Posted by Philippe Creytens at 14:41
Moving to the cloud is mostly initiated by moving ‘commodity IT’ services such as mail and calendar. Google, Microsoft and IBM/Lotus are trying to convince organizations not to have cold feet when thinking about moving to ‘the cloud’.
As cloud solution pioneers in the Benelux –we no longer have any servers– we see that after moving mail and calendars our satisfied customers, large or small, are eager to get more ‘cloud’.
In some cases the next steps are focused on implementing broad solutions such as CRM or document management solutions.
Others want to get rid of in-house developed applications and ask our advice on how to proceed. Clearly a difficult task that has some similarity with gazing into a crystal ball to come up with decent answers.
Amazon Web Services offers really great tools for developers: EC2, S3, RDS, and much more.
However you are stuck with managing infrastructure. Having to choose the right OS image, updating it and thinking about how your setup might need to scale, are only some of the side effects that come with going Amazon. Basically Amazon offers you a hypervirtualized environment on the public internet as a pay-as-go model where the developer needs to have a lot of knowledge on different IT topics.
MS Windows Azure is probably going to receive a lot of press attention in 2011. But it is, and will remain, primarily focused on .NET. Yes, I know Azure supports PHP, Ruby and Java, but I expect that PHP/Java/Ruby fan boys will prefer other solutions (see below).
Salesforce is ‘hot’ these days because of their recent acquisitions: Heroku, THE Ruby deployment platform, was recently acquired for $212 Mio. Soon Database.com will be able to provide cloud developers the power of an Oracle database for a minimal cost.
Even a blind man will see that cloud webdevelopers will be able to deploy very scalable applications with the trio Ruby on Rails, Heroku and Database.com without the infrastructure headaches.
And Google I hear you think?
Google prides themselves that Google App Engine is more and more being used.
But the success of GAE could be threatened by three factors: no support for SQL databases, Java and Python as the only supported development languages and a lack of an SLA with enterprise level support.
Nearly ten months after the announcement of SQL support and Google App Engine for Business, these announcements appear to remain vapourware.
As long as Google sticks to Bigtable as the only supported database, chances are pretty slim that enterprise applications, nearly all of them interacting with SQL databases, are going to end up on Google App Engine.
Our choice?
Not a 100% clear today, but Google App Engine in combination with Database.com is in our R&D plan for 2011. Unless if Google rolls out SQL support tomorrow.
Now, what would you choose and why?