Xyplex’s MAXserver 6625 X25 Gateway has been certified by British Telecom North America for use on the Tymnet world-wide X25 network. The certification is a statement of compatibility between the BT Tymnet network software and the Xyplex X25. Xyplex’s MAXserver 6625 combines the functions of a multi-protocol terminal server and an X25 Packet Assembler/Disassembler to form a high-performance gateway on a single card for the MAXserver LANbus intelligent communications chassis. MAXserver 6625 enables TCP/IP users on Ethernet networks to access resources on X25 networks, and users on X25 networks to access TCP/IP resources on Ethernet nets.

IBM Corp and Apple Computer Inc’s joint venture announced on July 3rd (Table 1) is clearly a big win for the Open Software Foundation and Motorola Inc. It should keep IBM and Apple very competitive in the worldwide market but won’t produce significant market share gains. Losers would be Motorola 88000-based desktop suppliers if there were any, writes Lewis D Brentano, vice-president, InfoCorp.

Table 1: IBM-Apple Joint Venture Overview

Product Features Timeframe*

1. Object Oriented Limited backward 1994-95 Layer above OS compatibility; for development/ forward execution on X86 compatibility; (OS/2), 680X0 (Mac) 90’s technology and RS/6000 (AIX) systems

2. Enhanced AIX to OSF/1 extended 1993-94 support Mac with Mac interfaces and functions – an programming as well upgrade of Apple as Motif A/UX (V.3.2)

3. Power PC RISC Reduces five chip 1993-95 microprocessor RS/6000 to one; version of RS/6000 expected cost chip-set reduction to $100 per chip for 1,000-up; Motorola to design and manufacture with at worst 0.5 micron features; may be desktop size only

4. Better SAA links 1992-93 for Apple Mac

*InfoCorp Estimates

International Macintosh Machines

IBM has an estimated 14% share by number of the worldwide personal computer, 12% of the workstation market in 1991; combined with Apple’s 9% share of the PC market in 1991, a joint effort could produce 50% of the PC and 30%+ of the workstation market on one object-oriented user environment by 1995, supplied by IMM (International Macintosh Machines). Such a scenario is possible as a result of the IBM-Apple joint venture, but unlikely in our opinion. Nevertheless, the IBM-Apple joint effort is one consortium that must be reckoned with and if it can deliver its promises on schedule will move Unix solidly into the personal computer market, making the RS/6000 number two microprocessor worldwide.

The joint venture can be a success

Table 2 presents our view of the strategic alternatives that can result from this joint venture. We believe alternative one is unlikely because market dominance – 50% market share – cannot spring only from an object-oriented user environment and a fast RISC chip – distribution channel strategies, application availability and migration, installed base size and loyalty, and cost of manufacturing must also be considered in forecasts. The IBM-Apple effort does not address these last four areas. In addition the ‘openness’ stressed in the companies’ press releases precludes either Apple or IBM from maintaining serious differentiation from clone or OSF/1-based system manufacturers for any length of time. We also believe alternative number three is not likely. Although past IBM joint ventures have not fared well, others have worked, most notably IBM-Microsoft from 1981 through to 1989 (until the O/S2 imbroglio). Culture differences between IBM and Apple have been raised as major impediments to a joint venture’s success; yet Apple is really not very different from Microsoft and good managers are what make joint ventures work, not company culture. There are good managers at both IBM and Apple that can make the joint venture work. There is always a risk of schedule lapses in this joint effort (common to every leading manufacturer), and we account for this possibility with our 30% likelihood estimate for alternative three.

Table 2: What’s the Market Impact?

Alternatives:

1. IBM and Apple will dominate the PC and worksta

tion market by gaining market share well beyond their current 23% share (10% likelihood).

2. IBM and Apple will remain strong in personal computers and become strong in Unix workstations; the RS/6000 chip will be number two behind the Intel X86 in unit ships (60% likelihood).

3. Like other IBM joint ventures (IBM-Rolm) this will have little market impact (30% likelihood).

Alternative two is most likely for several reasons:

1. Even if the object-oriented software never materialises, the AIX extensions and Power PC products are important enough to help IBM and Apple keep their strong (but not dominant) positions.

2. The enhanced AIX (really Apple’s A/UX upgraded to OSF/1 or 2 from its current System V.3.2 base) should be a winner since it will have Macintosh application portability as A/UX does. This brings 5,000 or so Mac packages to the Unix user with a Macintosh or Motif interface.

3. The RS/6000 is a strong technological product, early in its learning curve – Motorola is a perfect (maybe the best) choice to create a 0.5 micron feature size (or less) single chip version. The Power PC could well outperform the 1994 vintage X86 offering by a substantial margin.

For these reasons we think the IBM-Apple joint effort will succeed. Some might argue its success will give IBM and Apple dominance on the desktop, but this ignores the competitors each faces, who also can bring formidable products to the market. The best example is Hewlett-Packard Co, which already has the fastest single chip microprocessor CPU in Precision Architecture, which is 30% to 40% faster than IBM’s five-chip RS/6000 today and already has a joint venture with Sun Microsystems Inc on a common object-oriented user environment and offers the forward-compatible NewWave operating system environment on the X86, 680XO and Precision Architecture. Digital Equipment Corp and Compaq Computer Corp are also allied via the Advanced Computing Environment, and in a similar timeframe as IBM-Apple will deliver X86 and MIPS Computer Systems Inc forward-compatible operating environments with Windows NT and Open Desktop. As a result IBM and Apple, whose joint effort we believe has a slight edge in deliverable features, may gain slight market shares over today, but nothing that will drastically change the picture as of 1991.

Table 3: Migrating RS/6000 to Power PC

RS/6000 – 1991 Power PC

CPU chips 5 (2 opt caches) 1 (2 opt caches) Featre Size 0.8 micron 0.5 micron min Clock Rate 20MHz to 40MHz 50MHz to 60MHz Minmum MIPS ’92 ’93 ’94 ’95 Performance 30 40 52 68 90 Transistor 2m logic 3m to 5m Count 3m to 5m memory Pricing $800-$1,500 qty 1,000 $100 qty 1,000