Good comparative report. Let me make a few embellishments, then I will try to respond to your major question concerning the winner-take-all proposition.
My mention of WIND’s plentiful resources were mainly in regard to its relatively huge hoard of cash. You indicated WIND has $65 million vs. $12 million for MWAR (and $52 million for INTS). Actually WIND has about $95 million cash equivalents. If you re-check the balance sheet figures for the latest quarter, you will see about $27 million in long-term investments. The "investment" made me curious about the possibility of an early equity investment in another company, possibly as a precursor to an acquisition. But the investment turns out to be account-speak for treasuries over 12 months duration, i.e. cash equivalent. I definitely consider $95 million vs. $12 million a major difference in resource.
As you noted, comparative headcounts are tricky. For example, many WIND engineers classified as "service" contribute to R&D, since many service contracts result in additions to available products. Also, all three companies (WIND, INTS and MWAR) are adding rapidly to staff. But I didn’t mean to imply that WIND currently overwhelms MWAR, or certainly not INTS, in headcount.
About the winner-take-all proposition. This takes two possible forms: oligopoly and monoply. IBM was a monopoly; still is for mainframes. Wintel is now for PCs. IBM could not, and cannot, be dethroned from top of the mainframe computing paradigm. Wintel cannot be dethroned from the PC paradigm. These companies can only be surpassed by a paradigm shift, not direct combat.
Should the ubiquitous computing RTOS space consolidate instead to a small number of industry leaders, who will persist for the life of the paradigm, these leaders will form an oligopoly. If it doesn’t consolidate at all, it will just remain yet another competitive technology sector. There are many examples of oligopolies in technology, but one that may be a precursor for the ubiquitous computing paradigm is the RDBMS space. At the industrial-strength database level, ORACLE reigns as market leader, but it shares the market with other leaders: Informix and Sybase as ISD’s, and IBM (DB2) as a non-pure play. So far MSFT (Foxbase, Access and SQL Server) has not been competitive at the industrial-strength level, nor has Borland (dBase and Paradox are mainly personal databases, while Interbase has not been able to successfully challenge the market leaders). A host of next-paradigm databases (data warehouse and object oriented) are struggling to break through the high switching costs from legacy databases offerings by the market leaders.
Notice what has happened in the RDBMS space. The low-level, personal database vendors use cut-throat prices and struggle to survive. MSFT has dominated this space due to its PC monopoly. Meanwhile, the industrial-strength database vendors have been reduced to the three pure plays mentioned above. What happened to early contenders e.g. Ingres, Unify, Progress? What happened to relational query languages like "quel" that was wide-spread and arguably better than SQL?
The reason Oracle dominants market share is because it ported to every conceivable platform, and marketed aggressively, and slowly became fully featured. Thanks to Ellison’s technical leadership (note his interest in parallel processing via N-Cube led to Oracle’s technical leadership in parallel processing), Oracle finally became a solid, advanced, fully-featured, hard-to-administer, industrial-strength, market-leading SQL database. Hard-to-administer is important. It bestows priesthood on experienced and tutored professionals and guarantees absolutely huge switching costs.
Will Sybase or Informix ultimately go away? No. They also have high switching costs, and offer plenty of religious dogma of their own, with established disciples.
What about successful new entrants? Possible but very difficult. Unfortunately for new entrants, there will not be a rapid shift to a new database paradigm due to the enormity of legacy (now meaning SQL) databases. Moreover, Oracle, Informix and Sybase absorb new database concepts as they surface rapidly enough (after letting insignificant startups prove concepts and market acceptance) to maintain leadership. They currently boast of object-oriented databases as well as data warehousing functionality - all with the added benefit of "seamless" connectivity, or extensions, with legacy databases. (As an aside, Hitachi and WIND just announced THOR, an object-oriented database using VxWorks with a Sysbase-compatible front-end.)
Which version of winner-take-all will ubiquitous computing RTOS turn out to be, oligopoly, or monopoly, or will it remain totally competitive? The best guess is oligopoly, exactly like what happened to the relatively mature RDBMS space. But if one of the current leaders leaps ahead, there is the possibility of one company monopolizing the ubiquitous computing RTOS space. On the other hand, I do not think a totally competitive, better mouse-trap like market, can endure the barriers to new entrants in this space. (Most technology sectors are highly competitive due to simple selection criteria to win sales: cheapest product meeting specifications wins.)
As an oligopoly, you are guaranteed to do very well with WIND, MWAR and INTS investments, as long as each remains a market leader. As the leader, ORCL has done extremely well the past few years for investors, but even as a lessor leader IFMX has done sensationally for investors. Until lately, SYBS was also a stellar performing stock. If the ubiquitous computing RTOS space narrows to a monopoly, you have to hope you bet on the right horse. My guess would be WIND.
Allen, |
