by then. If your servers are relatively stable and you don’t plan to add new hardware or software to them, let ‘em run. Instead of the thrill of running the latest and greatest, you’ll have the satisfaction of simply running.

Enterprise P2P: A solution looking for a problem

The buzz around peer-to-peer computing in business has abated since it pegged the hype-o-meter a year and a half ago. That’s when Groove Networks released its eponymous application. While Napster was turning P2P into a synonym for “copyright violation” in the consumer arena, enterprise P2P applications revealed themselves as solutions looking for a problem.

As I first pointed out in a column last year, P2P doesn’t scale well beyond the workgroup because managing resources on a central server is more beneficial for large organizations.

Mind you, that doesn’t mean P2P can’t be a useful tool for certain purposes. In fact, there’s one criterion that, if present, makes a P2P product quite appropriate for a given problem: If you’re trying to tie together members of a group where no one is in charge or everyone is a volunteer, then P2P is the way to go.

Because most businesses have a hierarchical structure, server-based applications facilitate collaboration just as well as peer-to-peer applications do, and also offer the advantages of server-based authorization and centralized backup. But when organizations of equal status have to share information (imagine different branches of the armed forces, or a pharmaceutical company working with researchers in hospitals) P2P technologies can help them share intellectual and technical resources.

I’ve been talking about P2P as if everyone had the same understanding of the term, but it ain’t necessarily so. Many products that try to hop on the P2P bandwagon aren’t true peer-to-peer applications, in that they require a central server to mediate the peers’ communication. Consider Napster, for instance, as well as SETI@home and most instant messaging applications. None of them allows clients to communicate directly with each other. A true P2P application, like the Gnutella file-sharing program, doesn’t require a server to guide connected clients.

On the other hand, if you define P2P, as British consulting group Ovum does in Enterprise P2P: Flexibility and ROI as “any application or process that uses a distributed architecture and allows peers to provide and consume resources,” you open up the definition to a multitude of sinners.

Given that loose definition, Frost & Sullivan’s recent forecast that P2P components of centrally managed products may prove useful could be on the mark. Collaborative P2P environments like Groove can help supplement more traditional client/server collaboration products.

I’m glad developers are still working on P2P applications. Sometimes, as with instant messaging, it’s useful to go under the radar of the corporate Information Services department. But I still don’t see P2P providing a strategic advantage for most organizations any time soon. There are already client/server products that do what most P2P applications do, and that are more compatible with hierarchical corporate networks.