Sun’s bold strokes

I have been a pointed critic of Sun Microsystems for some time, at one point comparing it to Digital Equipment Corp., which rode its proprietary strategy into the ground in the early 90s in the face of overwhelming evidence that it was a wrong-headed approach.

But I have to admit some admiration for Sun’s recent moves to reinvent itself in the data center. I was in San Francisco this week for Oracle Open World and had a chance to hear Sun’s Scott McNealy outline the company’s comeback strategy. I was impressed.

While Sun has taken steps to make its UltraSparc technology more competitive, I was more intrigued by its intentions to put Solaris into the open source domain. This was a huge cultural hairball for Sun to swallow. Sun has maintained for years that Solaris was so superior to Linux that it justified the huge premium it commanded in the market. But users have increasingly had trouble buying that story. For the mass market, Linux worked just fine.

Sun has finally accepted the reality that Solaris was not going to win the battle against Linux in any but the uppermost reaches of the Unix market. This insures that Linux will have a potent high-end competitor for a long time to come. For Sun, the challenge is to insure that there’s a reason to buy Sun boxes to run Solaris instead of commodity hardware. That’s an easy argument to make right now, while Solaris is still mainly Sun code. It could be a tougher case a couple of years from now.

But that’s a battle for the future. Sun’s currnet bet is that an open-sourced Solaris will gain enough adherents that the revenue Sun can make from selling hardware to those people will exceed the revenue it would have made selling to a smaller and smaller captive Solaris base. I think it’s a good bet.

Linux needs a spoiler and open source Solaris can fill that role. No one seriously argues that Solaris isn’t a superior Unix. Sun has specialized in high availability, industrial-grade applications for years. The question was whether Solaris deserved the price premium it commanded. Increasingly, it didn’t. By open-sourcing Solaris, Sun is putting a potent Linux competitor into the market. That’s good for Linux and for users. It’s probably bad for Red Hat, Novell and anyone who has cast its lot with Linux.

I don’t see Solaris becoming a mainstream Linux alternative any time soon but for enteprises and those who demand enterprise-class reliability, open source Solaris will be an exciting alternative. If Sun follows through on its commitment to keep Solaris open source, then it will have introduced an exciting new alternative to the market.

The question now is what Red Hat and Novell should do. Both have cast their lots with Linux. But now they have a robust, industrial-grade alternative Unix that could create a profitable revenue stream. Do they stay loyal to Linux or become Solaris adherents, too? It’s an interesting problem…

Service or subversion?

TechTarget editors had an interesting debate last week over whether to publish information that could potentially cause harm in the hands of a malicious or reckless user but which could also do good for people who know how to use it.

It started with a tip submitted by Don Burleson, a respected and oft-published Oracle technical expert and a member of SearchOracle.com’s Ask the Experts team. Don wrote about undocumented features in Oracle that permit a user to manipulate memory to achieve significant performance gains. This technique could save time and money for users who can’t afford new servers or who don’t have the time to optimize their databases in other ways.

But there’s a catch. If applied inappropriately, ths technique can corrupt a database and cause data to be damaged or lost. Don was very up front about that and the editors on our SearchOracle.com site posted a prominent disclaimer at the front of the tip.

Some people thought that wasn’t enough. Tom Kyte, another respected Oracle expert, took issue with Don’s suggestions on his blog. He further suggested, disclaimer or not, it was reckless and dangerous for Don and for SearchOracle.com to post advice that could potentially corrupt data. Responses to the postings on Tom’s blog largely agreed with his position.

Other experts we polled were split down the middle, some thinking the tip was a valuable service to the Oracle community, others saying we were tossing a time bomb into a crowd. What’s the right thing to do?

In the end, the editors decided to keep the tip on the site while somewhat strengthening the language of the disclaimer. I agreed with this decision. Although there are no cut-and-dried answers on what is right in a situation like this, these are the factors I would consider:

• Is the information correct? No question of that in this case. No one disputed the accuracy of the tip
• Is the information useful? If it isn’t useful don’t publish it. I don’t think anyone argued that this advice wasn’t useful to some people. The debate was whether the potential harm outweighed the potential value.
• Is the source credible? There’s no question that both Don and Tom know what they’re talking about.
• Does the potential for misuse outweigh the value of appropriate use? The decision largely hinges on this question. In my opinion, disclaimers should significantly mitigate any potential damage.

On most points, then, the decision to publish the information was obvious. The language of the disclaimer was the only major issue in my mind and I believe the wording that the editors used conveyed the risk appropriately. Basically, anyone who was motivated and interested enough to employ this advice would read he disclaimer and be aware of the risks.

This is not the same as, for example, publishing an Oracle security exploit. In that case, there is little value to the user and great potential for damage. Nor do I believe should media organizations every post advice from anonymous sources unless the content is vetted thoroughly for accuracy. But when respected experts put forth advice that is useful to even a minority of the community they serve – even if there’s risk – it’s the responsibility of independent media to seriously consider publishing it. What happened last week was a debate but not a disservice.