Cloud Computing Interoperability Forum (CCIF) is dead and doghoused

While I wait for my latest post to the Cloud Computing Interoperability Forum (CCIF) to be vetted for anything negative about Enomaly I thought I'd take a few moments to give you a quick update in the cloud standards debacle. You probably remember last year's scandal where Reuven Cohen was evicted (along with myself and an unknown number of other bloggers) from his ownthe cloud-computing Google Group by his own employee, "Director of Research and Development" Khazret Sapenov. Reuven implies that he was evicted for firing this guy and as we still haven't heard the full story he must be still paying severance.

Anyway Khaz subsequently cleared up any doubt about their intentions for the list by [ab]using it to boot strap the Cloud Slam '09 virtual conference which, at around 50 bucks a 'virtual' seat (in addition to $10k extracted from sponsors AMD & Data Synapse) is still a pretty penny for an event largely organised and delivered by others. Congratulations for pulling off perhaps the biggest cloud computing related scam of the last year (I don't include Dell's trademark attempt because they didn't get away with it, and neither will Psion with "netbook" if I have any say in it). (see update 4 below)

Meanwhile some six months ago, upset about the loss of his prized possession and unfazed by the community's need for a clear communication channel (which is still under-served to this day) Reuven announced the Cloud Computing Interoperability Forum, where he carries the self appointed title of "Instigator" (the same title he carries for the CloudCamps which are also largely organised by others like Dave Nielsen and Sam Charrington). The group, which has the lofty goal of "[enabling] interoperable enterprise-class cloud computing platforms" (though its mission is now largely concealed by puffery) has so far produced nothing but noise and there's no reason to believe it's going to change any time soon.

Most of that noise sprouts from Reuven's own blog posts which are diligently blasted to the group (in case we missed them in our newsreaders) only to go largely ignored, no doubt because they have been increasingly nonsensical of late (culminating in this week's Pirate Bay DDoS report promptly denounced as a hoax). Aside from abusing competitors and academics, they follow a general pattern of introducing unnecessary three-letter-acronyms, name dropping seemingly artibtrary standards like RDF, MPLS, IPv6 and XMPP (and then being totally overwhelmed when it comes to implementing them, preferring to "crowd source"; his doublespeak for "outsource"), or both at the same time.

To his credit Reuven's got lots of good ideas (and a balance of bad ones) but he's more about the talk than the walk... which is fine (consultants make a good living out of this), but this is not going to be the source of a coherent suite of standards (as evidenced by the whiteboard snapshot above)... at least not on the timeframe that we need them (i.e. yesterday).

It's not like I'm the only one who's losing their patience with the Cloud Computing Interoperability Forum either. Chris Marino summed it all up nicely in his Cloud Computing Nukes the Fridge thread last month after he "just couldn't take it any longer", which quickly turned into a discussion about the definition ([over]done last year) and arguments about grid vs cloud, eventually being summarised nicely:

Chris put my frustration with this group in words. The group started out
fine but never had a single thread of worthy discussion. I alluded to
defining the objective of the group in my last email but it never happened.
-snip-
Reuven is great to start discussions but he always talks from a 10000 foot
view or at a very low level about XMPP and IPv6. Nothing in between. I agree
with Drue, time is of the essence.

The following unanimous consensus was "[not to] try to reinvent the wheel here" on the grounds that "companies such as [Cisco, Intel, IBM and Sun] are really picky about all these details, covenants, governance, IP intermingling provisions, umbrella cross licensing of embedded intellectual property, etc etc, are land mines", and yet Reuven and his clan are still flogging a dead horse by "creating a self-sustaining organization" and promising "some big announcements coming soon" (that's a month ago now and still nothing). Who will run said organisation I wonder, and will Enomaly still feature prominently in everything it does (as is the case for both CCIF and CloudCamp)?

Three things clarified the situation a lot for me recently:

1. A post in which I offered up the Cloud Computing Community Wiki as somewhere to finally start fleshing out some standards (after it was suggested by somebody else) mysteriously vanished. Twice. Dave later suggested that Reuven "somehow deleted both of them" while moderating (claiming that "most people are moderated by default"). A likely story.
2. An incomprehensible diagram of the [in]famous Unified Cloud Interface (UCI) appeared with the help of Pat Wendorf, one of Reuven's minions, sporting Enomaly's ECP first and foremost alongside Amazon and Google. Wishful thinking guys, how better to force your broken, insecure Enomaly ECP "product" on us than get it written into standards created by an "open, vendor neutral, not for profit community".
3. Despite earlier unanimous consensus not to moderate the discussion, another post I sent to the list outing a spammer vanished while others (according to mail headers, including the spammers') continued to be delivered instantaneously.
   Jason Meiers (in my opinion just another trademark troll having recently obtained trademark #77484486 on "monitoring as a service") was claiming that "XMPP is too stuffy" while offering his own commercially-backed protocol Simple Event Management Protocol (SEMP) as an alternative (on the basis that it is not extensible while XMPP is!?!)
   He was called on it at which point he tried to take the discussion off list, but more revealing was the fact that a number of the "experts" on the list (including Reuven himself) fell for it hook, line and sinker. That is despite the referenced Google Code project being empty except for the front page teaser and the (soon to be deleted) Wikipedia spam page declaring it "a component of the Internet Protocol Suite" and other laughable aggrandisements.
   Jason aka CAM Solutions aka CloudRequest aka UtilityStatus aka Cloud Autonomics claims to have a commercial implementation available here and he has a customer too: himself!
   

