JC - here's another cut at what I was saying, with a few facts. Before 1992, in other words, for the first 10 years SUNW was in business, 100% of revenue came from workstations. SUNW, along with the Unix movement, created a new paradigm of network computing where the linkage of machines was more flexible and abstract than the mainframe centric architectures, or the other primary competition, which was VAX clusters and DECNet. Sun helped fuel that whole notion of distributed computing with a variety of remote process models, NFS, and the like.
When HP introduced the 9000 series, and especially the 800 models, the game changed. HP was selling "big iron" into accounts which had previously been HP3000 shops, or VAX shops, or even mainframe shops. Sun did not play in that market, but they wanted to, so they drove SMP server development which resulted in the 490, 690 and other server products.
At the same time, their architectural direction started to shift away from a full-up distributed model, partly because of the performance hit that their Unix file systems took in comparison to centralized server models. Development of Solaris started about that time also. SunOS was never very good at SMP.
By the mid-90s, Sun had a respectable portion of revenue from servers - about 20% - and had already determined to drive up the food chain to counter the decline in workstation revenues. By 1998, the majority of revenue came from servers.
There's nothing wrong with what they did - it was exactly the right thing to do IMO. But still, it seems a little funny to see press claiming that it was always that way.
The Network is still the computer - more so now than ever. But it's a different network, and a different bunch of computers, and a very different architectural direction than it was during the first 10 years of SUNW's life. |
