The odds are good that the LAMP stack is running somewhere inside your company. The acronym refers to the foundational foursome of the open-source movement: the Linux operating system, Apache Web server, MySQL database and, collectively, the Perl, PHP and Python programming languages. Development tools such as Eclipse and application servers such as JBoss have also gained popularity—and trust—especially now that major vendors such as IBM, BEA Systems and Borland have adopted or supported them commercially. But what about the next step up the software ladder? Is open source ready for ERP, business intelligence or CRM?
Ready or not, it’s happening; the first industrial-grade applications in these areas are now emerging. And CIOs will soon need to decide how to approach these fresh options in their enterprise software catalog. As with the adoption of the LAMP players, these new open-source enterprise applications likely will find their way into the enterprise at a departmental or small-project level. As a result, “we don’t see [these applications] on CIOs’ agenda at all,” notes Michael Goulde, an open-source senior analyst with Forrester Research. But, he warns, “CIOs should sync up with their development teams to see [where such applications] might have payback to the organisation.”
However, CIOs should tread carefully on such open-source applications, advises Mark Lobel, a partner at PricewaterhouseCoopers who focuses on information security, including security for financial applications. One key concern is that applications tend to reflect and embed business processes and logic, which often are key strategic assets you don’t want to share with others—and open-source licenses can require such sharing if companies aren’t careful. Another issue is the long-term viability of open-source applications for specific functions. Open source depends upon volunteer developers for success, but the more niche a product, the smaller the potential pool of interested contributors. As such, grassroots support for specific apps such as ERP or CRM tools may look more like brigades than the armies now supporting broad open-source infrastructure such as Linux, Apache and MySQL.
Still, properly managed open-source applications can save enterprises money and time—as well as reduce dependency on specific vendors.
Finding a Fit
Financial-services giant Fidelity Investments has used open-source technology for about four years to reduce costs and dependence on vendors. “We started with Linux like everyone else did, but our intent all along was to see how far up the stack we could go,” says Charlie Brenner, senior vice president of the Fidelity Center for Applied Technology, Fidelity’s technology incubation group. After Linux, Fidelity adopted Apache and Perl, and then the Struts Web application framework and the Eclipse Foundation’s development environment. Fidelity is now looking at open-source database management systems and assessing what applications might make sense. The advantages of open source include widespread component reuse, better access to underlying code to customise interfaces across applications, and less complex systems to manage. “We’re heavy users of proprietary [software], and that won’t change, but there are times you need a motor scooter, not a truck,” Brenner says.
Others are less interested in picking the proper vehicle than they are in creating a uniform, inexpensive core on which to hang their IT business. At Midland Memorial Hospital in Texas, “we’re trying to get a complete open-source or public-domain stack rather than be proprietary,” says IS Director David Whiles. His organisation already uses the LAMP stack and is now deploying a public-domain electronic records system, the Veteran Administration’s Vista, for less than half of what a proprietary system would cost (even with the cost of hiring a consultancy to add features such as billing). Medical industry service provider Athenahealth, meanwhile, is using SugarCRM—an open-source CRM package. CTO Bob Gatewood says he had several reasons to switch from his current CRM provider, Salesforce.com. But he notes that making the change will save about $1 million over three years in per-user licensing fees, even after the cost of development and integration is subtracted. He expects to complete the migration in early 2006.
Beyond spending less, Gatewood plans to more closely integrate the SugarCRM code—which he can access directly—into his call-centre and other support applications, something not possible with proprietary software where code is tightly held by the vendors. Other IT execs seek the same benefit. “We can take the pieces we need [with open source],” says Bob Hecht, vice president of content strategy at specialised data provider Informa, which is investigating the Alfresco open-source knowledge-management application as an alternative to commercial enterprise content-management tools.
Informa is exploring Alfresco because a license for a commercial enterprise content management application for a company of its size would cost millions of dollars and would impose a single content-management model on the company’s array of publishing, training and events businesses. “We just won’t do that,” Hecht says. (It also helps that Alfresco was developed in part by former Documentum technologists, giving Hecht more confidence that the application will be enterprise-class.)
Open-source applications can make especially good sense for nonstrategic, fairly generic applications like reporting or sales-force automation. Departments that have unique technology needs and smaller companies with limited budgets are also more likely to consider open-source applications, says Forrester’s Goulde. “Larger companies are not about to rip out SAP. Plus the functionality and the integration are both more complex” for a large company than open-source apps currently can handle, he adds.
