When Google released its PaaS offering called Google App Engine, it attracted Web 2.0 developers in big numbers but it didn’t gain much traction like Amazon Web Services or Microsoft platform. In fact, in May 2010, Network World had an article quoting a Forrester Survey which put the percentage of developers using Google App Engine at a meager 8.2%.
Built for hosting Web applications, App Engine services more than 500,000 daily page views, but App Engine’s 8.2 percent usage rate, based on a Forrester Research survey of developers in late 2009, trails far behind Amazon.com’s Elastic Compute Cloud (EC2), which has nearly a 41 percent share. Microsoft’s newer Windows Azure cloud service edges out App Engine, taking a 10.2 percent share. Forrester surveyed 1,200 developers, but only about 50 of them were actually deploying to the cloud.
The IT category however remains the top performing job category with a 44% increase in provider earnings in Q2 year over year. Fueling this growth are companies embracing cloud computing platforms and mobile devices that need to tap into highly skilled and qualified work teams. After making an entry into the top 50 skills in Q1 2010, Google App Engine, the cloud-based application development platform, showed the single largest increase in demand with over a 10x growth quarter over quarter. With this increase, Google App Engine (#37) moved past Amazon Web Services (#40) as the cloud platform highest in demand in Q2.
With Microsoft pushing hard on Azure, we are in for some good competition. If this trend holds universally, we can see that a trend towards platform services exists moving developers away from the IaaS offerings like EC2 consistent with what I say about the future of platform services. I think the next two years will be interesting to see how the market develops for platform services, in general, and Google App Engine, in particular. What do you think? Will enterprises ever take Google App Engine seriously? With today’s vFabric announcement, how serious VMware will be in their support for Google App Engine?
While Azure should be my first point of call, GAE has become the first place I do things these days – because of the (lack of) cost along with scalability.
For example, the backend services for London Bike App (http://www.londonbikeapp.com) could be hosted on AWS, GAE or Azure – or even Dreamhost, where the website is. I chose GAE because the entry point was free – AWS would cost me a lot more than the app is making (sadly), and Azure’s pricing is all over the place. Dreamhost was an option (Rails or PHP), but I wanted something that could scale with little or no interaction from me.
Being LBA is a .NET app (MonoTouch), Azure would have been the logical place if it wasn’t for the stupid pricing.
From my perspective, if I want full control, I can get it and pay for it (AWS), and I’m about to login to a server there to do some “day job” stuff. It’s way too expensive for personal stuff tho ($90/month or so), or things which _might_ make some money down the line. It’s also a load of work to get a webserver, database (scalable), memcache, a queue etc all running.
I considered Azure, as I prefer C# to either Python or Java (I’m using Python on GAE at the moment, looking to move to Java or Groovy soon), but the pricing is crazy. I have to stick to MS’s API’s (like GAE), but pay for the app to be running all the time (like AWS), and the pricing is around the same as AWS. Definitely a MSFT Fail. The services/API’s are good, but they pooch-screwed the pricing (or at least the communication of it). I suspect this is so they dont “damage” the 3rd party hosting market. Typical MSFT.
AppEngine is priced right for me, as I’m usually in the free bracket., but if I need to scale up past 1m hits a day (which, frankly, would be fantastic!) I can, easily. The API’s are good, and they clearly explain what you can and can’t do. BigTable is a bit interesting if you are only used to SQL, but it’s not that hard if you stop thinking that way. The single-button deployment is brilliant.
So…. if Google added C#/Mono as a possible GAE language, I’d be using that in a shot. As it is, what they have is very compelling. If a company is already committed to Google Apps, it makes sense – same as Sharepoint makes sense if you have a MS environment.