The Inevitable Unbundling of Software and Support
Recent months reflect an important discussion of open source application software
for higher education. The success of Linux and Apache at the infrastructure
level and the generally collaborative nature of universities prompted Gartner
to forecast that higher education would be one of the early places for open
source applications, such as course management systems, portals, and ePortfolios.
The classic “Build vs. Buy” decision has now been augmented with
open source’s “Borrow” option—borrowing without an expectation
of repayment. Some argue that the lack of commercial support is open source’s
Achilles heel. Conversely, I assert that the emerging unbundled support model
for open source applications is actually a feature, rather than a bug for higher
Unbundling is inevitable. Traditional models for software have offered a for-fee
license to use proprietary intellectual property (IP) bundled with for-fee support
provided by the owner of the IP. Unbundling creates two distinct markets for
software and support. Thus, the following points merit timely scrutiny:
n Unbundling is a general economic trend that offers greater efficiency.
n Recent open source collaborations have triggered unbundling for higher education.
n Markets segments will choose.
Economic Trends Favor Unbundling
Unbundling is a general, evolutionary economic trend favoring greater marketplace
choice and efficiency. Wecan observe this trend all around us. Few of us buy
auto insurance from our car’s manufacturer or telephone repairs from the
phone company. We plan our travel with à la carte choices for rental
cars, hotels, and airfare. We buy our mobile, local, and long-distance phone
service from different providers.
Unbundling d'es not negate the possibility of offering bundles—note
the travel industry’s recent attempts to rebundle complete trips or the
telecom provider’s efforts to package all of your phone services—but
the options for unbundled choices remain plentiful because they are economically
It is tempting to dismiss these arguments by asserting that complexity in
large application software systems necessitates support from its creator. It
is indeed difficult to support a black box without knowledge of its inner workings—which
is today’s current situation with proprietary software.
Open source, however, changes the situation considerably. The inner workings
of the software are visible—and changeable—by anyone who chooses
to understand them. For-profit companies can build competencies in understanding,
supporting, and extending open source software to offer for-fee services without
holding proprietary license over the IP itself.
But will this unbundling work? Do these “IP-less” companies really
exist? The success of uPortal with well over 170 adoptions and multiple Open
Source Support Providers (OSSPs) demonstrates that unbundling works. The
r-smart group’s support for the Open Source Portfolio Initiative’s
Release 1 is another example. No less than a dozen well-funded open source application
projects in the past year ensure a healthy stream of open source products for
These include Sakai, Fedora, Navigo/AAM, ePortfolio, VUE, LionShare, and others—all
written for education, by education. Many of these will provide support markets
for OSSPs to achieve needed economies of scale. Projects that use the OKI Open
Service Interface Definitions (OSIDs) or the forthcoming Sakai Tool Portability
Profile will also help OSSPs build solid economics in support.
Not all open source projects will succeed, and not all OSSPs will develop
the competencies to compete profitably in this new era. Some administrators
perceive this as imposing greater risks on universities than the current bundled
model. I contend, however, that unbundling actually hedges a university’s
risk as it has more options. It can switch OSSPs, develop in-house competence
if OSSP pricing becomes unfair or they fail, or even leverage an OSSP relationship
to switch to different open source code should that ever be necessary.
For example, when Red Hat recently moved to increase academic support fees
and impose migration timelines, one major university dropped Red Hat’s
support services and contracted with a different vendor to provide security
patches and updates for Red Hat distribution at a considerably better price.
For application software, consider the actual lower risk of navigating among
unbundled options versus the costs of trying to switch among the leading bundled
application providers that pervade the marketplace today.
Market Segments Will Choose
Finally, assertions of open source versus commercial with only one winner are
overly simplistic and distracting. Open source and OSSPs are a complement to,
not a wholesale replacement of, traditional bundled software models. Open source
creates a continuum of IP and support choices beyond the traditional discrete
choices of “buy it with bundled support” or “build it and support
it all yourself.” Market segments will choose among open source, blended
(OSSP bundling an OS product), and traditionally bundled value propositions to
best suit their needs.
Bundled propositions from vendors—based on proprietary, open source
IP, or even a mix—will clearly prevail in some segments. Some universities
will buy a bundled open source application with a support package from an OSSP
while others will choose no support or only ala carte services from an OSSP
and do it themselves. Others may decide that the current bundle of proprietary
IP and vendor support serve their needs as well.
Unbundling is good for higher education. It will sharpen the value propositions,
increase choice, and provide a healthy set of options for serving our dual challenges:
working within sustainable IT economics while addressing the requisite frontiers
Wise administrators already reject the flawed assertion that open source will
make software free. Likewise, they should also reject the equally flawed view
that IP and support are only viable as a proprietary bundle. Now is the time
to align an institution’s IT strategy to take advantage of new options
for software and support.