Like many people, I did not think highly of the Open Cloud manifesto episode last week and now that some resolution seems to have come of it with the apology from CCIF , its time to leave the CCIF manifesto episode behind to the wider issues.
I have been trying to take a pragmatic view about the current and future role of standards for some time, albeit from the perspective of mobility where the issue has existed for a while now.
Standardization for the Cloud offers an opportunity to rethink standards themselves since many of the older paradigms do not apply in the new world dominated by the Cloud(or rather they become irrelevant/commodities).
So, much like the parable of new wine in old wineskins, there is little point in framing new discussions in context of the old.
Let me explain ..
In a widely publicised article, Larry Ellison – founder of Oracle corporation – criticised Cloud computing calling it meaningless. And Cloud computing is a trap warns Open source pioneer Richard Stallman.
Both Ellison and Stallman have the same thing to fear from Cloud computing – it makes them commodities.
For instance – when we use Google, we all become users of Linux(Open source) since Google servers run on Linux – but we don’t care that Linux(i.e. Open source) is used(and nor should we!).
With the Cloud, all you care about is the Service(The Service level agreement). Once the customers(business or consumer) have a SLA(service level agreement), who cares what database it runs on or if it uses Open source?
Essentially, if we want to enhance the uptake of Cloud computing – (and the precise definition of Cloud computing itself still remains a bit clear ..), there are four issues to consider(happy to add to this list)
a) Process level interoperability: i.e. a process or object on one cloud should be able to invoke a process or object on another cloud from a different vendor
b) Service level agreements – i.e. the relationship with the customer
c) Legal issues i.e. the legal implications of having data in the Cloud and
d) Data level interoperability
Depending on how cloud service providers set up their technology, they can make it harder or easier for applications living in one cloud service to interact with those living in others, or for their customers to shift to a different service provider.
What does this mean and more importantly How does it apply to the Cloud?
Process level interoperability (a) is hard and almost no one will be able to do it(example -Google’s cloud seamlessly calling a process/object on Amazon’s cloud). In contrast, Data level interoperability has always been possible. I used to work for PeopleSoft(now Oracle corp) – and migrating ERP systems – while not easy – was possible.
Tim O’Reilly has done some excellent thinking in this space which I have been following for some time (and I was truly disappointed that this was not linked by the Open Cloud Manifesto folk)
Tim O Reilly, speaks of the idea of ‘Reasonably open’ i.e. a more pragmatic view
This isn’t just a “moral” appeal, but strategic advice. The first provider to build a reasonably open, re-usable system service in any particular area is going to get the biggest uptake. Right now, there’s a lot of focus on low level platform subsystems like storage and computation, but I continue to believe that many of the key subsystems in this evolving OS will be data subsystems, like identity, location, payment, product catalogs, music, etc. And eventually, these subsystems will need to be reasonably open and interoperable, so that a developer can build a data-intensive application without having to own all the data his application requires. This is what John Musser calls the programmable web.
I think thats the kind of thinking we need ..
When I spoke at CEBIT I also mentioned the idea of a playframe instead of a cage(which was originally said by Robert Hamilton of Google in the session I chaired previously in Mobile World Congress – Barcelona called Business of Mobile 2.0 – i.e. when we mean ‘Open’ – we don’t mean the LACK of any structure – rather we mean the a structure which can evolve (allow users to experiment and reuse – like a playframe unlike a cage which restricts)
Gartner said over the weekend –
In all cases, there are claims of customer interests and the good of the industry, etc. etc. While there is some legitimate interest in standards, make no mistake that all vendors act in their own (and shareholders’) interests (as they should) and that if those interests line up with those of users, then great. But none are just altruistic and looking out for their customers only as they claim.
That was painfully apparent over the weekend.
There are many ways we could discuss this – starting with a taxonomy of ‘Open’ – the ten meanings of Open and understanding the various anomalies – ex Skype follows open source or open standards but wants the networks to ‘Open up’; Opera does not believe in Open source – but believes in W3C(web) standards.
And we could go on ..
I think it was Napoleon who said that all my troops are ready and prepared to fight the LAST battle – but not the NEXT battle ..
In other words, we cannot frame the discussion in terms of last years battle (Open source, Open standards etc) – and blindly apply it to the emerging paradigm(Cloud)
It seems to me that a discussion about the Cloud offers an opportunity to take a more pragmatic approach from first principles – taking into considerations what ‘Open’ means as in a Taxonomy of Open considering the different perspectives and then what Open means to the Cloud
There are many good ideas:
a) Define what Open means in terms of various possibilities(process level, Data portability etc)
b) Maybe create a more transparent process itself modelled on Wikipedia?
c) Discuss how to involve customers in the Open debate(again the wikipedia approach is transparent)
d) Discuss how the discussion could also include emerging economies? After all we have a G20 – and not a G10 (something which I am painfully aware of living in London this week!)
e) Consider the success of simpler initiatives like Microformats rather than considering complex, top down initiatives
f) Incorporate the ideas of ‘reasonably open’ i.e. playframe vs. rigid structure(obviously that’s an approach I have favoured)
g) Clearly we cannot ignore innovation over standardization. Ex – in the browser community – the most interesting developments have come from Google Chrome which has a completely different architecture which is a hybrid between a browser and an Operating system with process level isolation between browser tabs
But there could be many more
To conclude, The Open Cloud manifesto says that
The Manifesto was written as a rallying cry for the cloud computing community to come together around open Technologies .. .. I don’t like manifestos and rallying cries .. but I believe that its time to have a discussion around Cloud in context of a wider standardization process.
However, let us not succumb to the pitfalls of framing new discussions in context of the old paradigms
Comments welcome. Will add more to this later
Image credits: http://www.the1foru.co.uk/ekmps/shops/the1foru/images/action_tramps_2005_271%5Bekm%5D465x300%5Bekm%5D.jpg