Sun Swat Redux: A Strategic Roadmap [OSNews]
An interesting article by Paul Murphy on Sun's dilemma: it has the best software substrate named Solaris and awesome hardware yet it's losing market share. Sun's power needs some exposition:
The ability to treat a very large amount of memory as a single, symmetrically accessible, data store and an operating system capable of managing it with near perfect reliability for very long periods of time. There is a single wait-state penalty on memory accesses outside the local processor's 4GB base address space, but a Sun 6800 maintains cache coherency across all 24 CPUs and does so at about three times the comparable throughput rate on the maxed out 16 CPU, $5-million IBM z900 mainframe. Getting that to work on the Sun 6800 and its bigger cousins the Starfires is an unbelievable technical feat that's well ahead of the competition.
Neither Linux/x86 nor Windows nor other Unix vendors can touch this. But then what ails Sun?
Many problems. One problem is few people understand and harness this awesome power as the article points out. Hardcore people who appreciate what Sun can deliver are scarce and are disappearing. Also for lots of problems, Sun is overkill and overpriced at least initially. As data get larger and larger, Sun's unique ability to handle large memory looks to be a winner to some people. But most will just pile up lots of cheap x86 boxes.
Another problem is: Sun has done extremely fine job of building Solaris and SPARC but they didn't build on top of these great substrates. They are now trying to catch up on middleware (their new iPlanet is completely rewritten code) and database (by acquisition). By comparing Sun to IBM, it becomes clear how badly Sun is faring these days. Unlike Sun, IBM foresaw correctly high-level software platform (way above typical OS) is key to business and has built WebSphere and upgraded traditional transaction systems like CICS and integration platform like MQ. IBM also leads in web services space (as one half of duopoly with Microsoft) while Sun lags behind in current generation of data-centric web services. To add insult to injury, IBM has professional services that knows about other peoples businesses like no other while Sun doesn't. Finally IBM's R&D dwarfs Sun's.
When Apple came out with Aqua, I joked Scott McNealy should swallow his pride and license it to replace CDE and be done with it. Now Sun ships GNOME, a ho-hum Windows copycat whose only edge is that it's not KDE! (Of course KDE's only winning trait is that it's not GNOME). Another thing Sun could've done is to push its x86 version of Solaris (that provided source level compatibility with Solaris on SPARC) to expand its market and mind share by risking its workstation profit a bit. Sadly this fine OS seems to be in terminal decline with few users. Sun could have stopped erosion of its workstation market share by delivering a usable desktop and applications (something Sun stopped doing when it dropped NeWS and went with X many years ago). Desktop is not something you can give up as it is the first step and "last mile" for your users. Microsoft's making headway into server-side on the strength of its desktop monopoly starting from Outlook/Exchange, IIS and SQLServer. With fewer and fewer users facing Solaris directly, fewer and fewer will choose Solaris in the back office. When your own employees use either Linux or Windows or Mac OS X on their home computers, how can you persuade others to do otherwise?
Copyright 2003 Jay Han
Theme Design by Bryan Bell
