SOA after the hype

After more than a year of hype, SOA seems to be moving from concept to reality, but that reality appears to be a mixed bag of pluses and minuses, according to thought leaders.

By now most software vendors who could possibly justify fitting their products into the service-oriented architecture space have become at least buzz word compliant with SOA. The initial hype phase, characterized by a certain "irrational exuberance" and everybody rushing to offer an enterprise service bus (ESB) has passed and developers and architects are actually implementing SOA. To get an idea of where SOA is after the hype phase,...

we asked a group of analysts and thought leaders where they see it now.

Without solid architecture and governance in place SOA is basically a waste of time.
James Governor
Principal AnalystRedmonk

Riding the hype cycle: Gartner's view
Roy Schulte is a vice president and Distinguished Analyst at Gartner Research

The vendor hype favoring SOA is still pretty high, but it is now being countered by a significant amount of developer and management disillusionment. The real change is that companies are now using SOA in production–they are realizing first hand that it is not a panacea and that its initial costs are relatively high. SOA applications are composed of many "moving parts" so they are more complex than monolithic applications. Companies using SOA face significant challenges in governance, testing, configuration, version control, metadata management, service-level monitoring, security, interoperability and other disciplines. SOA is not the cause of these problems, it is part of the solution to them. The problems are inherent in all distributed applications. The main benefit of SOA is the shorter time-to-change. The implementation of the first SOA application in a given business area will generally take as long or longer than it would to build the same application using a monolithic, non-SOA design style. However, subsequent SOA applications and changes to the initial SOA application are generally faster and less costly because they can leverage the SOA infrastructure and previously-built services.

A mid-2006 Gartner survey asked 158 companies asked: "To what extent has your mentioned project had a positive or negative impact on your organization in each of the following areas?"

Results included:

  • SOA is not producing significant cost savings for companies. 32% of respondents indicated that they experienced some or significant negative impact on costs and 47% experienced no change in costs as a result of using SOA
  • But SOA is having a positive impact on business agility. Over 50% of companies indicated that SOA has a positive impact on their business agility, 17% of these said the impact was significantly positive.
  • Between Mid-2005 and Mid-2006, companies grew spending on SOA, Web Services, and Web 2.0 by 58%.

The findings are in line with the priorities from Gartner's 2006 CIO Survey, which for greater business awareness within IT. Gartner predicts that SOA will be used in some parts of more than 50% of new, mission-critical applications and business processes designed in 2007 and in more than 80% by 2010.

SOA reality brings SOA problems
James Governor, principal analyst, Redmonk

We're hearing too many organizations making really bad decisions -- such as rewriting their entire application portfolio to make it service-enabled. In other words SOA as a science project, which is exactly the wrong approach.

SOA needs to be a business driver and should be piloted, rather than adopted for its own sake. Indeed a touchstone of any SOA initiative is that it should involve a constructive conversation between IT and the business. Like ITIL (IT Infrastructure Library) in service management, SOA should be seen as a means to better alignment between IT and the business.

Another problem is that people are focusing on the "service oriented" rather than the "architecture." However, it's architecture and discipline that will enable SOA to deliver value. Without solid architecture and governance in place SOA is basically a waste of time.

SOA may well be supplanted by another buzzword, but that doesn't lessen the impact of the trend. We have been making slow progress towards SOA since businesses started using IT in earnest. Sometimes it seems like two steps forward and three back, and the progress is far from inexorable, but then progress never is. The fact is the industry is more standards-based than it ever has been before, which enables the latest iteration of SOA, and should enable further progress.

All that said -- big SOA is also being stalked by little SOA -- that is approaches that don't require some arbitrary "Web-services stack" defined by a few vendors -- in the shape of the WS-I stack. Thus, for example, the UDDI buzz has become far less, but WSDL is still used in many development shops.

Technologist accentuates the SOA positives
Eric Newcomer, CTO of Iona Technologies Inc.

The usual hype behind a trend with so much potential as SOA seems to be ending as customers start to make up their minds about how they want to take on SOA, if they do at all. I think customers are starting to emerge from their studies and their initial POCs with real projects in mind. One customer for example recently conducted a long study on reuse, including its ROI picture and prioritized list of candidate services. Now they are ready to engage on a real project. Others we talk to, particularly in the telecom industry, clearly seem to understand the benefits of SOA in delivering new products to market more quickly and in helping streamline the processing of order transactions for them.

