Who is building Oracle’s Internet Machine?
This is the title of a thread I started February 5 of this year, because of my interest in the underlying operating system, and probing to see if anyone else would see the virtues of VxWorks for the NC. (A serious embedded system RTOS like VxWorks is perfect as a base for NC, because: (1) it is multi-threaded, an absolute requirement to make Java efficient, (2) real-time, needed for multi-media and other serious I/O, (3) lean, a legacy of embedded systems but critical when memory is scarce, and (4) multi-platform, great to keep developers options open. Notice that these conditions do not preclude pSOS or OS-9.)
Posts on the thread hit all over the map, from ATTs Plan 9 to Acorn computer. It finally ran out of steam when it seemed that Oracle was making their own Network Operating System, which at the time struck me as odd. (Operating systems are extremely difficult to make from scratch, mainly because of all the I/O and other standards that must be implemented and the need to amass a comprehensive tooset for development and testing.) Needless to say, I got less out of the tread than I had hoped to get.
Today came the final answer to the thread: The winner is Wind River Systems!
Establishing a strategic relationship with a software company with the vision and presence of Oracle provides a priceless endorsement of Wind River’s tools and RTOS. When coupled with Intel’s I2O endorsement, and HP’s de facto strategic relationship to WIND, there can be little doubt that WIND is the market leader. Also it doesn’t hurt to anticipate the run-time license revenues that will follow a successful debut of NC2, Oracle’s Network Computer, 2nd Generation.
Finally, do not ignore the irony of a partial paradigm shift (PCs under attack by NCs) being implemented using ubiquitous computing. This is the similar to the irony of both PCs and network servers being dismembered from within by ubiquitous computing (in the form of distributed architecture with intelligent components such as I2O, intelligent printers, intelligent disk systems, intelligent graphics and wireless components.)
The final curiosity is that Oracle and the other database companies have been discussed recently on this thread as a model of how the embedded systems RTOS vendors might evolve, and that it would make sense for Oracle to enter the RTOS market through acquisition. It is now obvious that Oracle needs industrial-strength help perfecting a Network Operating System, but if Ellison would just extend his vision slightly he might see the larger opportunities that lie beyond.
Allen, |
