Last week, Microsoft released a new version of Azure with a strong focus on IaaS. Azure was previously more targeted towards applications running on the .NET stack with a wide PaaS approach but now it is a full-fledged IaaS cloud with features like Linux nodes, persistent VMs and the ability to control the network location of the VMs.
The (almost) joint announcement of two heavyweights of the Internet / Cloud to focus on IaaS instead of pushing only a PaaS model can quite naturally lead to the question: is the PaaS Cloud model about to be confined to the history books?
First of all, it is important to note that neither of these vendors have announced that they would discontinue their PaaS model. The PaaS model is very attractive in a context where an application is built using a fairly homogeneous technical stack with little integration with existing systems; it is also very suitable for organisations who don't want to worry about low-level intrastructure administration: they can focus on development and let the Cloud provider manage the rest. This model has been very successful with start-ups who have embraced Cloud very quickly and for whom the PaaS model is definitely of interest, since it allows focusing on end-user features development.
This shift of focus towards IaaS could be seen as a clear sign that large corporations are now seriously considering cloud; for these corporations who have existing systems, procedures, legacy systems and will require hybrid on-premise/off-premise solutions, the IaaS model is the only suitable one because it is the most flexible one.
With Microsoft and Google now fully engaged in the IaaS race, it will be very interesting to see the impact on the price of resources and potentially on the cost models that haven't been really revisited since the launch of Cloud. Google has already announced that one of the key features of Google Compute will be... to offer a 50% more competitive price point!