[occi-wg] Fwd: One Cloud Standard to Rule them All

[Sam Johnston]

FYI, many of you will receive this directly from the CCIF but I'm posting it
here too in order to give you an update on progress. With related
discussions taking place on other lists (e.g.
rest-discuss<http://tech.groups.yahoo.com/group/rest-discuss/>,
ietf-http-wg <http://lists.w3.org/Archives/Public/ietf-http-wg/>,
atom-syntax <http://www.imc.org/atom-syntax/>, etc.), on the weekly
conference calls and other out-of-band channels it can be difficult to see
that progress is being made but I can assure you that it is.
As soon as we get the documentation in line with the latest discussions
we'll crank the list back up as I believe most of us are back from summer
vacations now (the center of gravity of the working group happens to be in
Europe where the months of July and August tend to be useless for
productivity). Then we'll have to get to work on our PR as we've been
holding back while working on having something to talk about - in particular
I think the blog will be front and center and "guest" [re]posts will be
gratefully accepted.
Sam

---------- Forwarded message ----------
From: Sam Johnston <samj at samj.net>
Date: Fri, Sep 11, 2009 at 1:10 PM
Subject: Re: One Cloud Standard to Rule them All
To: cloudforum at googlegroups.com


On Wed, Sep 9, 2009 at 12:21 AM, Greg Pfister <greg.pfister at gmail.com>wrote:

>
> +1 on
>
> "at the end of the day, cloud standards seem to
> have little to do with customer requirements but instead more to do
> with marketeering and positioning for market dominance."
>
> That certainly aligns with my prior experience in several industry
> standards groups. A key goal of participants is always to steer new
> standards as close as possible to their own company's product roadmap.
> The normal order of the day is Intense politicking, behind-the-scenes
> maneuvering for votes, and whatever dirty tricks are needed to
> accomplish that alignment.
>
> If you don't get the desired result, you return home from standards
> meetings with a lot to answer for.
>

Dilbert captures this as only Dilbert can, and it's hard to believe it was a
coincidence for this to have come the day after the vCloud API annoucenment
last week:

http://www.dilbert.com/strips/comic/2009-09-02/

FWIW I'm fervent about openness and transparency and am 100% committed to
Open Cloud and the Open Cloud Computing Interface (OCCI). I'm also not
aligned with any of the vendors, nor do I have my own "horse in the race" so
to speak, rather I'm funded by strategic consultancy with large enterprises
who share my passion for openness (or should I say, distain for vendor
lock-in). For some background there was an article in Silicon (Germany) last
week which is translated to English below and there should be more like it
real soon now (note I wasn't involved in the writing of this article so
there's one or two gotchas - in particular "Open Cloud" in no way requires
open source).

My point is that all is not lost. OCCI's been a while in coming following a
change in strategy early on (I had planned to follow Google's example by
using AtomPub but it turns out people prefer a choice of formats) but it
will be here very soon and I can assure you it will be a best-of-breed API
as close as possible to the "universal interface" of HTTP. Not only have we
conducted a formal use case and requirements gathering process (the results
of which will be available for public comment soon) but we have had a number
of APIs submitted (hopefully soon to include vCloud) and are taking the best
parts from each rather than basing our development on any one API. The
process is backed by an established standards development organisation (OGF)
but is completely open - anyone can participate to the discussions and
contribute to or comment on the documents (which are now in a Google Code
repository <http://code.google.com/p/occi/source/browse/#hg/docs>). You can
also contribute extensions and new formats (we've done some work on Atom,
XHTML5, JSON & text renderings already) within the official spec or
independently - if you need protocol buffers for performance reasons for
example then you're welcome to write it up (assuming we don't get round to
it).

There are now 200 people in the working group, of which a number are
representing various vendors and academic projects on a formal/funded basis.
In the coming days we'll have a coherent specification and will need all the
help we can get in refining it so if you haven't already then please join
the list <http://www.ogf.org/mailman/listinfo/occi-wg> and get
involved<http://forge.ogf.org/sf/wiki/do/viewPage/projects.occi-wg/wiki/GettingInvolved>.
The OGF 27 event is next month and here's hoping we can get it to a publicly
presentable state by then.

Sam

http://www.silicon.de/software/business/0,39039006,41502041,00/open+source+macht+cloud_nutzer+frei.htm

*Open Cloud Initiative falls on fertile ground in Germany
*
by Ludger Schmitz*

Most concerns about cloud Computing are about security and governance. But
will a customer of a cloud stay free? Some "openness" is not enough. A
solution could be open source cloud computing. In Germany this seems to be
on the way.

If a corporation consolidates his globally widespread IT capacities in a
so-called „private cloud“, first of all there is talk about virtualization –
and VMware not far away. A „public cloud“ is called, where any company can
use an external IT infrastructure. Known vendors in this area are Amazon,
Google, IBM, Microsoft and Oracle, to name a few. In addition, several
smaller vendors are trying to attach to the trend Sooner or later a wave of
consolidation will sweep through this yet forming market. In a foreseeable
future there will be less freedom of choice for the customers.

