A few weeks ago I had a call with Intalio's Ishmael Ghalimi.? It was right before vacation, and the unfortunate side-effect is that I'm only now catching up with writing the post I meant to write when I got off the phone with him.
For the last couple of years, many of us in the BPM world have wondered what exactly Ishmael and his crew at Intalio were up to.? They made a series of acquisitions and investments, the purpose of which wasn't quite clear to much of the BPM community, because it included a CRM solution and a focus on cloud computing.? But when we finished talking, it was clear to me that Ishmael has simply made a different bet than most of the rest of the BPM community.? While the mainstream of BPM innovation has been focused on BPMN2, modeling and collaborating in the cloud, and tweaks to the method and thought process like ACM... Intalio has gone a contrarian route-? retooling and refocusing on BPM in the cloud - but not by dabbling, by jumping with both feet, so to speak.? This is the long game rather than the dink-and-dunk short game.
It is a bold move, and Ishmael makes the argument that if you really want to play with the big boys, you need to make the investments that Intalio has made in a vertically integrated stack for cloud computing.? Rather than writing all the code from scratch, of course, Intalio is relying on a bevy of open source software.? A quick look at this page reveals just how many open source projects are involved in Intalio's offering (and that's just on the application engines side).
A heavy focus on open source software might cause you to wonder how Intalio makes money.? Intalio contributes to some of these projects, and just leverages some others.? the distinction that Ishmael made is that while the open source project for, say, the process engine (Apache ODE and Drools Flow), is free and open source... The entire integrated solution that Intalio is offering is commercial.? You are, essentially, buying a professionally prepared 10 course meal, instead of just getting all of the ingredients for free, so to speak. As Ishmael points out, customers do not mind paying for commercial software today that includes open source parts (nearly all of the commercial BPMS offerings today include at least some significant open source project code).
Ishmael went into great detail about the investment required to make Intalio's PAAS offering virtualization- ready - the upfront investment to natively leverage SpringSource and the like is significant. But once there, it is much easier to scale application deployments.
We walked through a pretty compelling demonstration of the authoring tooling - which was browser-based, and he scored bonus points with me by running the whole stack on a Mac (attention commercial BPM vendors: your users and developers want Mac, iPad, and iPhone support).? The ability to model data, leverage it in user interface and process diagrams, and deploy it quickly, is compelling.? It doesn't *feel* like a BPM tool, it feels like an IDE for building apps in the cloud.? Perhaps that is the intent.? To make BPM a feature of the environment, so to speak, rather than the centerpiece of the environment.
I asked Ishmael, given that this seems like both a horizontal and global strategy and a great number of component projects that make up their product offering, how does Intalio decide what NOT to do?? We had discussed why he had made the investments he'd made, but how do they decide what not to do?? Ishmael's take is that they are throwing the long ball - seeing themselves as a much smaller version of a Microsoft of the 90's, where:
- The operating system is the cloud instead of Windows
- The "development studio" is Intalio Studio instead of Visual Studio
- The productivity suite is Intalio Docs instead of or in addition to Microsoft Office
Another analogy he gave is RedHat - providing the glue to pull together open source projects into coherent commercial offerings.? There are many execution challenges, but if you pull it off and execute, you're in very good shape because most competition is either too small to deliver, or too big to abandon their existing commercial software business (or not nimble enough). His example of too large:? he argues it would be very hard for IBM to come to market with an integrated product that is also virtualized for the cloud, because there are different P&Ls within IBM's software group, who aren't going to work together on one seamless offering. Others might argue these points differently, but given Ishmael's thinking, the strategy he's pursuing makes sense in that context.
The pricing model is free up to 5 users, and then scales up in price as you expand the solution.? This makes it easy to prototype solutions, and scale cost relative to the benefit achieved.
In some respects, what Ishmael and Intalio are trying to do is offer a compelling software stack for process applications in the cloud, but also removing many of the barriers to entry companies would normally face.? One could argue that they are trying to make BPM in the cloud simple enough that mere mortals can attempt it.? And Ishmael claims it isn't just a superficial attempt to band-aide these projects together - that they've thought and worked hard on every level of the stack - from hardware, to virtualization, to Infiniband networking, to the more obvious software layers that developers interact with.? The conversation with Ishmael about these topics reminded me of my conversations with Phil Gilbert of Lombardi about making versioning relevant to BPM authoring (that it isn't enough just to check things into subversion, that real versioning requires a much deeper understanding of process authoring).? Versioning was Lombardi's long pass in the last couple of years leading up to the acquisition by IBM. And the point of this kind of deep investment in both cases is to create simplicity out of complexity - to provide truly useful abstractions to the customers of your software.
Three is risk that Ishmael and Intalio may have made this investment "too early".? I recall in the 90's a firm that made a bet on server-side Java runtime environments a bit too early - betting on JRun, which was a pretty immature platform.? A year later, when the product suite was ready, the whole thing could have been accomplished in 3 months on top of newer commercial application servers (e.g. Weblogic).
Time will tell if Ishmael's long pass pays off - he's either stolen a march on the competition, by getting "fully cloud-enabled" early, or he's made the march too early and paid too heavy a price to get there.? If he's too early, we'll know because other "mainstream" BPM vendors will quickly follow with fully cloud-ready BPM stacks.? If he's stolen a march, then Intalio will use its technical advantages (perceived and real) to drive more business at the expense of others.