| By David Skok | Article Rating: |
|
| March 1, 1999 12:00 AM EST | Reads: |
9,155 |
The application server category is one of the more confusing markets to understand. In addition, the market changed rapidly during 1998, with a number of companies being acquired. This article aims to clarify the situation.
Software: A History of Consolidating Services on New Platforms
Looking back at the major shifts that have taken place in the software industry, a trend emerges that is a helpful indicator of what is likely to happen in the evolution of the application server market. As new platforms emerge, there are generally a number of players providing point products that address a specific niche area. As the market matures, leaders emerge who tend to incorporate most of the point products' functionality into their primary product. An example of this is the relational database area where data warehousing, object databases, full text retrieval, transaction processing and bitmapped indexing have been consolidated into market-leading products such as Oracle8. In another example, we see SAP take a leadership position by integrating functionality that was previously available only in multiple point products. On the client side we see Microsoft integrating all of the standard office products into a single suite, and in the process eliminating players that have led a single product category. I predict that a similar consolidation of services is about to take place in the middle tier.
The Next Consolidation: The Middle Tier Becomes Essential, Driven by Thin Client
The driving force for the shift to the next computing platform is clearly thin client and Web-based computing. Organizations frequently need to deploy e-commerce applications that are both consumer and business-to-business focused. Furthermore, there is strong pressure to provide ubiquitous access to applications at a low cost.
Supporting Web and thin client applications automatically implies a shift toward placing most of the processing in the middle tier. The Middle Tier Before the Web
Prior to Web-based and thin client computing, we saw a number of point product areas in the middle tier:
These products didn't provide facilities to support the generation of a user interface or what we refer to as "Presentation Services."
Web Application Servers
Starting around 1994, the major drive to create dynamic interactive Web applications spawned a new kind of middle-tier product: the Web application server. The first generation of this kind of product was simply a Web server running CGI scripts. The second generation focused on making it easier to develop and deploy dynamically generated HTML pages, taking over where CGI scripts left off. In this generation the primary purpose of these products was to provide presentation services (see Figure 1). A major limitation of this architecture was that business logic was being embedded in scripts inside Web pages.
The Next Step
As Web application servers evolved, they added important facilities to increase their usefulness (see Figure 2).
Well-designed servers support Java clients over secure HTTP and provide a three-tier data architecture (queries executed at the server, with result sets remotely accessible via objects at the client). Facilities should be provided for automatic distribution and updating of the client-side application for applications deployed outside the browser.
It's worth noting that EJB is a long-term vision. In version 1.0 the specification is immature and leaves many key areas undefined. Most players expect that the standard won't reach the level of maturity that will make it widely usable until version 2.0.
Web Application Servers Meet Distributed Objects
So far we've described two separate markets, the first consisting of object, transaction and connectivity services with no presentation services; the second, a market driven primarily by the need to deliver thin-client presentation support.
In mid-1998 these two worlds collided, resulting in major market confusion (see Figure 3). Leading up to the collision was increasing industry recognition that middle-tier distributed objects provided the right architecture for the creation of larger-scale sophisticated applications. Sun's Enterprise JavaBeans (EJB) specification was the catalyst for the collision. It provided a standard specification for how these middle-tier objects should be written, and made them far easier for the developer to write (the hard work would be done by the EJB container and server vendors).
A Market Thrown into Confusion
Both the application server vendors and the vendors of object and transaction services realized that they needed to support the EJB standard; at this time a variety of different products changed their name to application server. All of a sudden there were 40 application server vendors, each offering widely different types of functionality. Some were focused on Web applications, offering strong presentation services; many others were pure object servers with no presentation services at all.
What Does It Take to Be an Application Server?
As a result of the two markets converging, it is clear that we need a new definition of what is required to be an application server. Are presentation services important? Are connectivity services a requirement? Should the server come with integrated development tools?
The following section outlines where I believe the market is headed and explains the different component parts.
Where the Market Is Headed
I believe that customers are looking for an application server that combines all the services shown in Figure 4.
The rationale for each of the components follows:
Where the Players Are Today
Figure 5 shows three different types of players. Examples of pure presentation servers are Allaire's ColdFusion and Microsoft's Active Server Pages (ASP). Examples of pure object servers are products like Iona, Oracle and WebLogic. Few products combine both presentation and object services in an integrated package. SilverStream and NetDynamics are examples.
