Thursday, March 31, 2011

STKI Summit 2011 – SOA Perspective: Business Services or only Integration

On March 28th 2011 I participated in the STKI Summit by the local analysts company STKI. 
The company headed by Dr. Jimmy Schwarzkopf is focused on the Israeli market. 
Presentations were based on an Israeli market survey performed by the company.
The company's analysts presentations topics were Infrastructure Trends, Applications Trends, Office of the CIO trends and Infrastructure Services Trends, followed by wrap-up IT Trends presentation by Jimmy.
The survey may not be accurate but the information is valuable and probably reflects the overall market status.    

The theme of this year conference was: Manage Scattered.
The prime areas appearing in all presentations were: Mobile Computing, Social Computing and Cloud Computing

The combination of them is the basis for creation of totally different IT industry, as well as Business and Personal Interaction and Collaboration.

The theme is about IT efforts and solutions which are executed without the IT department involvement (e.g. SaaS and other Clod Computing solutions implemented by outsourcer, Inter Enterprise Collaboration by Facebook, Blogs by company's employees etc.). It was described by an analogy of a shepherd losing his scattered sheep. The The theme is about recommending to the CIO's to regain control over all scattered IT related efforts.

SOA Perspective 
Pini Cohen STKI's Vice President and Senior Analyst discussed Application Development and SOA in the Israeli Market. 

Unlike previous STKI summit events, SOA is no longer a prime conference topic.  

The Highlights of the SOA related part of the presentation are:

  • The SOA implementation levels (termed  in the presentation as SOA Maturity levels) are low in most organizations
  • SOA is used for Integration and not for Implementing Services
  • In most cases the implementation is ESB Centered, therefore the local SOA market was described as Integration Brokers and SOA Infrastructure.
  • Market Leaders are: IBM, Tibco, Oracle, Magic (Magic products are invented and built in Israel therefore the local market is biased in favor of that product in comparison to the Global market). The leaders are followed by Software AG and Microsoft

My Take
SOA Value Proposition is in the domain of Business and IT Alignment i.e. reducing the Business-IT gap. 
Usually it is not possible to realize significant benefits without implementing Business Services.

The Israeli SOA Market is actually an ESB market with limited Business Services deployment.
ESB is SOA based Integration.  Many SOA evangelists believe that SOA value propisition is mostly in Business Services and not in Integration. Some of them argue that SOA did not deliever its Integration promise. Read for example Loraine Lawson post

I already addressed the ESB-SOA issue in a previous post titled: ESB for an Orphan

Recently, I presented in the Israeli Systems Analysts conference. 
My Presentation was focused on Organizational and Methodological aspects of SOA and BPM. These aspects are crucial for successful SOA implementation. 

Pini Cohen's presentation is focused only on SOA infrastructure, mainly on ESB, not because he does not understand the importance of other aspects. It is focused on these aspects, because that is the focus of most SOA implementations in Israel.   

No wonder, that the ratio of successful SOA implementations in Israel is even lower than the same ratio in Europe and North America. 

 What is wrong in SOA implementations in Israel? 

The following highlights are generalizations, so there are exceptions whose SOA efforts are better than the average Israeli SOA implementation.    

  • The SOA initiatives are driven and executed by IT department with limited Business involvement and sponsorship.

  • ESB is the beginning or starting point of SOA

There are three main approaches for starting SOA initiative:

Approach 1: Planned SOA
Including training of key IT and Business staff, building an initial Reference Architecture and  selecting a small pilot project which will be executed with appropriate SOA tools or sometimes prior to purchasing SOA tools.

Approach 2: Pain points
The initiative begins by addressing a Pain point by small SOA implementation, hoping that success will promote other SOA projects and even planned SOA initiative.

Approach 3: Implementing SOA infrastructure products which facilitate relatively easy SOA projects implementations
It should be noticed that for some Enterprise a Planned SOA (approach 1) may include purchasing of SOA Infrastructure product in the beginning.

In most Approach 3 cases ESB tools is the first implemented product.
Beginning SOA implementation by implementing ESB is the best approach for the ESB Vendor. In many cases the decision to begin by implementing an ESB was based upon consulting of the vendor's  SOA experts.

The IT department also favors this approach: No need for collaboration with Business people is required, it is IT Department's SOA, which can be executed and controlled more easily.

The problem is that it is like looking for a lost coin near the street lampspot even though the coin was lost somewhere else.

After the ESB implementation in many cases there is no real Business Value and sometimes no Business Services implementation following the ESB installation.

In my opinion in many Approach 3 SOA implementations you should start with implementing service Registry and Repository product and not an ESB product.

If there are already many Web Services in production (and this is often the reality) implementing a Registry can save redundant efforts and stop usage or maintenance of unused or improper Web Services.

  •   Inappropriate Governance

  • The vendor has multiple roles: Infrastructure Products Sale and support, Planning, Consulting, Project execution, Training etc. 

  • No Architecture Group or Center of Excellence

  • Unrealistic expectations of completing transition to SOA in a short time and /or realizing SOA major benefits in short time.

No comments:

Public Cloud Core Banking: Hype or Reality? - Revisited

  More than 4 years ago I was asked if Public Cloud Core Banking is a Hype or a Short Term Reality? If you had read the post, you would prob...