Do you remember how DELL found it's way to where it is? DELL *believers* kept reaching for the falling knife that just wouldn't stop.
I don't think the comparison between EMC and Dell is apt because Dell's stock eventually reflected the changing dynamics of its core business.
Dell earned its premium valuation in the 90s because it was consistently able to grow faster than the rest of the PC industry with its highly efficient build-to-order model.
Unfortunately, even Dell couldn't avoid the inevitable slowing down of the rich North American market as the PC penetration level started to approach the 60% level. This, of course, does NOT mean that the PC business is dead. Rather, it just means that it's going to grow more slowly and differently than the last 10 years, where annual growth peaked at the 30-40% range in the early to mid-90s before slowly tapering off to the current teens.
Compounding this dilemma is the fact that the year-long delays in Itanium (formerly Merced) and Windows 2000 gated Dell's growth in the enterprise market. Up and until its most recent decision to OEM Unisys' 32-way server, for example, Dell had no participation to the $26 billion enterprise server market dominated by Sun and IBM.
Then there is the storage opportunity. Forrester Research recently forecasted that the top 2500 of the world's largest corporations will be spending an average of $19 million on storage in 2003. Dell's procrustean assumption is that storage hardware will eventually be standardized and sell like workstations. This may be true for the SME (small and mid-sized enterprises) segment, but the competition between EMC and IBM at the highest level is being waged on the basis on the ability to provide a seamlessly integrated device-sharing and data-sharing stack that addresses the fact that it is the management of storage resources that accounts for up to 12x the cost of the hardware.
For example, EMC has long had the basic ability to move any NT, Unix and mainframe data internally (disk-to-cache-to-disk) in a single Symmetrix or externally between networked Symmetrix/Clariion boxes in a SAN based on SCSI, Escon or Fibre Channel. This unique type of data extraction allows an NT or Unix server to read mainframe database files and vice-versa. This is extremely useful for the average large IT organization where 70% of the information still resides on mainframe systems (November 1999 survey by IDC) complemented by an average of around 4 different Unix platforms (1998 average from Ibexx survey), NT, multiple databases and multiple applications that sit on top of those databases. Note that the current database wars pit Oracle, which is trying to integrate all the applications with its database products, against IBM, Microsoft, Seibel, PeopleSoft, I2, BEAS, etc., which are trying to preserve the database-agnosticism of most enterprise applications. The escalation in this war between programmers means that the type of concurrent soft and hard locking mechanisms developed by EMC will become even more valuable in an increasingly more complex computing environment.
There is just no way to develop that type of capability overnight unless one has invested like EMC in a billion-dollar interoperability lab that contains 250 server models, 40 operating systems, 50 clustering environments and 145 networking elements all subjected to real-world I/O traffic. EMC started investing in their lab in the mid-90s when they made the transition from the mainframe market to the open systems market. EMC expects to spend billions more on interoperability in the next few years. |