For example, open-source tools “are not going to take the business-intelligence market because they are not yet competitive with commercial software,” says Eric Rogge, research director for BI and performance management at Ventana. For example, open-source BI applications don’t yet offer a comprehensive platform with reporting, ad-hoc analysis, online analytical processing (OLAP) connectivity, alerting, dashboards and workflow. Nor do they offer aids for developing user-interface controls, ad-hoc analysis against relational data sources or scorecard functionality with strategy maps, metrics management and collaboration features, he says. But Rogge does expect open-source applications to eventually make inroads in the BI reporting tool segment, since there are a variety of uses for basic reporting tools in an organisation where a costly, complex BI tool isn’t needed.
Furthermore, increased adoption of open-source databases should encourage the development of open-source reporting tools that take advantage of them, says Don DePalma, an analyst at the consultancy Common Sense Advisory. “Most database activity is about reporting, analysing and crunching the data, so [open-source reporting tools] would seem a natural development. Companies, universities or governments using open-source operating systems and databases would be a great audience for such software,” he says. DePalma doesn’t expect a popular reporting tool like BusinessObjects’ Crystal Reports, for example, to support open-source databases because of the vendor’s relationships with proprietary database developers such as IBM, Microsoft and Oracle. That provides an opportunity for the open-source community to create a Crystal Reports–like reporting tool, he says.
Open-source applications also make sense when there are regulations or other requirements common to an industry, where having a mutually supported tool would benefit everyone and not put anyone in the position of losing a competitive advantage, Goulde says. Analysts most often cite the health-care and financial-services industries as candidates for these kinds of tools, though liability concerns surrounding legal requirements make it critical that potential users understand the possible risks, notes Fidelity’s Brenner. It is also possible to imagine a large player in a specific industry making an open-source application viable, perhaps for some supply-chain management functions, much as Wal-Mart has done for RFID, notes Forrester Research ERP Analyst Ray Wang.
Gauging Open Source’s Risks
But using open-source applications does carry risks. One is that staff developers unfamiliar with the competitive value of various components might accidentally embed strategic business logic or processes into code that is then provided back to the open-source community, neutralising a competitive advantage.
But CIOs should be able to manage their strategic assets while still choosing open-source applications, says Eric Link, Diabetech’s CTO. Business logic, for example, should not reside in modified open-source code but in your internal rules base or in-house applications that call the open-source tools, as is common in commercial ERP systems, he says. “It does require careful thought to know what is strategic,” but any IT development effort should make such an assessment, whether it involves commercial, homegrown or open-source code, Link says.
CIOs should also be able to distinguish between applications and platforms and the issues that surround each, Brenner adds. Reporting tools and CRM are two examples of platforms that are often marketed as applications, he notes. The difference is that platforms typically don’t encapsulate specific business processes or logic, making them well-suited for open-source efforts—and less risky for the companies that use them, as companies using such tools will be less tempted to insert their own business logic into the products and unwittingly release it to the world. A reporting tool, for instance, might act on a company’s data, but it would never incorporate that data into its own code—and thus a company would never be required by the license to release the data as open source. (Another alternative is to go pseudo open source as in the Avalanche Corporate Technology Cooperative, which openly shares code on a variety of projects, but only among subscribed members.)
Beyond intellectual property concerns, another significant risk is an application’s long-term viability. Open source has worked well for widely distributed tools such as those in the LAMP stack that are typically run as-is and don’t need to be customised at each location. But for niche applications, the community of developers is necessarily smaller than for a piece of infrastructure, reducing the resources that contribute to the application’s development, maintenance and support. This could make it difficult for many projects to muster sufficient developer support to stay viable. The diversity of applications will be a difficult issue for the open-source community, says PricewaterhouseCoopers’ Lobel.
This limitation is exacerbated if companies don’t share their developments with the community for fear of releasing competitive business logic. “I can’t see it going very long if companies aren’t contributing back. An open system works only when it’s open,” Lobel says. Diabetech’s Link, however, believes that argument is overstated, since companies are typically happy to share infrastructure code with others, thus moving the application forward even while keeping their business-specific code to themselves.
Despite these issues, even cautious observers concede that open-source applications can make sense beyond the LAMP stack: And sensible CIOs should start paying attention