Breaking Away

• By Joseph C. Panettieri
• 04/01/07

Open source projects may free universities from expensive, rigid commercial software. But will the rewards outweigh the potential risks?

Brad Wheeler is breaking with tradition. As CIO of Indiana University, Wheeler is leading an ambitious move to open source-based applications that will manage the bulk of the university’s business and academic systems. One of the initiatives, known as the Kuali Project, involves multiple universities writing and sharing code for their financial and operational systems. Another, the Sakai Project, is a community source platform for academic teaching and course management (see “How Much Do You Know About…”).

If they succeed, Kuali and Sakai are bold initiatives that could free universities from rigid commercial software that’s often expensive to acquire and maintain. But the journey to open source freedom also includes numerous risks. While commercial software has a support network including thousands of hardware, software, and integration partners, niche open source projects may involve only a few hundred— or even a few dozen—software developers worldwide.

Still, more and more universities are seeing the upside of open source. In fact, right now, many universities are deploying or evaluating open source databases, e-mail systems, and business applications. Even voice over IP (VoIP) hardware and call management systems increasingly leverage open source software known as Asterisk (see “Know Your Options”).

SAKAI AND KUALI spearheader Wheeler: Promoting 'enlightened self-interest.'

Kuali and Sakai are unique, however, in that they are designed for universities, by universities. Unlike traditional, loosely organized open source projects which can involve the random contributions of programmers from across the world (some of whom are well-known, some lesser-known) and which can cause problems related to communications, project priorities, cultural issues, etc., the Kuali and Sakai projects have formal, organized teams stretched across multiple universities, with clearly defined responsibilities and goals. Active participants in the initiatives include Cornell University (NY), Michigan State University, San Joaquin Delta College (CA), and The University of Arizona, just to name a few.

“In essence, we’re pooling our monies to run a disciplined project,” says Wheeler, who also serves as chairman of the Kuali Project Board. “These aren’t tight consortiums where you’re locked in for five years: Everyone has walk-away rights, but we’re held together by enlightened self-interest. Each university can write incremental improvements and take advantage of other peoples’ research.”

Cornell, for instance, is writing a Kuali reporting module that IU plans to deploy. Contrast that with the commercial software market, where license agreements frequently bar universities and businesses from sharing homegrown software enhancements with one another.

Another key benefit of the open source movement: Universities frequently can make module enhancements without taking any systems offline. Notes Wheeler: “It’s the opposite of commercial ERP upgrades, where you often have to deal with numerous interruptions as you pay seven figures for an upgrade.”

Taking Control

After several years of development, many of the open source components are ready for prime time. Kuali Financial System 1.0 is out the door and drawing attention from smaller colleges. A nextgeneration release, to be dubbed 1.1 or 2.0, will debut this December. Assuming the upgrade meets Indiana’s scalability needs, the university will begin deploying it as soon as it becomes available.

By 2009, many larger universities will likely deploy Kuali, according to former university financial officer David Lyons, now senior fellow of the National Association of College and University Business Officers. He notes: “We have to execute well on the remaining module developments, and we need some successful deployments to point to” (that is, in order for Kuali to reach a tipping point and gain widespread acceptance throughout higher education). That’s already happening with Sakai, which is a bit more mature than Kuali. “Right now at IU, we’ve reached a critical mass of users with Sakai,” says Wheeler. “We’re looking to retire our old system at the end of this summer.”

Indeed, the university has 90,000 users on Sakai, and 74 percent of all course materials have moved over to the new system. “We now have absolute control of our destiny,” says the CIO. “Under the old system, we could only innovate as fast as we could afford to.”

Indiana University’s open source environment continues to borrow heavily from peer universities. For instance, IU has deployed a wiki component from the University of Cambridge (UK) and podcasting code co-developed with the University of Michigan.

Whitman College (WA) has also hopped aboard the Sakai bandwagon, according to Mike Osterman, a middleware analyst at the college. Whitman’s Sakai deployment, known as CLEo (Collaboration and Learning Environment online), has been a steppingstone to gradually migrating away from a Blackboard implementation, according to Osterman. “We started looking at alternatives in 2004, and Sakai was pretty raw at the time,” he says. “But we knew it would give us the flexibility to develop our own enhancements without spending a chunk of change to license more software.”

Risk vs. Cost

Many advocates believe open source projects can save universities money that’s typically allocated for ERP licenses and deployments. But universities have to look beyond the dollars and cents for even more benefits, say the pundits.

“Open source should not be about bashing the commercial software vendors,” says Lyons; “there’s an ongoing place in the market for them. The open source movement is more about design: Not only are these new applications designed for and by people in higher education; there will be hundreds and hundreds of new eyes looking at the code and the capabilities to ensure we’re meeting users’ needs.”

Still, moving to open source has its risks. When commercial software hits a bump, universities can turn to software suppliers for emergency service and support. If an open source code base fails, it can be difficult to track down the specific developer who can find and fix software bugs. In truth, moving to Sakai was “definitely a bit of a leap of faith,” Osterman concedes. “We decided to do a oneyear pilot. If it didn’t work out, we agreed we’d go back to the drawing board and start over.”

Whitman College launched its pilot in the fall of 2005. The effort included a gap analysis project that identified core Blackboard-type capabilities that the open source system would need to support. The university included 10 faculty members in the pilot’s first semester, then expanded it to 16 faculty members in the second semester. “We wanted to keep it fairly low-profile and make sure each of the individual testers bought into the project,” says Osterman.