The major difference is that we are seeing more customers coming forward with real projects. I believe that most customers spent the last year or so on research and on non-technical tasks such as thinking about the right approach, the question of having the right skill set available, how they wanted to organize their SOA initiatives and so on.

We say all the time that SOA is an approach, not a technology, but this means that customers have to take a step back from their discussions with vendors and make some decisions on their own about how they want to proceed with it. When it's more of a technology discussion, the vendor can take the responsibility -- say for figuring out how the particular feature or function maps to the customers' requirements, or it can be a shared activity between customer and vendor.

But when the new thing is really more of a different way of doing things, it represents more of a cultural change than learning how to take advantage of a new feature.

The impact of the maturity of the industry is that customers have enough features and functions, they have enough software, they have enough applications now (for the most part), but what they need is to improve upon what they already have and get their applications working together better. They need to constantly do more with less and many customers are having trouble reconciling IT budget with the bottom line, with getting the ROI right. SOA can help address these problems, but it also requires organizations to re-think how they do application development. Instead of building applications one by one as the need arises, department by department, they are starting to view IT as a potential collection of reusable assets across departments and sometimes even across companies or over the Internet.

SOA crossing the chasm
Ron Schmelzer, senior analyst, ZapThink LLC.

It seems that we have finally "crossed the chasm" of SOA adoption, with SOA now becoming a mainstream and increasingly well-adopted approach towards dealing with a wide range of issues caused by IT's inability to respond to heterogeneity and change. We know that SOA has reached mainstream acceptance because of a few key indicators in the market:

  • Consolidation of SOA vendor offerings and expansion of platform offerings - Whereas the first half-decade of SOA's presence as a major market force was dominated by the emergence of new vendors with point-solutions and new offerings, the last 12-18 months has been dominated by the consolidation and acquisition of these vendors and the significant expansion of large platform offerings in the space. Most of the vendors who have filled obvious gaps in large vendor product lines are now part of those large vendor offerings. It is likewise a more difficult market for new entrants as the existing market becomes mature and consolidated.
  • Dramatic increase in the number and size of SOA deals - The past 12-18 months has seen increasing size and scope of SOA deals with some exceeding $150 million for relatively short-term projects. In addition, the public sector (federal, state and local governments) have made significant and long-term commitments to SOA as the primary architecture of choice and, as such, SOA does not seem to be a passing fancy in either the private or public sector.
  • Increase in the depth of customer sophistication and SOA implementations - Rather than simply being about Web services or a collection of shallow service implementations, companies are looking at the more difficult and SOA-specific issues of security, management, reliability, quality, governance, process-driven composition and semantic integration. As such, it's clear that SOA has moved from an isolated, proof-of-concept or small pilot concept to becoming mission-critical and core to the long term efforts of organizations.
  • Tipping point reached as companies move from service production to consumption - The rapid emergence of consumption-side development through Ajax, RIA, enterprise mashup and enterprise Web 2.0 platforms signal the trend that companies have built a sufficient quantity of key services that the emphasis is now more so on composite app development more so than service exposure. This indicates we have reached a certain critical mass or tipping point with regards to SOA.
  • Significant increase in the demand for enterprise architect resources - The key issue now facing success of SOA is not its business relevance, availability of tools or maturity of standards, but rather the availability of resources and personnel to help expedite SOA solutions. This need shows that the demand for SOA skills outstrips current supply and, as such, the market will quickly move to meet this need.

For more information
SOA infrastructure: Mediation and orchestration

Why SOA needs IT management

SOA is not the end-all and be-all of enterprise architecture and surely something else will come along in years to come. However, SOA has proven to be significant, valuable and complex enough that organizations will be working on making their SOA efforts return value for long enough that we can count on the next new wave of architectural evolution depending on or building upon the success of SOA.

Dig deeper on Service-oriented architecture (SOA) Design

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchSoftwareQuality

SearchCloudApplications

SearchAWS

TheServerSide

SearchWinDevelopment

Close