Like many other big analyst firms, Forrester Research looked in their very own glass ball and made predictions for the cloud computing market in 2013 which will actually reach. From this ten predictions I took three and highlights them from my perspective.
Cloud and mobile will become one
What’s the value of a mobile app that doesn’t call out through the Internet to back-end services? Not much. And where will these back-end services live? Probably not in your datacenter — unless you plan to poke a big hole in your firewall to accommodate an unpredictable flood of traffic. More often than not, we are finding mobile applications connected to cloud-based back-end services (increasingly to commercial mobile-back-ends-as-a-service) that can elastically respond to mobile client engagements and shield your data center from this traffic. Nearly every software-as-a-service (SaaS) application has a mobile client now, which is proof of the model as well. As Forrester analyst Glenn O’Donnell puts it, cloud plus mobile is a classic “more than the sum of its parts” combination.
On this I have to disagree because this prediction has already occurred. Having a look on mobile apps on iOS, Android or Windows Phone, we quickly see that, instead stored on the local device, most of the data are already are strored on servers in the internet and downloaded on demand. Well-known examples are Foursquare, Dropbox, Google Drive, SkyDrive, Facebook, Twitter, Google Plus, Analytics Tools, Instagram, Evernote (hybrid), etc.
More on the issue at “The Mobile Cloud is the real mega-trend“, actually written in German in March 2011.
We’ll stop stressing about cloud SLAs
And recognize that apps have to protect themselves. The best practice for cloud application design and configuration is to build resiliency into the application rather than expect it from the cloud platform. This way you can achieve any service-level agreement regardless of the base SLA provided by the cloud platform. Getting the performance you need is an application-specific goal anyway. What’s the value of having your sourcing and vendor management team negotiate a high and tight SLA from the cloud vendor when only 10% of the applications deployed there need that level of protection?
The name for this is “design for failure”, because “everything fails everytime”, of course cloud infrastructures as well. Cloud computing per se delivers no high-availability for the customer. But it gives the ways and means to reach high-availability. Especially in the field of infrastructure-as-a-service you have to look on the horizontal scalability characteristics of a cloud infrastructure to design the applications and systems as available as possible. That means not to use just one server. Instead you should use several servers and distribute them over various availability zones (datacenter) and regions (global distribution). As a cloud-architect you have to avail oneself to the scalability of the cloud and therefore design the application for distributed infrastructures. It also means that a server has to start automatically if the load growth and shutted down as well when it is no longer needed.
However, regarding SLAs we should keep the providers in mind. The service level must be correct and complied with. In my view, the “pain” for the providers is not large enough when something happens. “Only” reimburse the costs, which were charged during the period of the outage, is not enough.
We’ll stop equating cloud with AWS
While Amazon Web Services has opened up a substantial lead in the cloud platforms market — arguably as large as 70% market share — in 2013 we’ll see that market position give way to a cadre of strengthening competitors and new entrants. Microsoft and Google have made significant improvements to their platforms, and by the end of 2013 we fully expect to see at least three substantial OpenStack-based clouds building strong positions. Look for a Forrester Wave™ of public cloud platforms in mid-2013.
The strongest competitor to Amazon Web Services are, as Forrester says correctly, Microsoft Windows Azure and the Google Cloud Platform. All others have to catch massively, because most of them just(!) offering infrastructure. Amazon has done a lot to build value-added services around the AWS infrastructure, which customers can profitable use to utilize the “stupid” infrastructure. It’s the same with Microsoft and Google. Other public cloud providers on the other hand have only computing power and storage, that’s it. Some talk a lot about that they have a new IaaS (quasi a next generation IaaS), but behind is just still stupid infrastructure a customer can not do much so far.
A good approach for a next generation IaaS will be a convenient IaaS like a Infrastructure-as-a-Platform. Because IaaS like AWS are really difficult to understand and therefore to use.