Before I start, I have to make a comment about the analyst dinner last night. I usually have a hard-and-fast rule about not blogging anything that happens when I have a drink in my hand, but I want to shout out to Heidi Bartlett for organizing the analyst summit yesterday and arranging for an amazing dinner last night. Not only were there excellent food and wines, but I sat with Christopher Ahlberg (Spotfire) and Bruce Silver so had excellent conversation as well about analytics and BPM.
The keynote session was hosted by Tom Laffey, EVP of products and technology — an engineer in the sea of sales and marketing people that’s typical for these events. After a brief intro, we heard from Vivek Ranadivé reprising his message from yesterday of Enterprise 1.0 being the mainframe era (1960-1980), Enterprise 2.0 being the database era (1980-2000) and Enterprise 3.0 being the event-driven era (2000-2020). Someone really needs to get the message to him, or whoever in marketing writes his stuff, that Enterprise 2.0 has a specific meaning in the current vernacular, and this isn’t it.
He does have a good message, which is that we’ve moved from having some of the information available in some places some of the time, to having all information available in real-time, on demand wherever we want it. In reality, we’re not there yet — my bank, one of the largest in Canada, still can’t post my banking transactions to the web in less than 24 hours — but infrastructure like TIBCO is going to help to make it happen by providing the mechanisms to tie systems together and perform complex event processing. He seeing the transition from the last 10 years to the next 10 years as being from static to dynamic, from database to SOA, from ERP to BPM, and from BI to predictive business. A modern-day event-driven architecture has an event-driven service bus as the backplane, kicking up events to the "event cloud" where they can be consumed and combined by analytics for visualization and analysis, rules to determine what to do with specific combinations of events, and BPM to take action on those decisions.
Ranadivé was followed by Kris Gopalakrishnan, CEO of Infosys (a major TIBCO integration partner), who talked about the changing markets, economies and demographics, and how enterprises need to change in order to respond to and anticipate the new requirements. A rapid consumerization of enterprise IT is happening, with greater demand for richer digital experiences both by internal enterprise workers and external customers. Process cut across systems and organizational boundaries, and need to be managed explicitly. IT systems need to be exposed as services that are aligned to business operating model in order to allow IT to respond quickly to business needs. Analytics need to be provided to more people within an enterprise to aid decision making, and there needs to be a convergence of BPM and BI to drive business optimization. He sees that the fundamental problem of information silos still exists: a point of view that I agree with, since I see it in client organizations all the time.
We then heard from Bob Beauchamp, CEO of BMC Software, to hear the customer viewpoint on IT process management using TIBCO’s products. The theme of the conference is "building bridges", with lots of pictures of the Golden Gate bridge and other famous bridges as slide backdrops, and analogies about building bridges between systems, and he used the Golden Gate bridge in another analogy about software: the bridge cost $24 million to build, and $54 million per year to maintain. This analogy is especially true of custom integration software, where in many cases you either effectively rewrite it constantly to keep up with other changes in your environment, or allow it to fall into disrepair.
In particular, however, he’s talking about how IT processes are the last to benefit from new technologies, since they’re too focused on providing these to (or testing them on 🙂 ) other departments within an enterprise. BMC is using TIBCO ActiveMatrix and some of their own technology to bring functions together in order to enable more efficient IT processes, including service support such as asset configuration, service automation such as auditing and compliance, and service assurance such as predictive analytics and scheduling. He sees this as transformational in how IT is managed, and believes that it will have a huge impact in the years to come.
Next up was Anthony Abbattista, VP of technology solutions at Allstate insurance. They’re a huge company — 70,000 employees and 17 million customers — and always felt that they were unique enough that they had to build their own systems for everything, a mindset that they’re actively working to change now. With a 7×24 operation that allows customers direct access to their back office systems, they had some unique challenges in replacing their custom legacy systems and point-to-point integration with standardized reusable components that gives them greater agility. They’ve completely rearchitected for data hubs, service bus and a range of new technologies, and taking advantage of standards to help them move into a new generation of systems and business processes.
“he used the Golden Gate bridge in another analogy about software: the bridge cost $24 million to build, and $54 million per year to maintain. This analogy is especially true of custom integration software, where in many cases you either effectively rewrite it constantly to keep up with other changes in your environment, or allow it to fall into disrepair.” – interesting analogy, unfortunately discredited by distortion of the scale. I am sure the quoted cost to build is an original investment without correction for inflation. If it cost twice as much to maintain, as it is to build, it should be rebuild a new as soon as possible. It is done in civil engineering all the time, but perhaps not as soon as it should be.
How many SOA projects, that involve replacement of legacy systems, do you see?
It’s true that $24M would be much more in today’s dollars, but my experience is that over the lifetime of custom software, you spend about 10x maintaining it than you spent building it in the first place. There’s huge inertia in replacing systems, which is why they tend to be (expensively) maintained rather than replaced frequently.