In any case the state of affairs at the CCIF right now is a dog's breakfast and there's no evidence that there's any light at the end of the tunnel. Reuven will likely prompty retort claiming that they've still got secret announcements in the pipeline but if that's the case then why bother with the "open community" sham?

Fortunately all is not lost as there's a bunch of "real" standards organisations working on these problems (thanks Dave for the list):

• Distributed Management Task Force - This group will focus on virtual machine formats building on OVF, groups of virtual machines (deployment groups), and VM Mobility. Lets hope they can get constituencies from VMware ESX, Redhat KVM, Citrix XEN, Oracle XEN, Amazon AMI/XEN, to a place where people can write tools that can easily work with all these
• Network Centric Operations - This is a group of US Govt and Govt contractors who for a long time have ben pioneers of realtime, distributed systems use for Command, Control, Communications and Intelligence (C3I). A subgroup believes that Cloud Computing is important moving forward and would like to see standard interfaces to clouds. Some folks from SRI and The Aerospace Corporation are the ones particularly interested in Cloud.
• Object Management Group - The OMG does not have a Cloud effort proper, is my understanding. However, the NCOIC guys are having a "Cloud Interoperability Session" in DC which they approached the OMG to help them run, as OMG had a DC meeting coming up anyway.
• Open Cloud Consortium - The OCC was started by a group in Univ. of Illinios who are focused on large scale data mining and business intelligence problems. They are one of many Universities to leverage the National Lambda Rail project which is dark fiber which goes around North America to be used for high speed networking research. Cisco has created a 10G LAN infrastructure overlay on a wavelength on the NLR called C-Wave which OCC is using. Univ of Ill. have an open source cloud project called Sector with a distributed storage system called Sphere which they have placed on the C-Wave as a distributed cluster. They have done good work in researching more efficeint protocols for distributed fast high latency networks for which TCP is not great, their project is called UCI and is in wide use for these kinds of use cases.
• Open Grid Forum - Open Grid Forum has been around some time and is a formal standards body for development of grids and parallel programming models. I believe they have some realtionship with the Globus Alliance.
• Open Group - Open Group is the former X/Open, who now have enterprise software architecture tools and best practices and conferences. They have recently concentrated on SOA enablement for large enterprises and have an enterprise architecture called The Open Group Architecture Framework (TOGAF). Recnetly they have hosted speakers about Cloud Interoperability as Cloud and SOA from an enteprise perspective overlap.
• Internet Engineering Task Force - IETF produce the RFC's which underly the Internet. For cloud there are several relevant efforts including LISP (a mobile IP addressing proposal, needed for VM Mobility and XMPP, a protocol for clouds to talk to each other upon).
• Storage Networking Industry Association - SNIA has an effort going to think about Cloud Storage standards
• Liberty Alliance - a widely accepted SAML-centric identity and authentication industry association with good interface documents and reference open source. Probably a good source for Cloud identity and authentiaction approaches
• FreeIPA - An open source project which combines many of the "root" functions a cloud interoperability federation might need such as DNS, Certificate Authority, LDAP, and NTP. Leverages the trusty and fast Netscape/iPlanet LDAP server. Fedora community.

More likely though, standards that don't already exist (e.g. HTTP[S], XML, REST, OVF) will (initially) emerge organically from the market place in a darwinian process of natural selection. This is what I said in August last year in Cloud Standards: not so fast... and that's exactly what is happening today in the absence of a coherent standards movement. It's a worry that one of the standards we'll need is emerging as the EC2 API (mostly because Amazon still haven't made their minds up about whether they're going to defend this interface as their intellectual property) but it's better than nothing, and nothing is what we'd have if we waited around for CCIF. It's telling that Enomaly themselves are looking into adopting it rather than eating their own dogfood.

Not wanting to miss this particular boat CCIF have churned out yet another acronym (UEC2) complete with yet another Google Group, but it's been repeatedly pointed out to them that "we [don't need] an EC2 abstraction layer, since there's already an API for EC2 and libraries for it for the major development languages". Surely that should have been obvious, right? An abstraction layer that implements an API and backs onto... the same API? WTF?

In the words of Larry Ellison, "What is it? It's complete gibberish. It's insane. When is this idiocy going to stop?"

And what does CCIF get for wasting everyone's time for six months? A chance to join Enomaly in the cloud computing doghouse. Anyone still serious about standards let me knowjoin the group below and we'll see if we can't make up for lost time.

Update: Things are really starting to get ugly - who'd have thought all it'd take would be one arrogant spamming trademark troll!

Update 2: Turns out it takes 10 seconds to create a Google Group so I've created the "no bullshit" cloud-standards group to "get stuff done" and a standards wiki. No invites. No resumes. No dedicated websites. No moderation/censorship. No fancy graphics. No sponsors. No added salt. Just standards.

Update 3: @ruv thinks it's personal. It's not and never has been. I'm sure he's a nice guy and I'm not one for ad hominem attacks anyway. I just want results and if CCIF and Enomaly are (even unintentionally) getting in the way of Open Standards and Open Source respectively by dragging their feet then I'm happy to be the one to say what (as it turns out based on the feedback) many others were already thinking.

Update 4: One of Khaz's co-conspirators felt that it was inappropriate to tar them all with the same brush and appealed for me to retract the "scam" statement. Fair call, and tangential anyway so retracting.