SQL Server Needs Revision

The aging database is dragging down other Microsoft servers.

From Windows server 2003 to exchange Server 2003 to SharePoint 2.0 to BizTalk Server 2004, some of the best technology coming out of Microsoft lately has been on the server side. But all of these great server products have one problem: They are standing on an aging, creaky and often damaged base—Microsoft SQL Server 2000.

As the name implies, SQL Server 2000 was released in the year 2000, along with Microsofts Windows 2000 and Office 2000. New versions of SQL Servers competitors were also released at that time, but they—and pretty much every other product released in 2000—have been upgraded in the four years since. Every product, that is, except SQL Server 2000.

Theres been a lot of complaining lately, in the pages of eWEEK and elsewhere, about the extended wait for Microsofts Longhorn, code name for the next full upgrade to Windows XP. But, compared with the wait for the next SQL Server upgrade, the wait for Longhorn wont be long at all.

Its true that big products such as databases tend to move in much slower product cycles than things such as office suites. But four years is a long time, no matter how you cut it. Five years is even longer, and itll be at least that long between SQL Server 2000 and the new release of the database, currently code-named Yukon. This could spell big problems for any company basing its infrastructure on Microsoft servers.

Think of it this way. Imagine you were going to build a PC system. You decide to build a killer rig, and you get the best motherboard, a big-and-fast hard drive, lots of memory, a top-of-the-line graphics card, even a DVD burner and LCD screen. But in this best-of-breed system, you install a circa-2000 Pentium III processor.

Thats what the current Microsoft server platform is like—lots of brand-new, high-performance products running on an old engine.

Some companies may decide they can no longer wait for that engine to be upgraded.

In fact, this may already be happening.

As part of my job, I talk to lots of people in IT departments. I always ask for a list of IT products in use, and many of these companies are predominantly Microsoft shops.

However, looking at my notes recently, I noticed a trend in these shops: a surprising absence of SQL Server in key spots.

Sure, these shops have some SQL Server installations, but usually only in cases where another Microsoft application requires it. Many times, these companies were using another product as their key enterprise database, often Oracle or DB2.

Microsoft recently announced that the next version of SQL Server wont ship until early next year. Even if the release date doesnt slip again (and would you really want to bet on that?), many companies wont be deploying the new SQL Server until 2006, especially if they follow the common practice of waiting for a products first service pack before full deployment.

Im not in favor of rushing products to market. In fact, I think vendors and customers are better off with a late product than with a bug-ridden product that hits its release mark. Customers should not be involuntary beta testers.

The feature set of the forthcoming SQL Server, which is expected to be named SQL Server 2005, looks compelling, especially for developers of next-generation business applications.

But five years is a bit much. I think I speak for a lot of SQL Server users when I say that a SQL Server 2003 release would have been a good thing. It didnt have to be anything major, just something that addressed security problems more effectively than a service pack and that made SQL Server better equipped to deal with the current reality of Web services and XML.

Instead, we wait. Or not. Because when it comes to a core enterprise database system, living in the past is not an option.

Labs Director Jim Rapoza can be reached at jim_rapoza@ziffdavis.com.