Standards: Some thoughts on the Open Cloud Consortium

The Open Cloud Consortium is a recent effort which aims at standardizing (part of) the cloud computing sphere. Here are some thoughts on how the OCC is organized, what they can offer and what they need to struggle with.

The OCC was initiated by universities in mid 2008 and currently has a very limited amount of members (3 full, 2 associates and 6 contributors). Yahoo and Cisco are the two most notable participants. That Cisco is a member (first industry participant to join in January 2009) of the OCC is not very surprising, they have a large stake in being able to provide interoperable cloud infrastructure to cloud providers. This is probably the reason why they were one of the signees to the Open Cloud Manifesto as well.

The OCC has an academic approach and strives to pursue the following goals:
1. Support the development of standards for cloud computing and frameworks for interoperating between clouds;
2. Develop benchmarks for cloud computing;
3. Support reference implementations for cloud computing, preferably open source reference implementations;
4. Manage a testbed for cloud computing called the Open Cloud Testbed;
5. Sponsor workshops and other events related to cloud computing
The development seems to be implementation driven with both point 2, 3 and 4 relating to implementations of technology. This is a sound practice for an area which is still very much under development and not clearly defined.

The membership structure is a variant which is consistent with the recommendations from Andy Updegrove on consortia formation. There are four membership classes with varying rights and fees. The two highest tiers will be allowed to vote on any standards developed and only the highest tier has operational governance. The membership allows for individual participation and free-of-charge participation for some organizational classes. OCC will use a two thirds majority procedure for approval processes. The general structure and processes of OCC are lightweight and streamlined, allowing for a smooth operational process. OCC also creates possibilities for individuals to participate which is good since open source development to a large part rest on individual contributions.

At the moment the biggest threat to the organization is that there might be little interest from the industry to participate in cloud standardization efforts and in order for standards to be successful a larger participation and higher degree of consensus is needed. OCC is also hurt by the fact that there is no consensus of what cloud computing relates to. OCC is currently very open and might need to come up with some “members only” functions to increase the strategic importance of participation. With a fully open process there are risks of free-riding and no central motivation for participation.

The OCC testbed is the most interesting feature the organization currently provides. Leveraging use of the testbed might interest additional participation as it allows for a unique testing environment which, according to the OCC, does not have any similar alternative elsewhere.

The OCC stance on interoperability is evident from the stated goals. Both using and developing open source products will probably be the main safeguard to ensure interoperability. As a young organization OCC has not yet have had to consider finer aspects of interoperability in products. One possible additional activity for interoperability is to use workshops to promote and display interoperability. Another natural addition would be to supply open source testing software for interoperability in cloud standards.

OCC will develop software distributed under OSI open source licenses. As of yet there is no information on whether there will be one or several OS license choices available. This is intended to attract academic institutions and individuals to the organization. This might also draw some interest from the open source movement which could create traction for the OCC. An open source solution for software is a good idea since it will allow for efficient scaling properties, which is one of the central tenets of cloud computing. Some more clarity on the form of open source licenses to be used would be useful, different licensing models have important implications for the cloud providers which use the software. A traditional tri-licensing scheme (MPL, GPL, LGPL) could be a good starting point.

To sum up, OCC is an interesting and well set up organization which focuses on problems of the future as much as of today. OCC might however be before its time. There is only a small membership which lead to consensus problems for standards, there is currently no definition of the cloud which makes standardization all the more tough and the industry is currently not all that hot on creating standards in the first place for cloud computing as evident from the Open Cloud Manifesto aftermath.