ZapThink 2004 SOA Retrospective and Looking Ahead to 2005

Well, it is that time of the year again — when thoughts turn from the tactical to the strategic as companies of all sizes plan their upcoming year and make resolutions anew. With 2004 in our rear-view mirror, what have we learned from the past, and what new resolutions can companies fulfill in the year ahead? What will this year, devoid of distractions of the political race, Olympics, and potentially in mid-economic recovery mean for companies investing in Service Orientation and the Web Services technologies that support them? What will newfound mergers and partnerships, continued consolidation, and maturation of different technologies mean for the market? In this upcoming year, what will companies come to expect from the providers of solutions, and what will those providers expect in return? And finally, what can you all expect from ZapThink as we continue to dig deeper into the widespread changes that Service Orientation is increasingly requiring of the enterprise?

How did we do in 2004?
Last year at about this time, ZapThink made a number of evaluations of the previous year and prognostications about what was going to happen in these past 12 months. How correct was our analysis, and what confidence can we give in of the accuracy of our future predictions?

Our first prediction was that companies would move from mostly implementing SOA pilot projects to implementing and rolling out their first departmental projects. We were right on the money on this one. Companies have been aggressively implementing and scaling up their SOA initiatives across the board. A number of large enterprises in a wide range of vertical industries, including financial services, insurance, aerospace, government, healthcare, pharmaceutical, retail, and manufacturing, have taken their limited Web Services pilots and turned them into full-fledged SOA implementations. These companies are now looking to spread their initiatives through more cross-departmental Service usage as well as some of the first truly enterprise-wide shared Services. As such, we predict that 2005 will see the blooming of a thousand SOA flowers — the implementation of SOA initiatives that will become strategic to the business as a whole, rather than tactical solutions to departmental needs. Yes, SOA will finally “cross the chasm” and become an approach that the majority of larger companies must implement, rather than an approach implemented by the few and far-sighted.

Another of ZapThink’s predictions last year was that the enterprise architecture team would emerge in earnest as companies realized the importance of having a mediated role between the developer and the business user in the SOA context. While our predictions have borne out mostly in that many enterprises do indeed now have specific roles with that title and/or responsibility, enterprise architecture is still a black art for many firms, and we can expect the roles, responsibilities, and growth of demand for this skill-set to continue to emerge over 2005.

Most importantly, we predicted that 2004 would be a do-or-die year for vendors, and that 2004 would be “the year of the SOA.” On both counts we were 100% right. Many startup vendors either ran out of gas, merged with other firms, or made significant adjustments to their strategy. Some vendors finally dropped their EAI messaging and adopted SOA strategies in earnest, whether or not their products actually reflected that strategic mental shift. The large, incumbent vendors solidified their offerings through new development and acquisition, and dramatically expanded their professional services capabilities in the space, leading to the belief that 2005 will be the “money year” for SOA consulting services. Finally, startups got into the acquisition or partnership game themselves, with a number of notable hook-ups resulting in a smaller base of vendors pursuing an ever-increasing market. 2005 will continue to see significant consolidation in the space as the market settles on its preferred approach for implementing, managing, securing, designing, and deploying SOA in the enterprise.

ZapThink, therefore, is batting 1000 on its SOA predictions. Here are our predictions for 2005 — hopefully as insightful as the last bunch!

Prediction #1: The WS-I Gets Going or Gets Out
In the past year, the WS-I emerged from their rough, politically-driven start and produced a number of significant deliverables, including the Basic Profile and self-compliance suites. However, the group has yet to deliver on the necessary profiles for security, process, and reliability — three significant minefields full of specifications that lack cohesive vendor support and have widespread disagreement. In order for the WS-I to prove their relevance, the group must not only produce profiles that cover those areas, but must also garner enough respect from vendors and implementers to make their word law in the SOA nation.

