The Cloud Infrastructure-as-a-Service (IaaS) providers offer businesses all over the world server and storage resources, that follow the Cloud Computing concept as defined by the NIST (National Institute of Standards and Technology). Cloud is nowadays a very generic and over-hyped term, but it is important to summarize what a pure Cloud offer must include: On-demand self-service (the ability for you to decide when you need the resources available), Broad network access (available everywhere, through the Internet), Resource pooling (the providers computing resources are shared amongst multiple customers), Rapid elasticity (scale up or down, quickly, when needed), Measured service (pay only for what you use in terms or resources units, like GB, CPU, …).
From these characteristics, I would emphasize three: On-demand self-service, Rapid elasticity and Measured service. When you choose your provider, make sure you have the ability to easily provision the services that you need (which includes creating and deleting servers, adding or reducing RAM, DISK and CPU resources, IP addresses, etc), when you need, and that you are also not stuck to a baseline configuration that you must pay, regardless of whether you use it or not. Elasticity is one key characteristic of the cloud and together with the “pay-per-use” concept, it means that you should be able to increase your resources when you need it (adding more servers or adding more resources to a server), but as easily, also reduce the resources when you don’t need it, which will allow you to adapt the costs of your cloud computing infrastructure to the ups and downs of your business or projects. For a real “pay-per-use” model, similar to what we have with utilities such as electricity, water and gas providers, your supplier must meter the resources you use by the hour and not per month. You don’t need to pay for a 16 GB RAM server for a full month if you only need it for one week, right?
Globally, Amazon Web Services (AWS) is the clear market share leader in cloud infrastructure services, with two main services: EC2 and S3, respectively, servers and storage. Many other players are now in the market, either with more expensive offers targeting the higher-end business market (mid-to-big companies, that need certain levels of security, compliance, support, reach, …) or more cost competitive offers for SMEs.
Hosting your servers and storage in the Cloud, means you are putting your data in a Datacenter. After all, the Cloud is physically in Datacenters, of course. Which takes us to an important aspect that some cloud providers have addressed, which is geographical diversity. Being able to offer customers the ability to choose in which region of the world they want to locate their computing resources and their data, is an important feature. Not only because it gives additional redundancy (not everything is concentrated in a single Datacenter), but it also addresses important technical aspects, such as network proximity (and performance) as well as legal and cultural issues.
This is where AWS doesn’t score so well, not globally (it has 8 regions across the world), but in Europe itself, where there is a single datacenter location in Ireland. Usually, US-based providers adopt the same strategy, with Europe being a single market and therefore needing a single datacenter or region to provide services from.
My point is that Europe is more diverse, culturally, geographically, legally and in terms of currency, than this one-size-fits-all approach seems to put it at. Cloud infrastructure services that are meaningful for European businesses should take this into account. A single provider does not need to have datacenters in every European country, as that wouldn’t scale well and would not have the same economies of scale of bigger global providers. But Europe is better served by 3 to 5 nodes that spread across its South, Central Eastern regions.
Why? Network-wise, this would guarantee faster access speeds to compute and storage resources to customers, over the Internet, because it would be closer to customers (from Russia to Ireland, the distance is more than 60 ms and can be a lot more, depending on the network you use). Also, because it would give geographical diversity options to customers, that might want to spread their infrastructure across different regions, for additional redundancy. Finally, although the EU is standardizing legislation regarding data locality, we are still not there. And some European countries are not in the EU. Therefore, having more datacenters in more countries, reduces some legal bottlenecks concerning where your data is located. Finally, the widespread cloud infrastructure needs to be well served by support teams that ideally speak the language of their customers, to really help them use the Cloud fully.
The Cloud is a smarter way of using Information Technology resources. The adoption curve by businesses is still at the beginning, but only by understanding the markets and business needs providers will succeed, helping customers benefit more from using the Cloud.