It's more than egos involved.
Obviously that is a factor but it really comes down to a fundamental difference in vision. SUNW is committed to the processor-centric model of computing which has dominated since the early mainframe days. While SUNW coined the clever slogan "The network is the computer" what they really mean is "The computer is the network", or perhaps "The network of computers is the computer". This worldview shows in SUNW's storage strategy, which is no different from the storage strategies of HWP, IBM, CPQ, and belatedly DELL: First, to prevent erosion of the existing install base and second to use "multi-platform" storage as an inroad to convert non-SUNW shops to the true faith. SUNW makes no secret about this strategy, as indeed it is the only sensible storage strategy for a processor vendor to adopt. Unfortunately, unless they abandon it they will find themselves on the wrong side of history.
The reason INTC got out of the memory business in the 1980s is because they saw that it was becoming a commodity business. Ironically, the reason INTC is now rapidly diversifying away from microprocessors is for the same reason. To be sure, INTC will not abandon the microprocessor market that it was largely responsible for creating, but if INTC sees the need to diversify itself that says something profound about the future of the processor business.
Processor vendors have long paid lip service to the notion of "portable computing", "write once programming", and "processor independence" all the while doing everything in their power to lock customers into their respective architectures. However the rhetoric is now becoming reality. The rise of the Internet and most recently B2B e-commerce has meant that with each passing year value is shifting away from processor-specific applications towards processor-independent applications designed (rather than retrofitted) to work in the connected world. This new world is inherently multi-platform and has two intertwined facets which characterize it: communications (information in motion) and storage (information at rest).
Ten years ago networking was viewed as part of a processor's I/O subsystem and IBM made a fortune selling expensive communications gear to customers locked into IBM's proprietary Systems Network Architecture (SNA). It was the genius of CSCO to realize that the networking function would detach from the mainframe and become a separate entity independent from the architectural control of any processor vendor. A decade later the same situation exists with regard to storage and it is the genius of EMC to have recognized early on that like networking before it storage too would detach from the processor and become a separate entity. Within a very few years from now the notion of attaching disks to servers will seem archaic. The diskless server becomes the new peripheral to the storage network.
The EMC thesis is that the storage market in 2000 is like the networking market in 1990--just starting to break away from processor captivity. Given the inherent inertia of the industry it may well take another five years before this separation becomes manifestly obvious to all just as it wasn't until 1995 that the Internet suddenly became visible. In the meantime processor vendors will continue to voice "storage strategies" which are little more than thinly disguised efforts to protect and expand their processor franchises. But that's OK. EMC shareholders can afford to be patient and settle for today's growth levels while waiting for the real growth to kick in around 2005. Because the real storage revolution has hardly begun. |
