Rash's Judgment: Official EISA Announcement Month
 
Microbytes Daily News Service Copyright (c) 1989, McGraw-Hill, Inc. It's
official. October became the official EISA Announcement Month for a number of
companies. One of those companies, Hewlett-Packard, is showing the first
EISA-bus machine to be seen in public, a machine that's also on the cover of
the latest issue of BYTE.
 
H-P is the first in what will soon become a flood of these new machines, all
aimed at becoming an alternative to IBM's Micro Channel Architecture. Already,
companies that were not in the original "Gang of Nine" that developed the EISA
specification have announced that they will make EISA machines. And the Gang
isn't standing still either.
 
As BBS News and BYTEWEEK reported last week, Compaq's Rod Canion has confirmed
Compaq's plans to announce its EISA machines on Monday, November 6. Zenith,
meanwhile, has made no secret of its plans to begin shipping an 80486-based
EISA machine in December. Olivetti has begun showing prototypes of its EISA
machine. The rush will continue after the first of the year.
 
IBM, meanwhile, is still pushing its proprietary Micro Channel. While there is
some activity by other manufacturers in Europe to develop a Micro Channel
machine, efforts in the US have been aimed at EISA. While there has been some
activity among peripherals manufacturers to support MCA, most of the products
developed so far are merely MCA versions of similar items that already existed
for the PC AT bus. To date, the rush to develop cards that take advantage of
the MCA's ability to move data or allow bus mastering has been underwhelming.
 
Whether the industry will approach EISA with a similar yawn remains to be
seen. Clearly, there will be no rush to develop things like EISA modems or
Arcnet cards. These already exist for the AT architecture, and there is nothing
to be gained by adding EISA capability. As a result, existing products can be
used immediately, which is, of course, the rationale for EISA in the first
place.
 
On the other hand, there is some indication that EISA is getting the attention
where it matters. DPT, for example, has already announced a caching disk
controller that uses the bus mastering capabilities of EISA to allow
significantly faster transfers of information to and from a hard disk. Since
the disk channel is one of the major bottlenecks in applications such as file
serving, this could be an important product indeed.
 
There has been less interest in supporting IBM in this manner, and part of the
reason is the company itself. Almost all MCA board developers face the daunting
possibility of having to compete in some way or another with IBM. For example,
IBM already supplies a hard-disk controller with its Micro Channel machines,
and while it's not the best controller in the world, it's not the worst,
either. More important, it is Already Here. The result is that there's little
incentive to develop superior disk controller technology for MCA. Likewise,
there's little incentive to move ahead with advances in video technology with
MCA systems. IBM already offers an optional 8514/A high-resolution graphics
adapter, and there are persistent rumors that this graphics adapter will
eventually -- like VGA -- be brought down onto the motherboard. An MCA version
of a high-speed, 32-bit VGA card like H-P's may never be developed.
 
At this point, the momentum of the industry outside of IBM is away from MCA
and toward EISA. But now that the EISA specs are out in the open, IBM will
surely try to point out its weak spots. And MCA, of course, has tremendous
inertia, so it's not clear that the EISA momentum will overcome it. What is
clear, however, is that EISA is a reality, that systems using it do work, and
they should be taken seriously. And simply because of the large number of
companies involved, it's likely that it is in EISA where we will find the
innovation.
 
                              --- Wayne Rash
 
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