It should be pointed out that no vendor has all the attributes required today, particularly as it will most likely require version 2.0 of the EJB spec to achieve this, and that is not yet available. Clearly, the right-hand group shown in Figure 5 is nearest to where the market is headed.
The Importance of Neutrality
Application servers are expected to integrate with the wide variety of existing architectures and standards that exist in today's typical customer environment. These include:
Based on the foregoing, customers are looking for application server vendors that are neutral and don't have a particular bias toward an operating system, data source, etc. Several of the application server vendors don't meet this requirement and consequently don't provide the necessary level of support for other competing products. This is generally considered a major drawback.
Summary
This market has undergone substantial change in the last 12 months. It is still maturing and will probably take another 12 to 18 months to sort itself out. Most companies will be evaluating and deciding on their standards during 1999 and 2000 as this is a time-critical technology that must be used now to remain competitive. Although a single winner has not yet emerged, the leaders are clearly separating from the pack based on vision, technology, ability to execute, support services and strength of installed base. A clear requirement is that the product must integrate all of the above services, in particular offering strong presentation services combined with strong distributed object and transaction services.
Published March 1, 1999 Reads 9,155
Copyright © 1999 SYS-CON Media, Inc. — All Rights Reserved.
Syndicated stories and blog feeds, all rights reserved by the author.
More Stories By David Skok
David Skok joined Matrix Partners as a General Partner in May 2001. He has a wealth of experience running companies. He started his first company in 1977 at age 22. Since then he has founded a total of four separate companies and performed one turn-around. Three of these companies went public.
Skok joined Matrix from SilverStream Software, which he founded in June 1996. Prior to its July 2002 acquisition by Novell, SilverStream was a public company that had reached a revenue run rate in excess of $100M, with approximately 800 employees and offices in more than 20 countries around the world. His work as a value added investor is best known for helping JBoss take its Open Source business to a successful exit with its sale to Red Hat, and for helping AppIQ, Tabblo and Diligent Technologies, which have all had successful exits, from their inceptions to their acquisitions by HP and IBM.
He serves on the boards of Digium (makers of the very popular Asterisk Open Source PBX/telephony software), CloudSwitch, Enservio, OpenSpan, Solidworks, VideoIQ, and HubSpot. In addition to his broad focus on enterprise software, he is specifically focused on the areas of cloud computing, Open Source, Software as a Service (SaaS), marketing automation, virtualization, storage, and data center automation.
- Three Tips to Successfully Load Test Adobe Flex Applications
- Patterns for Building High Performance Applications
- It's the Java vs. C++ Shootout Revisited!
- Asynchronous Logging Using Spring
- Graal, a Dynamic Java Compiler in the Works
- Cross-Platform Mobile Website Development – a Tool Comparison
- Write Once Run Anywhere or Cross Platform Mobile Development Tools
- Three Buzzwords That Every CIO Hears but One They Should Listen To
- Cloud Expo New York: The Java EE 7 Platform - Developing for the Cloud
- Immersing into JavaScript Frameworks
- Workday Reportedly Prepping to Go Public
- Book Review: Sams Teach Yourself Java in 24 Hours
- Three Tips to Successfully Load Test Adobe Flex Applications
- Book Excerpt: Introducing HTML5
- Five Years Waiting for JRE 7: Is It Justified? (Part 1)
- Book Excerpt: Java Application Profiling Tips and Tricks
- i-Technology in 2012: Five Industry Predictions
- Patterns for Building High Performance Applications
- It's the Java vs. C++ Shootout Revisited!
- OpenXava 4.3: Rapid Java Web Development
- The Next Web Architecture
- Asynchronous Logging Using Spring
- Java for Programmers (2nd Edition)
- Graal, a Dynamic Java Compiler in the Works
- A Cup of AJAX? Nay, Just Regular Java Please
- Java Developer's Journal Exclusive: 2006 "JDJ Editors' Choice" Awards
- JavaServer Faces (JSF) vs Struts
- The i-Technology Right Stuff
- Rich Internet Applications with Adobe Flex 2 and Java
- Java vs C++ "Shootout" Revisited
- Bean-Managed Persistence Using a Proxy List
- Reporting Made Easy with JasperReports and Hibernate
- Creating a Pet Store Application with JavaServer Faces, Spring, and Hibernate
- Why Do 'Cool Kids' Choose Ruby or PHP to Build Websites Instead of Java?
- What's New in Eclipse?
- i-Technology Predictions for 2007: Where's It All Headed?


















