MOC-up: Cloud Testing or Cloud Transformation?
How much does an "open” cloud cost? The Massachusetts state government is willing to invest $3 million in Boston University's (BU) concept of a truly open cloud, one the institution hopes will spur positive change marketwide.
Can this initiative really compete with offerings from big providers like Amazon, Google and IBM, or is it just another rebranding of cloud hype?
The Massachusetts Open Cloud (MOC) is a joint venture of BU, Harvard, UMass Amherst, MIT and Northeastern University, with test space provided by the Massachusetts Green High-Performance Computing Center (MGHPCC) and Oak Ridge National Laboratory (ORNL). If that sounds like a lot of academic firepower, it is — and it's expensive. In addition to the $3 million promised by Governor Deval Patrick to develop the MOC, the project is also set to receive $16 million in funds from federal, industry and private sources, according to HPCWire. But what sets the MOC apart from other public clouds?
It starts with the concept of "public" resources. Imagine popular commercial clouds as an inverted funnel; at the top is a single provider that owns and manages all infrastructure and resources. At the bottom are consumers, who have on-demand access to these resources but don't control their implementation. But, as noted by a BU white paper, the MOC concept relies on what's known as the Open Cloud eXchange (OCX), where “many stakeholders, rather than just a single provider, participate in implementing and operating the cloud."
In other words, the funnel expands into a cylinder of equal width at both ends; just as multiple customers can use the same cloud resources, multiple providers can leverage the same infrastructure. A recent BU Today article describes this as a kind of cloud computing mall, where providers are essentially storefronts, all chipping in to maintain a larger public space.
The ultimate goal of the MOC is to provide cloud resources as a utility that is analogous to power or water. Pioneers like John McCarthy and Douglas Parkhill predicted this kind of utility evolution more than 50 years ago, but the current market of public clouds actually serves the opposite purpose by forcing companies to rely on a single type of cloud power.
Slow and Steady
While the MOC project enjoys backing from vendors like Cisco, Red Hat and EMC, other large technology players are going in a different direction. IBM, for example, recently launched their Cloud Marketplace, which includes access to IBM resources along with selected offerings from partners and third-party vendors. Think of it as MOC-light — a wider funnel but still controlled by a single entity.
Many vendors are also trying to capture a larger section of the cloud market by deepening their "as a service" offerings; according to a recent Search Data Management article, the database-as-a-service (DBaas) market is set to hit $1.07 billion this year, and $14.05 billion by 2019.
Even education vendors, such as CompTIA, are starting to see the value in open offerings, giving prospective members a no-cost channel to access training materials, research and news.
What's the Holdup?
It's tempting to see industry reticence as nothing more than dollar-grabbing, but are there legitimate concerns about moving to a MOC-like model?
One possible problem is security. A recent ZDNet article reports that brute-force attacks on public clouds jumped from 30 to 44 percent in the last year as traditionally on-premises attacks "followed" the wave of public cloud adoption.
In the MOC's cloud "mall," attackers could target not just consumers but also multiple vendors, and potentially force total closure. If, however, BU and its partners can effectively leverage the accountability enjoyed by other open-source projects, consumers will likely force the hand of major vendors.
Bottom line? The MOC has the potential to transform public cloud computing by mimicking familiar retail experiences, delivering utility-grade resources and remaining accountable to stakeholders.