Too late for MacOS X Server?
AnalysisApple may have missed its February ship date for MacOS X Server 1.0, but for once a skipped deadline isn't the issue. True, many Apple-watching Web sites have expressed their concern over how the delay might damage the credibility of both Apple and its new operating system, but even a month's delay is as nothing compared to the set-backs Microsoft has frequently had to make to its OS ship dates. No, the real question mark hovering over MacOS X Server is whether it's worth releasing at all. It's not that the product is flawed technologically -- it isn't -- but in the face of that genuine computing phenomenon of the 90s, Linux, will it ever be anything more than a niche product found only in schools and publishing environments? Two or three years ago, about the time that Apple bought NeXT, it might well have gone on to do great things. Novell Netware and Windows NT were battling it out for medium to large-sized enterprise market, while Unix was only a valid option at the real high end. That left a huge opportunity to carve out a business at the bottom of the pile. Small business' networks and the burgeoning demand for Web servers, would have ensured a solid, cheap and, most important of all, non-Microsoft server OS could quickly build up a lot of user support. It didn't happen, of course, and instead almost everyone and their dog began to use Linux. For anyone who's been living under or a rock or in Mac isolation for the last 18 months, Linux is a version of Unix, originally written for Intel-based machines but now running on pretty much anything, including Macs. The Unix angle makes Linux both highly stable -- it takes a lot to knock over a machine running the OS -- and highly powerful. But that's not really what makes Linux interesting. The important point to bear in mind is that it's free -- all the code is available on the Internet for anyone to download, modify and use. It's an approach called 'open source' and hardcore programmers love it. They can build their own operating system and tailor it for whatever hardware they want to run it on. The DIY approach to Linux is very technically demanding, so various companies emerged to sell ready-to-run versions of the OS for a token fee. Freed from the need to compile and install the OS and applications manually, users have turned to these cut-price (few pay more than $50 for a full server OS) Linux 'distributions' in huge numbers, many of them administrators of networks both big and small, and Web server operators. Market research company IDC reported late last year that throughout 1998, Linux's share of the server operating system market grew a massive 212%. That left Linux with just 17.2% of the market, up from 6.8% in 1997. Windows NT lead the market with a 36% share, the same figure it achieved last year. NetWare dropped slightly, from 26.4% in 1997 to 24.1% in 1998. Other varieties of Unix had a combined share of 17.4%. In its survey, IDC only tracked sales of Linux distributions. The figures didn't include copies of the OS downloaded from the Internet simply because it was impossible to tell how many downloads had taken place and how many had been used for servers. That means Linux probably has an even larger marketshare than the 17.2% IDC quoted. That has meant it has attracted the support of some big-name software developers eager to cash-in on such a vast market and, in some cases, as a political move against Microsoft (step forward Microsoft-basher-in-chief and Apple board member, Larry Ellison of Oracle). Last week at the LinuxWorld Expo show in San Jose, IBM announced it would be offering technical support services to corporates running any of the major Linux distributions. It also said it will be bundling the version of Linux offered by the OS' best-known distributor, Red Hat, on various server and desktop offerings and that it intends to develop its own version for its PowerPC-based servers. Dell is already offering Linux to high-volume customers and will soon be offering it as an option for everyone else. So is Compaq. Netscape is porting all of its server software to the OS. You can already run Oracle 8i and Corel WordPerfect 8 on Linux. Heck, even The Register's server is powered by Linux. The point is, Linux is massive, and likely to get even bigger. So far most industry observers have focussed on the threat Linux poses to Microsoft and Windows 2000 (aka NT 5.0) in particular. And you could argue that if Linux is going to give a company like Microsoft a headache, its effect on a much smaller player like MacOS X Server would be fatal. Maybe -- and maybe not. Linux has a pretty major flaw: while it's great for servers, it's not too hot as a desktop OS. Even the latest Linux GUIs, like Gnome and KDE, look decidedly crude compared to Windows, let alone the MacOS. Even as a server OS, Linux can be tricky to set-up -- doubly so if you're not a techie. And there's a dearth of productivity applications. Developer Applix recently released a version of its Office-style suite, ApplixWare, for LinuxPPC, the leading Mac Linux distribution, but it's not an application you can just insert the CD and double-click the install icon. Linux is a work in progress, so all these issues are being addressed and over time solutions will emerge. In the short term, though, there remains an opportunity for an easy-to-use OS to make its mark. MacOS X Server is essentially BSD Unix -- another open source OS -- with a MacOS front-end plus the Yellow Box application APIs and Blue Box MacOS compatibility module. By the time MacOS X Client is done, Server will probably have been upgraded to match its stability and power. MacOS X in all its forms is a very modular operating system, and it wouldn't be too difficult to port its interface and APIs onto another OS kernel -- the core of the system that provides its most fundamental features. MacOS X's high level of portability comes courtesy of the work NeXT did to make OpenStep easy to move to different platforms. Apple has played down this aspect of the new OS, largely because it's keen to continue selling hardware, which is where the money is. However, hints have been emerging from Apple for some time that it's considering building MacOS X onto other kernels than the modified Mach 3 that it currently uses. For now it may be far-fetched to suggest a version of MacOS X that sits on top of the Linux kernel is on the cards, but if MacOS X Server fails to find an audience beyond schools who want to run a stack of net-booting iMacs, Apple at least has the opportunity to move with the rest of the industry and hop onto the Linux bandwagon. Linux need not be the death of MacOS X Server -- it could be the very thing to ensure its survival. ®