Cloudily

So where are we today? Well, just about everything can be done in the cloud nowadays. At least, there are SaaS services available that offer standard functionality to support most business processes. And then there are all the integration possibilities both between services (e.g. data-mining) and with business specific solutions (e.g. via Force.com or some form of REST API or similar). In short, it is possible for companies to choose their comfort zone when adopting cloud technologies.

We mustn’t forget that virtualisation is the backbone of all cloud infrastructures. Functionality like thin-provisioning provides economies of scale and a more efficient use of resources. Virtualisation started in the enterprise as private clouds and is still as relevant today even with the fantastic flora of public cloud offerings. So picking your comfort zone raises some important questions for companies, the most important being, what are the business drivers for adopting cloud technologies?

There is the obvious cost driver that can be realised with virtualisation coupled with hosting (aka IAAS). Virtualisation provides efficiency and hosting eliminates the need for resources for management and maintenance activities. This also brings us to the competence driver. IT organisations are implicitly expected to remain small and basically invisible to the rest of the organisation. Certainly in the Cloud Age if you need more IT people instead of outsourcing then the view is that you’re doing it wrong.

Virtualisation products also deliver a host of other functionality that can be important business drivers. High-availability, scalability and performance monitoring features are common to probably all IAAS platforms. So outsourcing infrastructure can provide cost savings plus lots of new features in a like-for-like migration, i.e. moving from on-premise to an equivalently performing cloud infrastructure with respect to the potentially different workloads of each IT system (e.g. IOPS vs. data-mining).

The next driver to look at is flexibility, the ability to be able to bring more resources online in response to a temporarily increased workload. The question is, does the organisation need flexibility, really? Yes, loads vary, but this must be weighed against the cost of hosting in an on-demand infrastructure instead of, say, a (virtual) private cloud (VPC). A VPC can of course be extended to meet increased workloads but this is a more long-term activity such as meeting the needs of a growing organisation. But, if just a single application requires extra capacity then this does not motivate putting the entire IT-environment in an on-demand IAAS solution such as AWS or Microsoft Azure. Certainly this is true of larger organisations which have a bigger and more stable infrastructure footprint.

Flexibility then is achieved using a hybrid solution. The backbone of the IT infrastructure should be virtualised as much as possible using a private cloud (on-premise or hosted). The private cloud will have a predictable workload and will expand or change at a strategic pace. The private cloud is then complemented with on-demand resources from an on-demand infrastructure supplier such as AWS for handling peak workloads. Finding a balance between these two costs can be achieved using the resource monitoring tools in both platforms. Vertical vs. horizontal scalability is also a deciding factor.

Security concerns are usually not a business driver. By this I mean confidentially and integrity of data when some or all of the IT systems are residing in one or more IAAS platforms. These platforms usually provide networking capabilities such as VPN to allow the corporate network to extend into one or more cloud infrastructures and so provide seamless integration between the various IT systems. Central authentication systems can also use the same VPN tunnels to control authentication and authorisation in all the relevant systems.

To summarise, this walkthrough of the IAAS landscape shows that you can pick your comfort zone with a proper analysis of the infrastructure requirements. The benefits are huge so some of the cost-savings should be used to over-dimension the infrastructure solution and so save time performing difficult workload analysis. Hopefully the company will grow into its new suit and this will also provide important input to future infrastructure investments.