25 Sep
2000
25 Sep
'00
10:04 a.m.
At 02:20 PM 9/25/00 -0400, Geoffrey G. Rochat wrote:
It's kind of tough for a
person [Ken Olsen] who has built a multi-billion-dollar company from
scratch over 20
years to realize, let alone admit, that the magic may well have moved
on, and that then-current realities may be different from those abroad
in "The Good Old Days."
Actually this is an extremely dangerous condition that kills companies all
the time. When you are a "pioneer" you create the rules, having created the
rules you forgot that you created them, someone who is trying to beat you
thinks outside the rules, and kills you.
Sun actually killed DEC, and sometimes I think having a collection of the
real milestones in that battle would be good. For example the Sun 3/280 was
the 19" rack mounted Sun 3 architecture machine. It typically had a couple
of SMD drives (Eagles) in the lower half and the card cage in the upper
half of a rack. Some had a tape drive on top. This machine was directly
aimed at the MicroVAX 3900 which was DEC's one-rack "killer" VAX at one
time.
Sun broke two key rules:
You can build "commercial" timesharing systems on a commodity
microprocessor.
You don't have to write your own Operating System.
By using the 68020 and using BSD, the Sun 3 only needed to pay for a modest
OS and hardware staff (and no chip design guys), whereas DEC had to cover
the cost of their custom chips, their own OS, and their own hardware and
busses. Thus Sun was able to sell "tims sharing servers" for less than the
equivalent VAX cost and still make enough profit to grow the company.
Interestingly, Sun now sits where DEC was (their own CPU
architecture/busses + OS) and so they are very vulnerable to someone to
come along and kill them with open software based X86 servers.
--Chuck