Since quite some time there as been wide agreement on some criteria for cloud based applications - namely that they had to support multitenancy, all the way down to the database level, they had to be deployed to a public infrastructure, they would only have one production code line and so on.
Already slaughtered - no customizing
For a long time the cloud application vendors have been maintaining, that they cannot support any form or shape of customizing - as they otherwise would no longer have a cloud application.
In my view this was a little bit of a self serving argument as it allowed the vendors to move fast and with little complexity from release to release. But to be fair more and more vendors start to support some more or less elaborate ways of customizing their cloud applications. So already in 2013 we did not hear the moniker - 'we are cloud we don't allow customization' (much) anymore.
The first to go in 2014 - database multitenancy
We have already written in mid 2013 that database multitenancy - as we knew it - as being a database containing rows of data owned by different clients - is largely an architecture of the past. It was largely required due to hardware constraints for the very first cloud architectures - but should not be deployed for a modern cloud storage in 2014 and onwards. Too many advantages speak for the end of database multitenancy - most prominently access security, predictable performance, and operational advantages.
You could argue that database multitenancy has already disappeared in new, state of the art cloud architecture - but probably 2014 will see the more wide stream adoption.
Next - public shared infrastructure
A table stake of cloud architectures used to be, that cloud applications had to be deployed on public and shared infrastructures. And while that is desirable for most applications, there are more companies out there, that do not want to have their applications being hosted on a public infrastructure. Some may say it may be triggered by the whole NSA / Prism sensibility, in my view the saturation of early cloud adoptions and the need of cloud application vendors to grow revenue wise, play an equal important role.
And the vendors are reacting and gearing towards that - the AWS government cloud is an example. Salesforce supporting the HP Superpod is a similar one. And then with most cloud application vendors embracing OpenStack, a deployment of their cloud applications on a on premise OpenStack infrastructure is technically possible and in my view - likely in 2014.
Of course cloud purists will now roll their eyes - and start to argue... we will see what 2014 brings.
And then - one release for all
As cloud applications get more and more adopted - it becomes more of a challenge to upgrade these applications centrally and synchronously for all customers. Cloud vendors have for the longest time argued (and in my view even a little hidden) behind the fact, that if a single customer would have their own version of code - they would be no longer a cloud application. That's of course not accurate... and with more flexible deployments in 2014 - we will see cloud vendors to begin supporting different code levels by customer.
Implications for customers
It will be key for customers to make sure their vendor supports the more complex code deployment landscape that results from slaughtering some of the sacred cows. Do not take 'that's not cloud' as an answer anymore in 2014.
Implication for vendors
If you are not revisiting you code delivery and configurability, 2014 maybe a rude awakening for you. Better to disrupt and be early on these trends than be disrupted by the competition. Look at OpenStack as the easy way out, that a number of the larger cloud vendors have already adopted or at least are heavily looking into.
MyPOV
Nobody knows what the future holds - otherwise those who knew would play the lottery and win every week.... but it's time for criteria that defined a cloud application for the longest time - are being revised by the market. Even sacred cows do not live forever.
In my view - a lot of that will happen in 2014.
P.S. Don't miss the 2014 cloud trends that fellow Constellation Research colleague Ray Wang and me have put together here.