Basically, the group, mostly vendor-driven, still doesn’t have enough respect and pull in order to enforce the fundamental notion of interoperability for Web Services. In 2005, the group will get even more serious, put out profiles that solve the critical interoperability problems, and hopefully with it, any fires among the vendors that choose to continue to produce specs that focus mostly on competitive differentiation rather than industry-wide interoperability. In 2005, the WS-I will get serious or will fade from relevance.

Prediction #2: Companies will Make More Money on SOA Services than on SOA Products
Many have said that the only people really making money on SOA are the analysts, media, and consultants. In many ways that was true of the early years, but increasingly there is a real market for both products as well as implementation services for SOA. More importantly, ZapThink is seeing rapid growth in the market for real enterprise architecture support and guidance in building the right services that enable loosely-coupled, coarse-grained, event-driven SOA.

The past nine months or so have seen a dramatic increase in the number of professional service firms that have either opened a new division solely focused on SOA or significantly expanded an existing practice.

ZapThink has also seen some new Professional Service Organization (PSO) startups that have gone from just a concept and no revenue at the beginning of the year and are already bagging multi-million dollar contracts. In fact, it seems that in 2005, PSO revenue might exceed and even dwarf that of all the SOA vendors in the market today.

Prediction #3: ESB: Use it or lose it
One of the more troubling trends of 2004 is the emergence and rapid market confusion around the term “Enterprise Service Bus.” Once a vaguely-defined term that nevertheless carried with it some specific functionality requirements and implementation implications, the ESB term now refers to a hodgepodge of mostly unrelated products, disparate feature and function sets, and confused vendor product messaging that seems to be more a definition of chaos than a specific functionality set. Some ESB vendors include a message bus infrastructure, while others don’t. Some ESB vendors provide business-process modeling and runtime capabilities, while others simply provide an event-driven runtime infrastructure. More gravely, some vendors are distancing themselves from the idea that ESB is even a product category — but rather an architectural framework for SOA. As one vendor recently told us, “If SOA is the will, ESB is the way”.

As a result, ESB has now come to mean “any technology that is required to implement an SOA.” As such, without a more concrete definition of the capabilities and functionality that an ESB provides, the term will either become meaningless, or will be given a more narrowly defined and concise definition. However, don’t hold your breath — vendors and other analyst firms are only increasing the noise pollution over this term by broadening its definition, watering down its promises, and making it utterly useless to the enterprise end-user that has no idea what to expect from their ESB vendors. So, in 2005, the ESB term will either shift to a realistic, understandable, and meaningful meaning, or disappear from the SOA lexicon forever.

ZapThink Take
In addition to all the above predictions, there are a number of key trends that bear watching through 2005.

The rich client is beginning to emerge as a viable approach for consuming and composing Services that users can interact with in lieu of more staid, tightly-coupled approaches such as web-based portals.

While enterprises will replace very few web-based portals with rich client offerings in 2005, they will no doubt find that the ability to consume Services in a stateful, sometimes connected, and event-driven manner is more appealing than the stateless, poor user interaction capabilities of today’s web-based browsers.

In addition, metadata management and SOA governance will finally come into their own in 2005. Companies will realize that achieving the benefits of reuse and dynamic composability of Services is a significant challenge without a means to manage the Service contracts and policies that govern their reuse. Metadata management, which was given short shrift and little attention in 2004, will be the star of the ball in 2005 — giving enterprises the real means for making SOA work in production. In addition, companies will realize that their SOA initiatives will stop short of the strategic if they can’t simultaneously address their governance issues. While 2004 was the year that SOA governance first emerged as a topic of conversation, 2005 will bear out as a year of implementation for SOA governance-related issues.

While ZapThink doesn’t have a crystal ball or any other mystical means for divining the future, we do see the writing on the wall. Our conversations with literally hundreds of end-user, professional service, and vendor customers as well as a deep understanding of the technological and cultural shifts that SOA is imposing on the market really leads to only one set of conclusions. The real question is how long will it take the market to realize and implement those required changes. On that note, we can only hope that politics, economics, and yes, even sports, don’t distract the enterprise yet again in 2005.