There's even less freedom when it comes to applications. Within a cloud
there are almost no competing offerings. Today's clouds are exclusive
circles, open only to those who are willing to accept the respective rules
and – even more important – the limits. Cloud computing rises the dependence
of the customers to a level that reminds of the eighties. Last year Dell put
it to the extremes by trying to get the trademark for cloud computing.

Many potential customers fear that cloud computing, though marketing tries
to associate it with freedom, might push them back into former dependencies.
Sun Microsystems seems to have realized this danger – and consequently
called it's own offering Open Cloud. Actually in this offering open source
software had a key role. As it has at Google, but here the APIs of the
Google Application Engine form the cage.

Early this year there were extensive discussions among some vendors, how
much freedom would be necessary to keep the cloud offerings attractive. The
discussions resulted in an “Open Cloud
Manifesto<http://www.opencloudmanifesto.org>”,
published on April 1st, 2009. Among it's principles is the sentence: “Cloud
providers must not use their market position to lock customers into their
particular platforms and limit their choice of providers.” Therefore
standards should be used and if needed be developed in a common effort.
Every work should be under the imperative of customers needs.

Many vendors signed the manifesto, among them several from the open source
community with Red Hat and Novell on top. Of the potentially most important
players at this time IBM, SAP, Sun and VMware stood behind the manifest.
Amazon, Google, Microsoft and Oracle are missing. Microsoft criticized it
had not been involved in preparing the document and was only asked to sign
it. In an open source community this procedure would not be unusual as
always someone starts something, other get involved and later influence a
project by their activity. But the manifesto was not only made for the open
source world.

At just about the same time Sam Johnston <http://samj.net>, an Australian,
analysed the various forms of clouds and came to four different ones: The
first, the “Closed Cloud” is not open accessible and usually referred to as
private cloud. The second is the “Open Cloud” with open APIs and open
standards. Somewhere between these two categories today's cloud offerings
can be located. As a third form there could be talk of an “Open Source
Cloud” where open source software is used beyond open standards and APIs.
Where additionally even data is open this would form the “Free Cloud”.
Wikipedia gives an idea.

By using precise definitions Johnston goes more into details than the Open
Cloud Manifesto. According to him open APIs mean that all relevant functions
must be implemented on the basis of open standards for open interfaces.
Userdata and even metadata must be available in open standard formats. In an
Open Source Cloud only software may be used that is under a licence approved
by the Open Source Initiative (OSI). And in the free cloud form the data
must be under an open knowledge licence.

When it comes to the central question, what open standards are, Johnston
goes once more deeper into details than the vague manifest. He demands that
an open standard is in all details public and can be used at no cost.
Therefore copyright fees are not acceptable. Similar with patents that must
be irrevocable available free of charge. Trademarks may not be used in a
discriminating way. Last not least all implementations of an open standards
must have a OSI-accepted licence, so be open source – or public domain.

With this last point Johnston sets fire to what he has described before as
Open Cloud. He seems to do so willingly, because an Open Cloud is obviously
not enough, neither for him nor for others in the open source community.
This became apparent on April 6th, 2009 when Johnston published an “Open
letter to the Community <http://www.opencloudinitiative.org/openletter>” to
introduce the “Open Cloud Initiative <http://www.opencloudinitiative.org>”.
It's abbreviation is only OC to prevent confusion with the Open Source
Initiative <http://www.opensource.org> (OSI).

The OC has several aims: It wants to ensure that the expression open cloud
is not diluted and the customers under this label get basics freedom rights
like the one to access their data in open formats via open interfaces at any
time. As guard of the open cloud the initiative also wants to keep the cloud
market open to all vendors and the entry barriers low. Finally the OC wants
to make sure that in an open source cloud the principles and definitions of
the OSI are respected.

According to some insiders the ideas of an Open Cloud Initiative fell on
fertile ground in the German open source community. There seems to be
massive work at a reference stack of open source solutions, that are needed,
integrated and reliable connected for an open source cloud. The work is not
done yet, quite a lot technical problems still need to be overcome. But
using open source software the companies committed to open source cloud
computing have some advantages: They work with readable source code, with
open interfaces and standards. It looks as if they are about to present a
cloud offering, that would be far more extended, flexible and free than
anything a proprietary vendor can offer as cloud. This would change the game
in the cloud market.

An open source cloud at this time would be a milestone in the open source
history. Because up to now open source itself was a trend but lagging years
behind proprietary products for other IT trends. This time it could turn
into a neck-and-neck race.

*Ludger Schmitz is a freelancer in Munich, Germany, specialized on open
source developments.
Contacts: l.schmitz [at] open-source-presse.de
Translation German to English assisted by Michael Kromer, Millenux GmbH,
Munich.

