Is Monterey Unix's revenge?
It could be shaping up to be the real threat to Win2k
Stealth marketing is evidently being used for an alternative to Win2k that did not even get a mention in the oral evidence at the Microsoft trial. It is uncertain if Microsoft wanted no mention of it, or whether Microsoft's lawyers were unable to understand that the real threat could be Project Monterey. This industry standard Unix operating system initiative for the IA-32, IA-64 and IBM Power PC processors already has a lot of weight behind it, and could prove to be Unix's revenge. The OEMs giving support are Acer, Bull, CETIA (a Thomson-CSF subsidiary), Compaq, IBM Netfinity, ICL, Motorola, Samsung, Sequent, and Unisys. There is support from 33 ISVs, including Baan, BEA, Computer Associates Compuware, DataPro, Informix, Infospace, Micro Focus, Netscape, Novell, PeopleSoft, Pick, Progress, Rational, Real World, Risk Management, Software AG, SAS, and Take Five. Cygnus Solutions is developing, in conjunction with IBM, GNUPro tools for AIX and Project Monterey on IA-64. SCO, as the volume market leader for Unix with UnixWare7 is collaborating with IBM to enhance IA-32, and to co-develop for the IA-64 market. Meanwhile IA-64 Linux, Trillian, development is being led by VA Linux Research with support from Cygnus, HP, IBM, Intel, and SGI. Within the last month or so, both 64-bit W2K and Monterey/64 were demonstrated on Merced without a software emulator, so all things being equal they may well be released around the same time. Sun has only got as far as demonstrating Solaris with an emulator on the IA-64. Rajiv Samant, IBM's Unix brand general manager claimed that Monterey/64 is ahead of HP-UX for the same reason, although Intel says it is serious about its support of HP-UX. IBM currently offers three Unix systems, AIX, SCO, and Linux, and the company's vision is of a triangle with Linux at the base, running on Netfinity servers and mostly used for file and print, mail and collaboration. The second layer is AIX + RS/6000 (or UnixWare + Netfinity) being used typically for transactions such as retail and branch banking. The top layer is AIX + RS/6000 SP for ERP transactions and e-commerce. The boundaries between the three layers are not intended to be solid, but in general, transaction value is likely to increase up the triangle. IBM has recently clarified how it will manage Sequent, a leader with high-end Intel-based Unix, which it bought for $810 million. As the developer of DYNIX/ptx, its jewels are the NUMA-Q 1000 and 2000 servers and the NUMACenter, which IBM and its partners will start selling immediately. On the development side, NUMA technology should help IBM to offer larger SMP processors as synergies are identified. NUMA works with Intel and Power chips. Before the merger, IBM and Sequent were collaborating in Project Monterey, so both companies are aligned. The NUMA-Q team will remain intact, but the Sequent name looks like being subsumed in the IBM brand. Sequent has a development team in Eau Claire, Wisconsin that will continue its collaboration with Microsoft. IBM plans that the Linux application execution environment will provide most Linux APIs and ABIs for AIX 4.3.3 + Monterey/64. Linux applications, which IBM says can be readily recompiled for high performance on AIX/Monterey (or can run as binaries without change) will go some way to providing an application portfolio (for which read "overcome the applications barrier"). The target date is the first half of next year. IBM will offer the Red Hat Linux distribution for its ThinkPad 600E, and plans to offer the Caldera, SuSe and TurboLinux on other ThinkPads. It also has a Developer Kit for Linux in alpha, with Java 1.1.8 and a preview version of VisualAge Java for Linux. There were around 50,000 downloads of IBM's DB2 for Linux in the first nine months. There seems to be quiet confidence, at least in the IBM camp, that only Monterey/64 will give high performance on Merced from day one. However, it will effectively be a new operating system, with the probability of an extra portion of bugs initially, so take-up is likely to be somewhat slow at first, until the reports of early users are positive. Microsoft may well out-market and out-ship Unix in terms of the number of systems running Win2k, but in terms of sales revenue, Unix may well continue to hold its own or even accelerate. The big opportunity will arise if W2K is found to be significantly unreliable. Linux is being positioned by IBM as a worthy junior partner in Monterey, and not as a serious competitor in the same way as Win2k. For price-sensitive organisations without too many high-value transactions (like the oil company Amerada Hess that reckons it has saved $3 million by moving from Unix to Linux), Linux will prove attractive as it works well with less-expensive servers. Using Linux on the desktop does not give such dramatic financial savings, and such users are mostly the Microsoft-averse. Despite the chumminess of the Unix camp towards Linux, Unix still looks somewhat divided while Sun and HP keep remote from the Monterey/64 effort. Some cosmetic collaboration could make a considerable difference to the perception of the nearly-unified Unix, but perhaps that is too much to expect. ®