Next Generation Integration Architecture for Smarter Grids
This is one of the favorite topic all across the board. Each and every utility is brainstorming and each and every vendor is proposing next generation Integration Architecture.
Reason? Well the roadmap of smart grids begin with integration of heterogeneous applications, systems and blurring line between OT and IT applications and systems. Hence the obvious challenge is what should be the integration architecture to achieve the goal of reaching Smart Grid Vision 20XX (is it 2020 or 2030).
There are several reference architectures available in the market place and each one addresses some unique challenges, but they all go on tangent at some point of time depending on the objective.
I would like to share my thoughts and provide food for thoughts for discussion in this (and perhaps other industry) forum(s).
As per my experience and understanding it is premature to say what will be the next generation integration architecture, but we definitely can say what is required out of the next generation integration architecture. At least my background is not in Enterprise System Architecture hence I will focus more on the requirements rather than suggesting the integration architecure itself. But we at infosys are investing on the next generation integration architecture for smarter grids and I would like to present the thought process behind our next gen integration platform in this and future blogs.
Let's take a step back, and think what is really required here? Few general requirements to name:
1. Interoperability between IT and OT systems and subsystems within IT and OT world
Now are these not the same requirements exist in any integration problem? So what's different? The difference in the context of the Power Systems Domain is that the IT systems as well as OT systems are fragmented due to the fact that many IT systems were developed in-house, almost all of the OT systems have their own legacy. It is getting more complex with the fact that newer concepts and systems are being introduced both in IT as well as OT side of the world in phase-wise approach.
Very soon we will see newer systems with the objective of interoperability but won't be able to interoperate within themselves forget about the legacy systems. OK, some of the reader will say, "This is exactly what we are addressing in newer implementation", but is that 100% correct?
In order to complete current phase of the project on time, aren't we ignoring the interoperability between the integration systems being implemented in a parallel phase or next phase? The schedule and budget pressures have always been the enemy of the interoperability and homogeneous system implementation.
Let's come back to the objective of this blog, to list the properties or requirement of the next generation integration architecture.
- The integration in smarter grids won't be just based on data exchanges but will have to be based on events that are flowing between the systems and subsystems. Hence the next generation integration architecture must be even centric, i.e. the Complex Event Processing is going to be the key component of the integration architecture.
- The data coming from the field devices is no longer just the real-time data but also the data relevant to the enterprise systems, hence the integration architecture must support the differentiation of data and information integration based on the following time-line and usage:
- Near Real-time
- The time critical data coming from field devices and systems will contain high volume transactional data that means the integration architecture should incorporate the data convergence, aggregation and disaggregation features.
- Another aspect is the analytics (near real-time and real-time) requirements, hence the integration architecture should provide the data in such a way that it enables analytics to be built seamlessly.
- The most important part is interoperability. Now this is critical because no matter how much efforts are put in interoperability, we are at least 5-7 years away from the plug and play kind of interoperability. And even that to be achieved the onus is on the integration architecture. Apart from providing support for all interoperability standard based protocols, in my opinion CIM is most important component. The integration architecture should have it's own CIM based data model so that as and when CIM enabled systems are implemented the integration becomes seamless.
Some of the aspects I did not touch in this blog, but I will write more about next generation integration architecture in next blog, hence would like to stop here for the comments, feedback and discussions on this very important topic.