The recent Cloud Connect Conference in Santa Clara, CA was a very sobering experience for me. While I knew cloud computing was way overhyped, I thought that there was one or more standardards organizations that claimed ownership. I also thought that all the functional requirements and specifications done for grids, web services, and SOA (e.g. distributed management, federation, SLA requests and validation,etc) would not have to be re-invented and redone for clouds. Wow, that’ll be a huge undertaking.
It’s my belief that for at least the next five years, each cloud provider will define its own set of user interfaces, SLAs, performance parameters, security methods, etc. The more cloud providers, the more chaos and confusion will reign.
To a much lesser extent, it reminds me of my 1st experience in standards. In 1978, each X.25 Public Packet Switched Network had its own specification, which was loosely base on the 1976 CCITT X.25 recommendation. IBM authored a game changing paper that year, which showed it was not economically practical to build a single X.25 DTE (e.g. host computer or workstation) that would operate on more than one of the four X.25 public packet switched networks studied- Transpac (France), Datex-P (Germany), Telenet (U.S), KDD (Japan). As a result of that paper, ANSI X3S3.7 in the U.S. and CCITT internationally worked with great urgency to complete a truly global X.25 standard (the ANSI Standard referred to the CCITT recommendation) in 1980. I was part of that effort as both a U.S. contributor and CCITT SG VII/ WP2 secretary. I represented the U.S. in the ITU-T as a subject matter expert till 2003 (Optical Networks was my last project there).
Maybe such an urgent, accelerated standards effort is needed for (at least) the network aspects of cloud computing, e.g. UNI and NNI, SLAs and validation/compliance. I would’ve thought by now that the major players would’ve gotten together to create such an organization or combine several interested standards bodies/forums/alliances to make one.
While the Cloud Computing market is forecast to be very big by IDC, Gartner Group, etc, there seems to be a lot of confusion regarding the service delivery method. For example, the venture capital firm, the Sand Hill Group, has concluded that cloud computing represents one of the largest new investment opportunities on the horizon. At the Cloud Connect conference, a represenative of that firm asked IBM’s VP of Cloud Services Ric Telford what he thought: “I have no problem with those numbers (40% in three years; 70% in five) as long as you include the caveat, it could be any one of five delivery models.” So maybe we also need to define and standardize on those methods of delivering cloud applications to users,
What do readers think? Would love to hear your opinions. Here are a few comments from conference participants:
From Kevin Walsh of UCSD:
I think that establishing a unified cloud standards framework under the auspices of a standards organization such as the IEEE deserves further discussion. (I know it would be well received by my government customer.)
which is covered by this ISO section –
The ACM is also in the mix with a conference planned for early June. See
see the accepted papers thus far –
In general, I like the IEEE process, and the organization is well respective from my point of view. Their standardization process is mature. See
From Robert Grossman:
The url for the new IEEE Cloud Computing Standards Study Group:
I’ll send a separate note to Steve Diamond, who is coordinating it.
I am looking forward to following up with you regarding the virtual networks effort.
There is also RTF research group on virtual networks called:
Virtual Networks Research Group (VNRG).
From Gary Mazzaferro:
I enjoyed your comments about the need for cloud standard initiatives. I”m of the same opinion and have slowly been moving towards a more collaborative initiative. The largest challenge is balancing time and funding. I have little time because of the lack of funding for the project. 🙂 I do have an idea how to make it work and gain participation by the user community
The bigger message was that there is still much work to do in this area. With a ton of standards bodies emerging today, and vendors coming to market with their own unique APIs, it’s becoming difficult to have one voice.
If cloud is going to gain any kind of traction, let alone achieve the nirvana of the Inter-cloud, then we must have some level of standards in place to make it happen. As we’ve seen historically, not having standards in place has created challenges around interoperability, as well as vendor lock-in. The value proposition around cloud computing is negated if interoperability is not possible. It’s as simple as that. No ifs, ands, or buts.
NIST, a federal agency that has been instrumental in defining cloud computing, will take on an additional role as a central publisher of cloud use cases accompanied by a recommended reference technology implementation. “But the airing of strong use cases where a technology set is deemed suitable for a particular problem could lead to a specification for a standard, a NIST representative at the Cloud Connect show in Santa Clara, Calif., said Wednesday in an interview.”
Please feel free to leave a comment below or email me and I’ll include it
Here is a link to view what others have written about the Cloud Connect Conference I attended in Santa Clara, CA: