OVF 1.0 and beyond

OVF 1.0 just got released as a DMTF standard. Here is the specification and its companion white paper. After a quick scan I didn’t see any major change from the submitted version, which is consistent with the content of the “preliminary standard” from last year.

The interesting question is what comes next, especially with regards to VMWare’s vCloud. The VMWare press release stated that “as one of the original authors of the Open Virtualization Format (OVF) standard now released from the Distributed Management Task Force (DMTF), VMware will build upon that work by submitting a draft of its VMware vCloud API to enable consistent mobility, provisioning, management, and service assurance of applications running in internal and external clouds” and Drue Reeves at the Burton group commented on this (Drue, we’re still waiting for part II). I see no reason to believe that VMWare is going to stop playing by the Microsoft playbook in DMTF as it appears to be quite successful so far (I’ll pat myself in the back for predicting over a year ago that “OVF might only be the beginning” for VMWare at DMTF).

This results in what looks like a landgrab from DMTF in Cloud standards. Meanwhile, in Washington DC yesterday, the Strategies and Technologies for Cloud Computing Interoperability (SATCCI) workshop took place. At this point all I know about it is the report from Reuven Cohen that I just read (hopefully Stu, Krishna and other bloggers who participated will provide additional perspectives). From Reuven’s report, Winston Bumpus (Director of Standards Architecture at VMware and President of the DMTF) described OVF as “an ideal cloud migration and deployment package”. Which may be true but is a pretty recent repurposing (the spec and the white paper don’t even mention this application). And while the DMTF is going full speed ahead on this, Reuven reports that “Craig Lee, President of the Open Grid Forum suggested that we need to take more time to examine the overlap between various standards groups, mapping the opportunities for collaboration”. Sure thing. The old timers might remember that when the DMTF decides to run with Microsoft’s WS-Management it wasn’t just OASIS (where WSDM was created) that eventually got hosed but also OGF (then called GGF) which relied on the WSRF/WSDM stack. At the time too there were discussions to identify and reconcile the overlap, for all the good they did (disclosure: I have some history there).

We’ve seen this in the WS-* game before. At the end it’s not so much a matter of what the standards bodies do (and even less of what they say), it’s a matter of what the big players do and where they choose to take their marbles. To the extent that you can separate the two, which becomes tricky in the case of vendor-run bodies like WS-I and DMTF. As I have written before, “at the end, it comes down to what [you think] a standard should be”.

[UPDATED 2009/3/26: Stu has now written a report on the SATCCI meeting.]


Filed under Cloud Computing, Conference, DMTF, Everything, Grid, IT Systems Mgmt, OVF, Portability, Specs, Standards, Utility computing, Virtualization, VMware, WS-Management

5 Responses to OVF 1.0 and beyond

  1. William,

    Sorry for the delay on part two. The Cisco announcement and IBM/Sun thing took over the blog for a while. Plus, Burton Group is launching our new cloud service in a few weeks…so we’ve been bee-like in prepping our cloud content.

    But — I will have it out today.

  2. There is a lot of public references from major industrial players stating their interests in using OVF for clouds. Apart from the vCloud API reference you mention, the Citrix partnership with rPath to “extend Kensho [the flagship OVF-related public Citrix project] to support the deployment of OVF appliances in infrastructure clouds, starting with Amazon EC2” and the Oracle and Intel collaboration to “extend standards that enable portability of virtual machines images such as Open Virtual Format (OVF) and also create Web services standards for provisioning and management of cloud-based services” are other examples.

    However, I feel that these are not more than “declaration of intencions” surrounding the current hype in cloud computing and OVF. We should take into account that OVF was not designed with cloud computing in mind and that there are some issues to solve when is applied to describe services to be deployed in clouds. For example, OVF is designed for static deployments, so how it could be adapted to elastic clouds (where the number of virtual machines grows/shrink following the service demand)? What about self-configuration (the standard OVF self-configuration is based on the fact that configuration parameters are known in pre-deployment time, which is not true for clouds, e.g. the user issuing the OVF won’t know the IP address assigned to VMs in the cloud)?

    The bottom line is that I think it could be much more interesting to have technical work from the industrial player on these and other topics related with adapting OVF to cloud, instead of “press-ware” about alliances and intentions :)

  3. Good points Fermin. Most people who latch on to OVF don’t necessarily understand its capabilities and limitations, especially as applied to Cloud situations – which is not what it was designed for. They just want to make a statement of intention to support standards, and OVF is what is the most convenient to use for this. Of course that’s a great situation for those who understand and drive OVF.

  4. Pingback: William Vambenepe’s blog » Blog Archive » DMTF calls the ball on Cloud standards

  5. Pingback: William Vambenepe’s blog » Blog Archive » Cloud API: what’s cooking between IBM and VMWare?