So the dust has settled. The first doting users have Power Macintoshes sitting on their desks and colleagues hanging over their shoulders demanding to know how fast X-Wing and Word for Windows runs. But while the desktop Power Macintoshes hogged the limelight there was precious little in Apple Computer Inc’s announcement to cast any light on its plans for new servers or its A/UX implementation of Unix. Apple’s plans for A/UX have been the subject of long and impassioned debate among the user community, the whole thing stirred up by the news that Apple was killing development of its current code-base, in favour of a bright, shiny, new A/UX – PowerOpen-compliant and based upon IBM Corp’s forthcoming AIX version 4.0. The coup de grace was due to be issued as we went to press: Apple has just announced A/UX 3.1, the last incarnation of AU/X as we know it.
Squeeze the last ounce
The company is also announcing an accompanying release of AppleShare Pro 1.1. Users won’t notice many differences other than improved speeds. Apple has apparently tidied up the kernal and done everything possible to squeeze the last ounce of performance out of the new software, but there are no major new features. Apple UK’s Steve Everhard (a man who simply has ‘Servers R Us’ on his business card where the job title should be) says that together, the new A/UX and AppleShare Pro should generally be about 40% faster. In special cases, such as the enumerate function, the server will be 140% faster – good news for those that obsessively recalculate their folder sizes, supposedly. Very nice, but all the improvements quoted are for the two new pieces of software when used in combination. Information on how much faster A/UX is by itself is not readily available, an indication that, at the moment Apple still sees Unix as a system for hosting AppleShare Pro, and little more. Next stop on the A/UX road will appear in the first quarter of 1995, when Apple is set to unveil the PowerOpen-Application Binary Interface compliant AIX-based offering. Whether the A/UX name will survive has yet to be decided, says Everhard. Will it look like today’s A/UX? Don’t bet on it. The primary interface will be Mac-like, and Apple will continue its tradition of what Everhard accurately calls masking the worst excesses of Unix. However there are no guarantees that the exact look-and-feel of A/UX will be preserved across the transition in the same way, say, that System 7 was preserved in the change from the 68000 family to the PowerPC line of RISC chips. Certainly A/UX applications will need to be recompiled, although AIX applications will not – so stand by to run those RS/6000 applications on Apple hardware. Son-of-A/UX, let’s call it PowerOpen for short, will run on new specialised server machines due to appear from Apple in the first quarter of next year.
By Chris Rose
They will, says Everhard, be built to an architecture radically different from anything that the company has produced before. He’s loath to give details at this stage, but acknowledges that the architecture will handle symmetric multi-processing, support memory in the Gb range and have multiple, high-speed communications buses. One of the things that Bull brings to PowerOpen is their great knowledge of multiprocessing software we intend to take advantage of that he says. It will also have swappable processor boards so that users will be able to plug in the latest processor boards as they emerge. It will not comply to the PowerPC Reference Platform. It is unlikely, however that you will be able to get the new Unix to run on existing Macintosh servers, Everhard says, although Apple is making the necessary stories about providing an upgrade path. While Unix has a secure future on Apple servers, it looks increasingly unlikely that the company will be offering the operating system on its desktop machines. There is, says Everhard still an open discussion within Apple about whether PowerOpen will appear on the desktop machines, but the balance of opinion is against it. The desktop Unix market is too crowded, to
much of a commodity, for Apple to be able define a niche: We are trying to keep an open mind, but at this stage of the game we are struggling to find what part we can play, Everhard says, adding that there is a lot of competition for research and development dollars within Apple. Instead of developing desktop Unix, the temptation is to spend the money on the Macintosh Application Environment, software already picked up by Sun and Hewlett-Packard which enables Unix users to run Apple applications in an X Window. But whichever way it jumps, we know that we cannot play a half-hearted role [in the Unix Market] like before Everhard admits. On the face of it, the same arguments apply to the server market too – if Apple has difficulty finding a case for going up against the Suns and the Hewlett-Packards on the desktop, then going up against the Suns, Hewlett-Packards, IBM, Bulls, DECs, Ncube’s et al in the server market cannot be easy either.
Bigger fish to fry
Everhard doesn’t see it that way and argues that the same attention to detail and ease-of-use that Apple bought to the desktop, will let it make its mark in servers. Partly, the firm intends to build on its traditional strengths, providing back-end engines for desktop publishing and the like. However, it also intends to break away from the Macintosh world, gaining more general industry acceptance. It also has an eye on this month’s hot technology – video servers – and is working on multiprocessor, image distribution servers using its QuickTime movie software. It is an audacious strategy and takes Apple into uncharted territory, in both technical and marketing terms. The company’s entire ethos is based around the desktop and its expertise with graphical user interfaces and close hardware-software integration. Big, multiprocessor boxes running AIX – does that sound like the Apple you know? – no wonder Apple’s Unix business does not want to squander research and development dollars on the desktop: it has bigger fish to fry.