Smart move: Thanks to faculty testing and feedback, Whitman College was able to move CLEo into a production environment in the fall of 2006. Osterman says the system offers three clear benefits:

• Code transparency—an IT environment that’s truly open for enhancement without any need to negotiate with vendors.
• International reach. The CLEo project helped Whitman College to work more closely with leading institutions worldwide.
• Surprisingly strong reliability. Since larger institutions tested and deployed Sakai slightly ahead of Whitman College, Osterman’s team gained the benefit of their reliability work.

History Lesson

This journey to open source would have been unimaginable even a few years ago. For decades, universities have relied on completely homegrown information systems or complex commercial software to run their financial, academic, and development operations. But ERP and other commercial applications were frequently horizontal by design, and lacked deep vertical features written specifically for higher ed. Often, universities had to pay vendors extra fees for custom modules that assisted student recruitment, enrollment, and campus services. Even at a small college with fewer than 10,000 students, annual license and maintenance fees can cost $100,000 or more. “That’s a lot of money for bug fixes that should have been corrected before the products even shipped,” quips Ed Golod, president of Revenue Accelerators, New York-based technology consultants.

Indiana University balked at that approach in the 1990s and instead worked with American Management Systems (acquired by CGI Group in 2004), a major IT consulting system, to design and deploy a client-server system. The system proved reliable and scalable, but like many client-server systems that were written for specific operating systems and application programming interfaces (APIs), the software lacked modern features for the internet age.

While universities were wrestling with internal development efforts or commercial ERP deployments, the open source movement quietly began in the early 1990s. From the Linux operating system to the Apache web server, developers worldwide agreed to write and share code enhancements for the greater common good. Still, open source wasn’t an overnight sensation. Even the most promising projects required many years to fully shine. It took nearly a decade for Linux and Apache to gain critical mass on corporate servers. Apache is now the internet’s dominant web server and Linux commands more than 30 percent of the server operating system market, according to market research firm International Data Corp.

More recently, open source applications have gained traction in businesses and on university campuses. Zimbra, an open source e-mail provider, says its e-mail inbox software has been downloaded from the web more than 6 million times. Many of those downloads landed within university settings, according to a spokesperson for the company. Other increasingly popular open source applications include Centric CRM, MySQL, JasperSoft (for business intelligence), and SugarCRM.

Even Apple is benefiting from the open source movement. Apple’s Mac OS X code is based on Unix, a close relative of Linux. That makes it easy for programmers to write Linux applications that they can test and deploy on Mac OS X. In fact, in February, IBM unveiled an open source desktop initiative that will provide tools for writing applications that run across Windows, Linux, and Mac OS X.

Smaller vendors are also working to fan the open source flames. In February, Centric CRM and a dozen other open source companies formed the Open Solutions Alliance. The group aims to promote interoperability and enhanced support for open source applications. Major integrators including Unisys have lined up to support the alliance, though it’s unclear how soon their work will result in new integration hooks that universities can leverage.

Look Before Leaping

As the open source sector continues to grow, the ERP market continues to consolidate through numerous mergers and acquisitions, leaving Oracle, SAP AG, and other companies such as SunGard as the primary choices for such commercial ERP systems. To be sure, the remaining commercial options continue to offer clear benefits. Generally speaking, they are highly scalable, incredibly reliable, and secure. Plus, they are backed by thousands of integrators and valueadded resellers (VARs) that can service and support the deployments.

By contrast, open source solutions typically have only scattered support in the VAR and integrator community. Even market leaders such as Red Hat have only a few hundred Linux VARs across the globe—compared to tens of thousands of partners supporting Microsoft’s Windows Server, SQL Server, and related applications. And while server hardware suppliers such as IBM, Dell, Hewlett-Packard, and Sun Microsystems support Linux, that doesn’t mean they’ll be familiar with niche open source projects and modules that pop up across the globe.

KNOW YOUR OPTIONS

“Just about any database or CRM integrator worth his salt will be familiar with Oracle and Siebel solutions,” notes Golod. “But the open source sector is hit-and-miss when it comes to finding partners that can help you.” (In the higher ed space, rSmart is the leading vendor stepping forward to provide services to colleges and universities that want the assistance of a commercial vendor for community source implementations; for example, the company offers boxed versions of Sakai and Kuali as well as consulting help.)

What’s more, many open source systems don’t have all the features and functions that universities take for granted in their current commercial systems. And it can take many years to fill in those holes. Even the most vocal open source proponents sometimes struggle to get code upgrades out the door. Novell, provider of SUSE Linux, is still striving to enhance the operating system with high-end clustering functions that many Unix providers have offered for more than a decade.

Universities also can suffer if an open source project stalls or is aborted. “If you hitch your wagon to the wrong open source project and it loses steam, you could wind up with a dead-end technology,” says Golod. “At least in the commercial software market, you can watch a publicly held software company’s financial performance, in order to to estimate a product’s long-term viability.”

Still, commercial ERP and academic systems also have their downsides. Deployments can cost millions of dollars and frequently require years to complete. Universities that write homegrown systems face their own challenges, including the need to employ and train teams of programmers to perpetually maintain and enhance the code.

For the most part, universities will continue to rely on a mix of commercial and open source solutions. Kuali and Sakai, in particular, prove that there’s plenty of room on campus for open source.

::WEBEXTRA :: The next wave of Kuali: a student information system. Find out more at Campus Technology 2007 in Washington, DC, July 30-Aug. 2 :: Learn more about open source BI.