In many commercial computing domains, there are two purchasing options: buying from a single provider or constructing a niche, or “best of breed,” array of components.
Each has its advantages: Purchasing an integrated suite from a single vendor, for instance, is often simpler and quicker, and it involves fewer surprises than dealing with multiple sources.
In the ALM marketplace, customers can now enjoy the advantages of both approaches: leading products integrate a wide range of functionality yet can cooperate with tools or components from other vendors. ALM is important now and will become more so over the next few years. One engine for that growth is the openness of commercial products.
Those are the conclusions from recent discussions with market watchers and analysts; more than that, the same analysts report these as conclusions of the market itself. As Theresa Lanowitz, founder of independent analytics group voke inc., puts it, all the leading technologies, including ALM products from Hewlett-Packard, IBM, Microsoft and MKS, are “open to integration.” Lanowitz and other analysts see the marketplace confirming this on three levels:
- Even the ALM framework providers that sell integrated systems provide “hooks” to connect to third-party materials;
- The most exciting and rapidly growing tools companies, including Coverity, Electric Cloud and Replay Solutions, are precisely those that have invested in integrating well with ALM frameworks; and
- The most satisfied customers are those that take advantage of framework openness to tie ALM deeply into organizational processes.
It doesn’t require insider perspective to understand: With few exceptions, an open approach to integration is winning ALM competitions.
Not an obvious bet
It’s rare for any business choice to appear so clear-cut, and ALM engineering hasn’t always tilted in one direction. In a time when cloud and as a service are IT’s leading buzzwords, it would make sense that companies are considering buying all-in-one ALM packages in which each function is certified to fit well in the workflow. Thomas Murphy, a research director for Gartner Inc., emphasizes “the burden of complexity and value of quality” in software management and how much organizations need the benefits of ALM without the cost of more complexity and maintenance. IT departments have their hands full focusing on automating domain knowledge on lean budgets; it’s unrealistic to ask them to study up on ALM and assemble their own niche stable. Most major ALM product frameworks originated as end-to-end, “fully baked” systems, or at least were marketed that way.
Several business trends have promoted openness in ALM technologies, though. First, “we’re all in software now,” as everyone from Lew Moorman to David Kirkpatrick has observed, and therefore software management has become the fundamental differentiator and competitive skill. Second, product and organizational cycles only shorten: Companies are subject to merger and acquisition over just a few years. Murphy and other analysts emphasize that divisions don’t simply “dump $3 million invested in tools and training” for one particular technology when acquired by a company that uses a different ALM system. Finally, the open source community is on a two-decade winning streak of demonstrating that protocols and interfaces matter more than implementations.
The result: No product development team, not even those at the giant ALM vendors, can keep up with all the change and complexity modern ALM needs to control. Vendors’ greatest value is to define best practices, provide an out-of-the-box experience that models a valid workflow and engineer enough configurability to connect to all the legacy tools that are indispensable to any organization. This enriches the whole ecosystem of framework vendors, customers, specialty product providers and consultants.
There’s still a place for end-to-end ALM products, of course, particularly in regulated industries that rely on embedded software, such as medical equipment and aviation, and also in greenfield construction of geographically distributed teams. Murphy notes the likelihood that different constituencies in the same organization have different requirements of ALM:
While a “line worker ... wants fast and straightforward” out-of-the box behavior and management requires “reporting and regulation,” developers often adjust and configure ALM to match processes already in place.
The wider context
Also crucial to a proper understanding of the role of integration in these frameworks, says Murphy, is to distinguish ALM from SDLC, or software development lifecycle. Sometimes decision makers use the two interchangeably. As Chris Birmele and Mridula Velagapudi explain, properly understood ALM includes not only all of SDLC, but also the business analysis before any development has begun, as well as all the use after development has ended. Moreover, comprehensive ALM must allow for multiple SDLCs: different acquired business units; software destined for embedded systems, Web applications or desktops; software oriented to “big data,” accounting or domain techniques; and software coded in Java, SQL or Ruby. Only at the ALM level can these differences be consolidated and effectively managed. In many cases, products advertised as ALM actually take on only narrower SDLC responsibilities. This underscores for Murphy the importance of clarity in goals for ALM roll-out, a topic he elaborates in such Gartner publications as Selection Criteria for Success in Choosing ALM Products.
It’s a good time for ALM customers. Healthy competition among leading vendors has brought the market to a point where several different products pair good integration with openness to customization. You should expect to put a lot of work into defining the precise criteria for your ALM adoption; you’ll receive even greater value once you make a choice and put it into practice. Another way this theme of openness benefits customers is in its reduction of risk. Such areas of computing as enterprise resource planning, or ERP, and advanced storage systems have emphasized costly “big bangs,” where years of effort should lead to revolutionary process improvement. While ALM also targets qualitative and quantitative improvement in manageable processes, the detailed implementation of ALM can be incremental. This gives organizations the opportunity to choose the level of change and risk that best suits them. The business consequence of openness, therefore, is finer-grained control over planning and execution.
About the Author
Cameron Laird is vice president of Phaseit Inc., where he implements software projects and publishes articles about the results. He is a longtime developer, manager and author, and his current work emphasizes performance and security in HTML5.