Since the launch of SQL Server Magazine in March 1999, we’ve been tracking SQL Server 7.0’s progress in the market. In particular, we’re looking for significant trends that will affect momentum of SQL Server as a platform for business-critical applications, e-commerce, and knowledge management. Here’s what we found.
The Transaction Processing Council (TPC) has long been the industry standard benchmark for database performance. You can find the "Top 10 TPC-C Results by Price/Performance" at TPC’s Web site, http://www.tpc.org/new_result/ttpp.idc. SQL Server 7.0, running on Windows NT Server 4.0, holds all 10 spots. When factoring price/performance into total cost of ownership (TCO), it's no wonder SQL Server 7.0 comes out a winner.
However, what happens when you factor out costs? If you look at the "Top 10 TPC-C Results by Performance," SQL Server holds none of the top 10 positions, and only position 10 is held by an NT database—Oracle. The top-performing database is Oracle8i, running on a 64-CPU Sun Starfire Enterprise 10000, at a cost of more than $12 million dollars. If money isn’t a concern, you can buy performance. But for most SQL Server Magazine readers, cost is one factor that led them to SQL Server.
In June 1999, Microsoft, Unisys, and Compaq announced the record-setting results of running SAP's sales and distribution benchmark on 8-way Intel Profusion-based servers. The results show a continuing trend of pushing NT/SQL Server into enterprise performance territory once held exclusively by UNIX. Prior to the Intel Profusion-based servers, SQL Server Magazine couldn’t recommend purchasing an 8-way server. It looks as though the Wintel community has made a scalability breakthrough, making 8-way servers feasible and providing the extra capacity enterprise customers need. The bottom line is that SQL Server 7.0 wins the price/performance market on the 2-8 CPU system running on NT.
One of IT’s toughest problems is getting high-quality data into the hands of the business decision-makers who need it. Today, data warehouse/data mart solutions aid querying and reporting tools and allow data optimization for retrieval. Prior to SQL Server 7.0, companies spent upwards of $100,000 for an OLAP engine. Microsoft includes a free OLAP engine with SQL Server 7.0 so mid-market customers can experience the benefits of OLAP. And, with Office 2000 and Excel, Microsoft is giving away basic analytical tools for querying with the OLAP engine. These changes have made Microsoft a player in the business intelligence market and caused Microsoft's partners to reposition their products to add value to the existing free tools. The interest of building business intelligence solutions on the SQL Server Platform is at an all-time high among SQL Server Magazine readers. Interest in supporting SQL Server 7.0's business intelligence platform is also high among software vendors.
Microsoft has integrated SQL Server 7.0 into its development platforms, such as Visual Studio and Visual InterDev. In addition, Microsoft has included a desktop version of SQL Server 7.0 called Microsoft Database Engine (MSDE), with unlimited licenses for users of Microsoft's development and Office tools. For example, when you configure Access 2000, you can choose to use the legacy Jet engine or the new MSDE engine. If you choose MSDE, you ensure your upgrade path is to SQL Server because SQL Server 7.0 is what’s under the hood. Developers can choose MSDE as a Web database engine for small Web sites and scale up to a full version of SQL Server 7.0 as needed.
Knowledge management (KM) is one of the latest buzzwords to hit the IT market. In simple terms, KM integrates structured and unstructured data into a form that business decision-makers can use. In reality, it’s one more step in the querying and reporting evolution. Whereas business intelligence deals with structured data, KM adds the unstructured component, usually by tackling documents and emails. KM created a huge opportunity for Microsoft to position its BackOffice suite as an integrated KM platform. By combining Exchange (unstructured data), SQL Server (structured data), Internet Explorer (user interface), Site Server (the back-end Web engine) and Office 2000 (Outlook and Excel as knowledge portals), Microsoft provided a path for its development community to build KM solutions. If KM takes off, Microsoft might eventually merge the disparate Exchange, SQL Server, and Internet Information Server (IIS) development communities into one cohesive community that develops Web applications the Microsoft way.
Web sites, especially e-commerce sites, are database-driven. If you're an emerging .com business, you require that your site is up 24 x 7. You might have dreams of becoming the next Amazon.com, which means you need a highly scalable system. And you don't want to redesign your Web site because you've maxed out your database server. SQL Server 7.0's fate is directly tied to the Windows NT/2000 server, because this platform is the only one it runs on. Lately, Windows NT has taken a beating on the reliability front. Linux on the low end and UNIX on the high end have put the squeeze on NT and sought to exploit the perceived reliability weakness. Many Internet companies are loaded with venture capital or IPO money and are willing to spend as much as 10 times the price of NT/SQL Server for the availability and scalability features found in a Sun/Oracle8i environment.
Today, SQL Server 7.0 can scale to eight CPUs on a single box. Although SQL Server 7.0 supports failover clustering, it does not support load balancing. So, if your database outgrows a single 8-way box, you're stuck. However, according to major enterprise resource planning (ERP) vendors, such as SAP and Baan, SQL Server 7.0 on NT can handle 90 percent of all customer requirements. Will SQL Server on Windows 2000 be able to handle 100 percent of everyone's requirements? Probably not. However, the ability to address 80 to 90 percent of the market is a huge opportunity for Microsoft and its partners. Partners who guarantee reliability or provide tools to increase reliability have a huge opportunity in today's market.
|"Windows 2000 is the largest investment for Microsoft ever. Microsoft is literally 'betting the company' on Windows 2000."
Microsoft plans to spend more than $100 million on marketing the launch of Windows 2000. Microsoft knows that the success or failure of all of its BackOffice products, including SQL Server 7.0, will ride on the success of Windows 2000. If Windows 2000 is perceived as being more reliable and scalable than Windows NT 4.0, Windows 2000 will enjoy success. If that happens, BackOffice products will get an immediate lift in the market because of the integration of BackOffice with Windows 2000. For example, the next version of Exchange, code-named Platinum, has shed its own directory database in favor of Windows 2000's Active Directory. In other words, you can't run the next version of Exchange without having a good installation of Windows 2000. Exchange is fully integrated with Windows 2000. Likewise, you can expect SQL Server to integrate even further with Windows 2000, providing the basis for an Internet-based file system over time. If integration is important, BackOffice has an edge over competing products.
Microsoft is being challenged on all fronts—from Oracle, Linux, DOS, UNIX, and even the Web itself. The Web has changed the game on several fronts. First, developers want to develop native Web applications. This endeavor challenges the domination of Win32 applications on the server and desktop. Second, application service providers (ASPs) are providing an alternative to building your own applications. ASPs can host applications on the Web for IT organizations that can’t build, manage, or keep up with the increasing demand from end-user communities or external customers. In both cases, Web-based applications will become the norm and will require extreme reliability, availability, and security. Microsoft can no longer slowly ease its way into the enterprise. These server-based Web applications have requirements equal to or greater than enterprise-level customers who want these applications now. Enterprise-level customers seem willing and can pay top dollar to build a platform that will get them into the new e-everything world.
This server-centric application platform presents an incredible opportunity for Microsoft and SQL Server to capture a large share of the new market. Microsoft already has the components it needs to build the next generation of applications. However, Microsoft finds itself in the position of protecting its legacy desktop-centric business model instead of being the aggressor in this major shift in computing. Microsoft used to laugh at IBM for being slow and protecting its aging mainframe-centric model. Now that the tables are turning, Microsoft is in the hot seat. In the Internet world, you either lead, follow, or get out of the way. If Microsoft decides to lead, SQL Server can help its customers get to the new server-centric computing model.