While not every organisation wants to risk outsourcing its vital IT services to a public cloud provider, more and more are taking advantage of the flexibility and scalability offered by a private, in-house cloud. Indeed, according to Technology Business Research (TBR), the private cloud market will generate $69 billion (£42 billion) by 2018. The benefits of a private cloud infrastructure are clear: IT services can be easily created and provided to users as and when they are needed, while at the same time organisations can avoid the very real concerns of compliance, security and control that can accompany handing IT services over to a third party.
For any organisation, implementing a private cloud depends on having the infrastructure to support it: in general, this will result in a single, centralised data centre (or group of data centres) holding all the servers needed to provide cloud services. While this provides the necessary computing horsepower more efficiently, it is also a glowing weak spot: a failure in this infrastructure could remove IT services from the entire organisation for however long it takes to resolve.
As a result, when adopting the private cloud, organisations need to be completely confident in the underlying infrastructure. Whether building the private cloud on top of an existing environment or taking the opportunity to perform a refresh, this infrastructure will almost certainly be virtualised. Essentially the ability to quickly and consistently create and allocate new machines as and when needed, without worrying about physical limits, is one of the keys to providing a cloud infrastructure.
However, businesses have so far shown a lack of capability when it comes to ensuring virtual infrastructure can survive in the event of a disaster. Too often, they assume that the same approaches that work for a physical environment will work just as well with virtualisation. Yet the technology operates on a very different set of rules meaning new approaches are needed. Without these, attempting to keep services running will be like performing surgery with a hammer.
Server downtime costs £250,000 per hour
This incorrect approach is already showing worrying results. A recent survey of 500 CIOs across the UK, US, Germany and France found that over half of businesses encountered problems simply with backing up and recovering their virtual infrastructure. Considering that server downtime was estimated to cost $409,531 (£250,000) per hour, and that the average time to recovery was five hours, we can see that downtime can have a huge cost: even for organisations that may not necessarily be using a private cloud. The research also found that, even if it takes hours, recovery of services is not always guaranteed. In one in six cases, businesses stated that they experienced problems when attempting to recover their virtual servers. Again, when an organisation’s services are reliant on this infrastructure, it needs to be sure that any service that fails can be brought back online.
The fact is that this shouldn’t be a problem with virtualisation. Indeed, if used correctly it should make private cloud infrastructure even more resilient. For example, virtualisation makes it extremely simple to use replication, where a duplicate IT infrastructure is created that can be switched to in the event of a disaster. Indeed, creating this infrastructure should simply be a matter of copy and pasting virtual machines across. If it is consistently updated, then a failure should not cripple the private cloud. Similarly, the ease of creating new infrastructures means that an organisation can, with the correct tools, create a space to automatically test backed-up machines and be sure that they will recover as expected.
The Big Questions:
All of these considerations boil down to one thing: confidence that the private cloud infrastructure is resilient. Whether a business’s infrastructure is being implemented in-house or by a third party supplier, it should ensure that it can answer all of the following questions:
- Can the most critical services be back up, and running, near-instantly in the event of failure?
- How long will other business services take to restore: minutes, hours or worse?
- How often is backup performed: every hour, every day, every week?
- Can the recoverability of backed-up machines and services be guaranteed?
- When backup is performed, how quickly can it be done: will it restrict access to services in the meantime?
Without satisfactory answers, businesses will find it increasingly difficult to have confidence in the infrastructure that supports their private cloud.
About Alex Ball and Veeam
Alex Ball is Manager UK & Ireland, for Veeam a cloud security business founded in 2006, Veeam currently has over 20,000 ProPartners and more than 80,000 customers worldwide. Veeam’s global headquarters are located in Baar, Switzerland, and the company has offices throughout the world.
Previously to Veeam Alex Ball had various international roles in technology companies, including Blue Coat, AppSense and Microsoft. As UK & Ireland Regional Manager.