On-premise. Co-location. Public cloud. These terms are being bandied about as if they are ‘either/or’, whereas the reality is that many organisations will end up with a mix of at least two of them.
This is why it is becoming important to plan for a hybrid datacentre – one where your organisation has different levels of ownership over different parts of its total data centre environment. It is time to move away from the concept of the singularity of the ‘datacentre’ and look more at how this hybrid platform works together.
This makes decision making far more strategic. If the organisation has reached a point where its existing data centre is up for review, it may make more sense to move to a co-location facility, which then enables greater flexibility in growth or shrinkage as the organisation’s technology needs change – and indeed, as the technology itself continues to change.
However, although co-location takes away the issues of ownership of the facility and all that it brings with it (the need to manage power distribution, backup, auxiliary power, cooling and so on), it still leaves the ownership and management of the IT equipment within the co-location facility to the end-user organisation.
This can, in itself, be expensive. Few resellers offer an OpEx model to acquire hardware (although capital agreements with the likes of BNP Paribas Leasing Solutions can turn a CapEx project into more of an OpEx-style one), so with co-location organisations still have to find the upfront money to purchase hardware. Then there are the costs of licences and maintenance for all parts of the software stack – as well as the human costs for applying updates and patches to the software and for fixing any equipment that fails. However, where an organisation believes that it can provide a better platform for its workloads than a third party platform provider can, colocation is a great way to gain better overall security, scalability, performance and flexibility.
If ownership of the stack offers no particular benefit, this is where public cloud comes in. Infrastructure as a service (IaaS) gets rid of the CapEx issues around hardware; platform as a service moves further to get rid of the costs of licencing and maintenance of operating system and certain other parts of the stack; software as a service (SaaS) gets rid of all these costs, rolling everything up into one subscription cost.
None of these choices is a silver bullet, however. For whatever reasons – whether these be for solid, logically thought through or more visceral ones – many organisations will end up with some workloads where they want more ownership of the platform, alongside other workloads where they just want as easy a way to deploy as possible.
Planning for a hybrid platform brings in the need for looking well beyond the facilities themselves – how is high availability and business continuity going to be achieved? How is security going to be implemented and managed? How is data sovereignty impacted? How is the end-user experience to be optimised, monitored and maintained?
The world of the datacentre is going through a period of rapid change not seen since the hegemony of the mainframe was broken by the emergence of distributed computing. Few can see beyond an event horizon of just the next couple of years – virtualisation and cloud are still showing how a datacentre needs to shrink; other changes in storage, networking and computing power may yet have further impacts in how future datacentres will have to change.
At this year’s Data Centre World, attendees will be able to engage with vendors and customers involved in this journey.