TIBCO: could these padded jackets reduce M&A insanity?
I worked for Arthur Andersen a few years ago, specializing in systems integration using Middleware platforms such as TIBCO, IBM WebSphere MQSeries, Vitria and SeeBeyond. Since 2001 I shifted my focus from systems integration technologies to Knowledge Management integration strategy (more on that shift in a later post), but I do keep an eye on progress on systems integration where I can. For TIBCO integration, Rourke McNamara’s TIBCO (& cooking!) blog is one great vantage point from which to watch.
Reading http://www.rourkem.com/2006/12/04/tibco-activematrix-one-container-to-rule-them-all/ I was struck by the significant progress that TIBCO has made in the last few years, in particular the further abstraction of the Enterprise Java Beans concept to the point where modules written in practically any language can be hosted and flexibly deployed into the TIBCO Active Matrix Service Grid:
[Say] you break your application down into four pieces, or services. You decide to build one of those services in .Net and the other three in Java. You deploy all four the same way, using AM Service Grid. Your application can now run as one OS level process with all the services communicating in-process even through one of those services is .Net. You have an XA level transaction that spans all four services and you didn’t need to do anything fancy to make that work. Database connections are shared across all four services. When your .Net service invoked one of the Java services it did so with a single line of code, as if it were simply invoking a function.
Four months have gone by. Another department in your company wants to re-use your .Net service as part of a new application they are building, but they don’t want to host their application on your server. No problem here - they still invoke your service with one line of code, but now that call flows from their node to your node, and the response comes right back. Time to put their new application into production. Maybe the sysadmins decide that this new application should be deployed to the same node as your application. No problem, no code needs to be changed, and now the communication is in-process again.
Two years later your company’s compliance officer tells you that all the social security numbers flowing between the application services must be encrypted and a log must be created with a dump of all updates to a certain list of employee records. Still not a problem. You purchase the ActiveMatrix Policy Manager, create a policy that does exactly what you’ve just been asked to do - but graphically, without writing code - run through a quick test cycle in your QA lab, and then deploy this policy to your applications without changing a single line of your service code or redeploying your services.
From a technical standpoint, TIBCO is providing unprecedentedly abstraction in platform capability for building new interoperable applications. From a business standpoint, it looks to be a winner for revolutionizing and rationalizing an organization’s existing technology portfolio, and contributing new heights in flexibility for ongoing works.
What do I mean by this? Well, a huge factor in making real successes touted in Mergers and Acquisitions (M&A) activities is whether the technical platforms powering the prospective firms will jive. Whether an M&A activity is deemed worth it depends on cost savings and new opportunities afforded from increased automated interactions across the firm. Just as good technology speeds the flow of information, poor technology implementation seriously hinders flow and is a crippling drain on manpower and cash alike.
The neat thing about the TIBCO Service Grid is that it wraps brittle technologies and code fragments into smooth padded coats of loveliness. Applications that were designed a long time ago for limited contexts can be reused and orchestrated into grander schemes alongside the whizziest of SOA buzzword compliant modularity.
This padding that software components get in this environment allows them to ride along next to one another without the usual casualties, emergencies, sticky-tape and glue you get when practicing ad hoc-integration.
So TIBCO Service Grid should further help reuse, scale and make interoperable old works. In turn, the new confidence that problems can be solved should accelerate rationalization, and therefore improve the accuracy of returns predicted by M&A practices. From a strategic point of view, I hope fewer technology constraints will allow firms to make M&A deals along vectors that makes best sense from company value-proposition point of view.
In the billion dollar M&A deals that typically mandate a sudden acquisition of these technologies the stakes are high and pressure is intense to deliver rapidly. Integration architecture solutions do have high price tags, but they are known price tags and in exchange they promise substantially reduce uncertainties about whether the deal will work and what it will cost.
These infrastructures are becoming increasingly ubiquitous: all large firms should have an integration capability at its core. With integration at hand, firms can afford to be both more sharply focused and more loosely coupled, able to automate the efficient exchange of operational data with a multitude of different partners. As such, they reduce the need to conduct M&A in the first place, instead allowing firms to innovate across loosely coupled business webs that reconfigure as business conditions morph.
In both contexts, M&A and loosely coupled business webs, solutions such as TIBCO are indispensable for keeping your staff, and shareholders, from tearing their hair out.
