It’s the stuff of nightmares: After months of preparation and marketing the launch day finally arrives, the “Go” button is pressed and your creation is released into the wild. Shortly after that moment your website starts failing; too many people are trying to log on and the site simply cannot cope with the extra traffic.
Whether it’s a consumer electronics company announcing its latest smartphone or tablet, a band streaming its new album or an online retailer announcing a massive sale, spikes in traffic and usage can cause all sorts of problems. Sales can be lost and reputations can suffer. UK broadcaster ITV, for example, has suffered a couple of times in recent years with streaming content online - there was so much demand for the first episode of the second series of Downton Abbey that ITV’s online player simply couldn’t cope.
When infrastructure is pushed over its capacity it can be frustrating for all involved, but the simple answer of buying more and better infrastructure isn’t really suitable.
Traffic spikes are generally one-off or very rare events such as those listed above and do not justify the cost of additional hardware that would not be used when the traffic spike recedes, particularly as tightening budgets and increased competition are forcing many businesses to sweat the assets and get more from their existing infrastructure.
This doesn’t just apply to websites though; any application or service is at risk from collapsing if it hits peak capacity.
That’s where cloud bursting comes in. During peak periods an application that is running in a corporate data centre or in a private cloud can “burst” into a public cloud, providing the extra capacity needed to keep services running smoothly. It also means the company will only pay for this extra capacity when it is used, keeping costs down.
This works by abstracting the application delivery requirements from the underlying infrastructure, enabling the applications to basically span physical and virtual infrastructure in the data centre and in the cloud, as demand dictates. Increasing an application’s available resources by dynamically redirecting workloads as needed will result in a more stable and reliable service for end users, which will benefit all parties.
However cloud bursting relies on data centre agility, which can sometimes be impacted by the network. Using something like F5’s Cloud Bursting technology can eliminate those network bottlenecks, by using metrics of real-time service behaviour to deliver demand-based workflow routing. Doing this over public and private data centres eliminates the restrictions of physical device, connectivity and capacity experienced within data centre silos.
F5’s Cloud Bursting solution also ensures that all relevant security policies are enforced when an application is burst to the cloud, meaning regulatory requirements will still be met.
Cloud bursting is a great way of ensuring applications can keep running through huge spikes in demand without forcing a business to pay for infrastructure that does nothing for long periods at a time. The business will only pay for the additional capacity it uses and end users will be able to access what they want whenever they want to, even if huge numbers of others are doing exactly the same thing at the same time.