Problem solve Get help with specific problems with your technologies, process and projects.

Thinking beyond 2012 and SOA, event processing takes the fore

As we move from 2011 into 2012, Todd Biske predicts the next big thing in enterprise architecture -- Event processing.

As we move from 2011 into 2012, what do you predict will be the next big thing in enterprise architecture?

This question is a chance to ponder what the next thing might be. What is happening today that is positioned to move from the bleeding edge to mainstream adoption by 2022? If you've followed my blog, my answer will come as no surprise: events. In fact, my very first blog in 2011 was titled, "Make 2011 The Year of the Event." While I certainly won't claim that 2011 lived up to my request, I do believe that there are several forces coming together that will result in a stronger emphasis on event-driven systems in the next decade. 

The first is the attention being gained by node.js. Node.js is a set of server-side JavaScript libraries that runs within Google's V8 JavaScript engine. Why is this important to events? Node.js uses an event-driven, non-blocking I/O model.  It is asynchronous. This leads to very good performance and scalability. When LinkedIn release their overhauled mobile application back in August, they also stated that they had moved all of the supporting server side functionality to node.js.

While most of the stories around node.js I've seen come from startups or the technology community, rather than the corporate IT department, I'm hopeful that it will find its way there. I view this in a different light than a move from Java to PHP or Ruby. Those languages still followed the familiar request-response paradigm of web applications. If node.js does continue to gain traction, we'll see an increase in the number of server-side developers who understand event-based programming.

The second is mobile. The use of mobile devices is clearly much further along, but the thing that is still in the growth stages is the use of notifications and location-based services. Today, most of these mobile applications first have to be launched by the user, and then request access to location, and then do something intelligent with it. There's no shortage of people talking about a world where that approach is flipped on its head. The phone's location is detected, and some application tells the user, "Hey look at me! There's something I want to show you that's important based on where you are!" This is an event-based approach. 

In order to do this, clearly, we need an event stream. The foundation for that has been established through Facebook and Twitter. These are streams of "events" all just waiting for someone to take action in response. With any event stream, there are plenty of signal-to-noise ratio concerns. I'm sure all of us know someone who has looked over our shoulder as we viewed our Facebook wall or Twitter stream and said, "Why do people post this stuff?" While it would be great to make improvements on this, the advances in processing "Big Data" have given us the tools necessary to process this data. 

Finally, from a business standpoint, in most cases, we've progressed past the point of making information available, and now the emphasis needs to be on finding the relevant information for the tasks at hand. Gartner was on the right track with this back in 2008 when they introduced the concept of "Context Driven Architecture." Context is what must be used to make things relevant. It's not any one piece of information, but rather, a combination of demographics, time, location, historic activity, peer activity, recent actions, and much more. We can build up a model of context for a user or a system by tracking events and the ensuing responses. In the past, analytics may have focused solely on outcomes (e.g. transaction log) but losing all of the events that led up to the decision.  Analysis of click-stream and HTTP logs improved things, but there's still much more than can be done. 

There are more pieces than this, some mainstream (e.g. BPM engines) and some that have strong niche application (Complex Event Processing systems), and of course, all of this will require having a robust portfolio of services, since that's where the actions happen. But when I think about all the pieces required for this, it's hard to find an area that isn't gaining traction. Perhaps the only thing wrong with my blog post at the beginning of 2011 was the timeframe. SOA didn't become mainstream in a year, but you could easily say that the 2000's were the decade of the service. Now, let's make the 2010's the decade of the event, with event and context driven systems and processes becoming mainstream, driven through the success adoption of event, mobile, social, and big data technologies.

Dig Deeper on Topics Archive

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.