New Chip Portends Major Changes for Software Sales Groups

By Geoffrey James

This segment of the newsletter usually reviews several announcements from large software organizations. Recently, however, a product was announced that completely dwarfs in significance everything that’s happened over the past year.

In February, IBM, Sony and Toshiba announced a CPU chip that uses what they call a multicore architectural design. The chip provides supercomputer-like floating-point performance with observed clock speeds greater than 4 GHz. The chip, code-named Cell, is the first chip in more than a decade that threatens the Microsoft-Intel axis (aka Wintel) If Cell proves successful, which appears quite likely at this point, it will rapidly shift attention and then sales away from software that crutches on Microsoft server architectures, such as .NET, in favor of server software built atop Microsoft’s main competitor, Linux.

Here’s why Cell is important. To achieve speeds beyond 3.6 GHz or so, CPU chips generate so much heat that the transistors and other components inside the chips no longer run reliably. This has caused enormous problems for Intel, resulting in the cancellation and delay of several highly visible chip projects, followed by a new strategy. Rather than attempting to make single-core CPU chips run faster, Intel is touting dual-core architectures that run cooler because they couple two smaller CPUs rather than one big, hot one to achieve the same overall computer power.

Cell, with eight cores, and more to come, leapfrogs Intel’s latest designs, providing more compute power in a chip package that runs surprisingly cool. More significantly, because Cell will be the CPU for the Sony PlayStation 3, Cell will be manufactured in quantities large enough to keep the price competitive with Intel. In addition, IBM designed the chip from the start to be optimized for Linux, which means IBM soon will announce Cell-based servers that will be cheaper and faster than comparable Intel offerings.

Even the rumor of such machines will upset the cost-benefit equation for server software. One of the reasons Wintel servers became so popular was that they were cheaper than comparably powered servers running UNIX or another proprietary operating system. Linux is taking market share from Microsoft in several server segments for the same reason. Because Linux is free, it’s usually cheaper than Microsoft server software, even if the two systems are running on the same Intel hardware. Thus, unless IBM bungles the pricing, the Linux-Cell server combo will be an order of magnitude cheaper than the Windows-Intel server combo.

The Cell is like an earthquake at the bottom of the ocean that grows into a tsunami. Because of this, software sales groups must take action now! Here’s what you must do.

  1. Evaluate your current server product offerings to ensure you are sufficiently deployed in the Linux segment.
  2. Build a contingency plan to shift resources away from Wintel sales and toward Linux sales as the Cell begins to gain momentum.
  3. Create a plan to rapidly retrain sales, sales support, marketing and technical support, and for the conversion of sales materials and sales tools.
  4. Closely monitor announcements from IBM to determine when the affect of Cell will begin to change the software market. We’ll keep you posted, naturally.
    1. The above is based on a series of conversations with several industry analysts, including Shane Rau (IDC), Dan Hutchinson (VLSI Research), Roger Kay (IDC), Rob Enderle (Enderle Group), Cary D. Snyder (, Boris Petrov (Petrov Group) and Dave Mack (Technology Business Research).