SOA Dead? Not If You Are A Utility!
Is SOA (Service-Oriented-Architecture) top of mind for Utilities?
I’d have to guess a big “Nope!” For those of you that know me… I was involved in the early days of J2EE/Application Server craziness and then the next wave of early days SOA – so this post comes from the heart! SOA was the next “Big” thing to hit enterprise IT after the app server. The allure of loosely coupled and re-usable “services” that could be quickly assembled into “composite applications” to address the needs of the agile business users within the enterprise - sounded great! And we all thought that “services” would take over the world! Even Gartner Group was hot on the topic! Key events were held each year for the early adopters of SOA – to gather and hear wise and privileged secrets as to how best start implementing services and becoming “service-oriented”.
Sounds great eh? Well, the only problem was that everyone thought it was easy. And we thought it would happen FAST! Everyone involved had just come off the .dotcom era of fast buy, quick deployment, and a “do everything/anything” to get new applications launched! Money flowed freely – and water-tight ROI’s were as solid as the business plans that a lot of start-ups during that time were built on! In the end.. no surprise – things crashed and everything slowed down. It was about this time – that we (software vendors) decided the next thing to sell to over-bought-software stacked customers was a new “architecture”! Enter SOA!
I am a fan of architectures, don’t get me wrong. Like a map, a blue print, a template, or a a guide – they tell you what the big blocks are, how they fit together, what needs to talk to what, and what the communication protocol should be. The missing link to any architecture though in my humble opinion is “TIME!” Architectures are not 3 dimensional in that there is no Z-axis to highlight a time line that it takes to actually implement. And this is where SOA went off the rails… everyone failed to realize how long a process it would take for enterprise to:
- Understand SOA…
- Determine the requited existing/lacking skills sets that would be required
- Designate the “SOA” Champion or Group of Champions to lead the transition
- Assess internal environments to determine what core functions could be exposed as services
- Gain support for internal pilots – both in time to investigate and funds for seed pilots
- Map pilots to useful business use cases – which had deployment ROI attached
- Engage early SOA vendors to roll pilots, learn, re-align strategies and goals
- Battle the myriad of standards that cropped up -> ws*.* nightmare
- And finally create services, create repositories, creates policies, create simple 2-3-4-5 services based applications
Do you feel the pain my friend? I sure did!
So where am I going on this SOA bandwagon with respect to the current Smart Grid, Smart Cities and Smart-Wannabe-Utility companies? They are going to be faced with the same challenge that the enterprise was faced with 10 years ago. And it is “smart” if we all understand that it is going to take time and talent to make all this “smart” stuff work. If you look at the above TC57 Reference Architecture [EPRI - Harmonization of CIM with IEC Standards: Draft Report ] I feel like we are headed for the “Perfect Storm“! Is this architecture scary? Not really… let’s see what EPRI is up to…
EPRI has sponsored development of a number of international standards that provide the basis for information exchange to support power system management. One of the most important is the Common Information Model (CIM), which is rapidly gaining acceptance throughout the world as a common semantic model to unify and integrate data from a myriad of systems involved in support of real-time electric utility operations. As its acceptance as the basis for information integration grows and areas of application expand, it has become increasingly important to ensure that other standards used for the sharing of operational data are harmonized with CIM. Standards include the Utility Communications Architecture (UCA) standards for architectures and interfaces within substations and distribution feeders.
This report, the first in a series dealing with enterprise-wide harmonization, describes the work accomplished to date in achieving this desired harmonization. The report analyzes the most important standards to identify commonalities and differences that need to be harmonized and identifies specific areas where additional work is needed. Recommendations for a harmonized future reference architecture and specific actions needed to resolve existing differences to achieve an enterprise-wide unified information infrastructure are presented.
Web services will play a role within this architecture. ESB’s (Enterprise Service Bus) will start to pop up although maybe they will be called Utility-Service-Buses! The fact is plain and simple – the utility back-end needs to change in order to accommodate the new technologies and applications that the Smart Grid will require to operate and properly service end-user consumers. (I haven’t even touched on HAN – and all the home consumer apps that will also spring up…)
The good news is that SOA has matured, talent pools are available, and the F1000 enterprise helped all the SOA vendors shake down their products and capabilities. The one factor though that still remains a challenge – is time.
This change my friends will take some time… so the next time you are with a Utility prospect and someone says “SOA” – please buy that a person a beer!
They are the next-generation champions!
Everybody like a party right? Well I had the chance to attend the past SVLG event in Silicon Valley (where else?) hosted at the rocking NetApp campus. Nice digs! Super presenters and crowd.