Greg "experience breeds cynicism" Pfister
> http://perilsofparallel.blogspot.com/
>
> On Sep 5, 1:14 pm, Reuven Cohen <r... at enomaly.com> wrote:
> > Lots of discussion recently on the the topic of Cloud standards and a
> > potential Cloud standards war emerging. Thought I'd give you a quick run
> > down.
> >
> > In an article written by Tom Nolle for Internet Evolution he asks if
> Multiple
> > Standards Could Spoil Cloud
> > Computing?<
> http://www.internetevolution.com/author.asp?section_id=561&doc_id=181129>.
> > In the post he says " too many standards are worse than no standards at
> all,
> > because these efforts can stifle innovation and even implementation. In
> the
> > case of cloud computing, there’s also the big question of whether
> standards
> > being pushed for private clouds will end up contaminating the successful
> > Internet model of cloud computing."
> >
> > Tom also gives some love to my Unified Cloud Interface
> > concept<
> http://www.elasticvapor.com/2009/02/semantic-cloud-abstraction.html>saying
> > "The
> > best hope for a unification is the Cloud Computing Interoperability
> Forum.
> > Its Unified Cloud Architecture
> > <http://code.google.com/p/unifiedcloud/>tackles standards by making
> > public cloud computing interoperable. Their map
> > of cloud computing shows the leading public cloud providers and a
> proposed
> > Unified Cloud Interface that the body defines, with a joking reference to
> > Tolkien’s *Lord of the Rings*, as “One API to Rule them All". So then,
> why
> > not have that “One API” rule private clouds, too? A single top-down
> vision
> > of cloud computing for public and private clouds has to be a better
> > approach." -- amen brotha.
> >
> > (As a side note, I'm still committed to the concept of Semantic Cloud
> > Abstraction and the Unified Cloud Interface Project
> > (UCI)<http://code.google.com/p/unifiedcloud/>,
> > but lack the time to do much with it given all my various commitments. So
> if
> > you're interested in helping do something more around the concept of
> > semantic cloud API's, please feel free to get in touch. Currently I would
> > describe the UCI as a science experiment at best.)
> >
> > Elsewhere on the standards front, on Friday the OpenNebula project
> announced
> > that they had made available a prototype implementation of the OCCI draft
> > specification <http://www.occi-wg.org/>. In case you're not familar with
> > OCCI, it's a simple open API specification for remote management of cloud
> > computing infrastructures focused on multicloud interoperability. In the
> > note to the OCCI mailing list they stated that they believe that it is
> > important to have an implementation in order to demo the standard, which
> > they think will provide value by being able to show people how the
> standard
> > will function in a real world environment. The prototype includes a
> server
> > implementation, client command for using the service and enabling access
> to
> > the full functionality of the OCCI interface.
> >
> > Some other interesting recent cloud standards commentary includes a
> > post by Simon
> > Wardley with an assertion that a standards war is now in full
> > swing<
> http://blog.gardeviance.org/2009/09/cloud-computing-standards-war.html>.
> > In the post Wardley outlines several considerations for cloud standards
> > including;
> >
> >    - A specification can be controlled, influenced and directed more
> easily
> >    than an open source project.
> >    - A specification can easily be exceeded providing mechanisms of
> lock-in
> >    whilst still retaining compliance to a *'standard'*.
> >    - A specification needs to be implemented and depending upon the size
> and
> >    complexity of the *'standard'* this can create significant adoption
> >    barriers to having multiple implementations.
> >    - Open source reference models provide a rapid means of implementing a
> *
> >    'standard'* and hence encourage adoption.
> >    - Open source reference models provide a mechanism for testing the
> >    compliance of any proprietary re-implementation.
> >    - Adoption and becoming de facto are key to winning this war.
> >
> > Christopher Hoff also chimes in
> > saying<http://www.rationalsurvivability.com/blog/?p=1316>"A Cloud
> > standards war? War is such an ugly term. It’s just the normal
> > activity associated with disruptive innovation and the markets sorting
> > themselves out. The standards arena is simply where the dirty laundry
> gets
> > exposed. Get used to it, there’s enough mud/FUD flinging that you can
> expect
> > several loads"
> >
> > Personally I would say that at the end of the day, cloud standards seem
> to
> > have little to do with customer requirements but instead more to do
> > with marketeering
> > and positioning for market dominance. If there truly is a cloud standards
> > war emerging then the way it will be won will be based solely on the
> > standard / platform with the broad adoption. The technology (API,
> Platform,
> > etc) with the broadest market penetration will ultimately win -- this is
> a
> > certainty. Although personally I would prefer to see the "most open" and
> > interoperable approach win, lets be realistic, the winner will most
> likely
> > be the one with the largest share of cloud computing revenue.
> >
> > --
> > Reuven
> > CCIF Instigator
> --~--~---------~--~----~------------~-------~--~----~
> You received this message because you are subscribed to the Google
> Groups "Cloud Computing Interoperability Forum (CCIF)" group.
> To post to this group, send email to cloudforum at googlegroups.com
> To unsubscribe from this group, send email to
> cloudforum+unsubscribe at googlegroups.com<cloudforum%2Bunsubscribe at googlegroups.com>
> For more options, visit this group at
> http://groups.google.com/group/cloudforum?hl=en
>
> -----
> Join our Twitter Group at www.twitter.com/cloudforum
> Or Our Linkedin Group at http://www.linkedin.com/e/gis/927567
> -~----------~----~----~----~------~----~------~--~---
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/occi-wg/attachments/20090911/76547b7a/attachment.html