Hewlett-Packard Co has been lifting the veil on its future hardware plans, particularly in the area of high-end systems where its business has been soggy of late. The top-end HP3000 960 is to be followed by the first multiprocessors in the 3000 line – the RISC Precision Architecture is designed to support four-way multiprocessing – […]
Hewlett-Packard Co has been lifting the veil on its future hardware plans, particularly in the area of high-end systems where its business has been soggy of late. The top-end HP3000 960 is to be followed by the first multiprocessors in the 3000 line – the RISC Precision Architecture is designed to support four-way multiprocessing – and are likely to include HP9000 Unix as well as HP3000 MPE/XL machines. Next year, a processor faster than the current 15 MIPS one is promised. At the low end, while the company will continue to sell the 16-bit Micro 3000, which sells for $11,000 to $16,000, the company plans a low-end version of the 32-bit RISC CPU for a 3000 that will be in the $15,000 to $30,000 price bracket, to be followed in 1991 by an entry model supporting eight to 16 users and starting at $11,000. And the 32-bit version of the operating system is due for a facelift, with release 2.0, due in November, remedying two glaring omissions from the origianl version – direct connection to X25 and to bisynchronous communications lines. In 1991, the Posix applications programming interface will be incorporated into MPE/XL. The company believes it needs the proprietary operating system as well as UNix because MPE offers about 50% better througput in transaction processing. On the 68000 family workstation front, the company has two separate development efforts for the 68040, one for its own HP9000 300 family, the other for the Apollo Domain line. The Apollo version will have limited expansion capability while the 9000s will be modular with a greater range of storage and memory options. The two versions of Unix, HP-UX and Domain/OS will continue to be offered side by side for a couple of years because the planned successor to both from the Open Software Foundation is not expected to have all the things Hewlett needs before 1991 – and with the notorious delays that seem to be built into any software development dependent on IBM, it may well be later than that. Meantime both new machines are being designed to support both operating systems. On the people front, growth below the planned 20% in the HP3000 business – it is still at over 10% means that Hewlett is 600 people, 10%, too heavy in that area. About 200 are on short term contracts that will not be renewed, another 200 will be shed by attrition, and 200 will transfer to other divisions. Direct sales is being restructured into multi-user systems and workstation groups: the division had been commercial and technical.