As the shocking events in Canterbury have brought home this past month, disaster can strike anywhere, anytime. Thankfully, there was no loss of life in the Canterbury earthquake, despite widespread destruction, but it serves to highlight the need, wherever we are, to plan ahead and be prepared.
In any form of disaster, many organisations would face ruin if their computers were knocked out of action for longer than a day or two. As more and more businesses replace traditional paper-based supply chains with electronic links to customers and suppliers, disruption anywhere along the chain can have major consequences.
At the same time, with consumers turning to the internet as a convenient way of buying goods and services, the new businesses that have sprung up to supply them — often boutique operations with slender resources —are entirely dependent on uninterrupted computer and network uptime. This is nothing new for established large organisations. Ensuring their mission-critical data centres are fully redundant and backed up by an exhaustive disaster recovery (DR) plan is a fully budgeted cost of doing business.
For many businesses, however, there may have been a tendency to view DR as a luxury they couldn’t afford. With earthquakes, floods and run-of-the-mill power failures a matter of ‘not if but when’, the reassuring message IT providers can offer smaller organisations, is that server virtualisation is an affordable way of building resilience — and DR disciplines — into their systems. Virtualisation itself isn’t a DR strategy, but it can simplify implementation of one.
Before virtualisation became a common way of provisioning server resources, the starting point of a DR strategy was to duplicate your infrastructure in a second data centre — not an affordable option for many organisations, although they might have got part-way there by mirroring their most important systems.
What virtualisation does is turn servers into a resource that is not tied to a particular piece of hardware in a particular location — processing can be readily switched from one data centre to another. The key thing virtualisation does is it makes your application workload portable. Where previously you may have needed to physically deliver backup tapes to a secondary data centre, laboriously recreating an application server, virtualisation encapsulates an application in a portable file that can instantly be started on any virtual infrastructure platform. Depending on the nature of the disaster, you could potentially point and click to move that application workload to another physical location.
By virtualising applications, organisations give themselves the ability to take that workload and move it within a short space of time — that’s minutes, in some cases, rather than the days it could take to rebuild a physical server.
The cloud — the ultimate resource pool, with a location invisible to the user — could be a key ingredient in this. It could be a private cloud aggregating the organisation’s own computing resources, or a public cloud subscribed to on a pay-per-use basis. Either way, virtualisation is seen by many IT experts as the foundation for cloud computing.
When it comes to getting this across to clients who’ve previously thought comprehensive DR was beyond their means, a number of points can be brought out. Aside from the highly attractive cost advantages of virtualisation in DR provisioning, another thing to emphasise is that virtualisation is now a mature infrastructure architecture.
Independent software vendors (ISVs) have played a big part in helping virtualisation become respectable, both by using it inside their own development shops, and by certifying their products to work in virtualised environments.
Where once they would have dictated a particular hardware setup for their systems — and perhaps understandably so, when the range of variations they would have needed to support was enormous — there are obvious benefits for them in merely having to support a handful of virtualisation vendors. Some early adopters of virtualisation solutions had concerns about performance, when numerous virtual servers on a single piece of hardware were competing for resources. Today, virtualisation is a mature technology and capacity planning is a well established discipline in the virtualisation market. Perhaps foremost in the disaster recovery benefits of virtualisation, is cost. DR used to be something only large enterprises could consider, but virtualisation has brought the cost down. Using virtual infrastructure, businesses can improve many aspects of business continuity and disaster recovery, realising faster, more fl exible and reliable disaster recovery at a lower cost by removing the dependencies of operating systems on specific server hardware.
With even smaller organisations increasingly dependent on 24/7 system uptime, and the ever-present threat of calamity, virtualisation offers them the DR assurance they need at a price they cannot afford to ignore.