Thinking big comes to crunch time

The computing power of the cloud is perfectly suited for the Herculean task of crunching big data, writes Rod Newing

Big data, any technology officer will tell you, is mainly unstructured information that is too large to be managed by a traditional relational database. It is defined by large volumes, a range of data types and fast growth – all features that make it perfectly suited to the cloud.

Here, more than anywhere, the cloud competes on cost. “Procuring and managing on-premise storage in order to glean information from big data is dauntingly expensive,” says Leena Joshi, senior director of solutions marketing at Splunk, which indexes machine data.

What the cloud can do with this mass of unstructured information is important because big data is getting bigger.  An October 2012 report from Talend, which builds data management software, found the practice of collecting diverse stacks of data for further analysis is happening in 48 per cent of financial transactions, 48 per cent of social media and internet text, 36 per cent of web logs, and 28 per cent of call detail records.

Organisations that have been accumulating this complex data for years can now access technology to manage and analyse it, and are climbing aboard. According to research by the Saïd Business School, on behalf of the IBM Institute for Business Value, 28 per cent of respondents are piloting or implementing big data initiatives and a further 47 per cent are planning them.

Bill McGloin, storage and data optimisation head at Computacenter, an IT infrastructure services provider, explains that organisations start by uploading their static data, more than half of which typically hasn’t been accessed for a year or more.

“It is sitting on their most expensive storage,” he says. “Moving it off the premises lowers the cost of managing and backing it up regularly. It frees up staff to be more productive, allowing them to focus on live data, as well as reducing environmental costs.”

Often the data has been sitting idle because no one has figured out how to mine it. Once on the cloud, it can be poured over by a swarm of analysis and business intelligence tools offered by the cloud provider or a third party, far exceeding the resources available to the firm itself.

On-premise storage in order to glean information from big data is dauntingly expensive

The cloud also copes well with seasonal shifts in demand for data crunching. These are popular in retail, e-commerce and the charity sector. “Cloud analysis is burstable and can be turned on and off again,” says Jonathan Wisler, general manager for Europe, the Middle East and Africa (EMEA) at SoftLayer. “To get similar speed of analytics in-house, you would need to be a very large company with a lot of resources.”

Many organisations start big data initiatives with a private cloud and then expand it to link in with a public cloud in a hybrid environment. This increases capacity, reduces costs and boosts performance. Mindful of national data privacy laws, clients now have a say over where providers store their data.

Which data ends up where in the hybrid cloud depends on volume and sensitivity, says Andy Jolliffe, business development director, data centre, EMEA at Juniper Networks. Real-time critical analysis can be kept close to the business and lower-cost cloud services used for batch-based analysis.

The quicker big data gets off clumsy local servers and into the cloud, the better, he concludes. “The underlying storage infrastructure, already struggling to cope, and the most bloated IT spend are like concrete galoshes compared with the cloud when ‘going big’.”


Racing ahead in cloud

Cancer Research UK, a leading charity, organises Race for Life in which thousands of women run to raise money for the charity. Activity on the race website is cyclical as publicity builds ahead of each summer’s events. This spike in activity placed pressure on its third-party hosting resource.

“It became clear that the old expensive system was not fit for purpose and needed a major overhaul,” says Mick Briggs, head of infrastructure at the charity. “Savvis was the only provider flexible enough to offer a non-standard service.”

The charity implemented a hybrid solution, with nine-to-five development and disaster- recovery services on a shared infrastructure, managed by the charity, and web hosting on dedicated cloud service, which the host monitored and supported 24/7.

The charity has since broadened the scope of the arrangement to include other mission-critical web applications, as well as increasing its co-location capacity. Mr Briggs believes that as cloud adoption matures, such solutions will become an industry standard.

“We created a hybrid cloud, bundling the Race for Life project with other managed services, for 50 per cent less than the majority of other service providers were offering for the Race for Life website hosting alone,” he says. “The flexibility it provides allows us to make decisions much faster, while ensuring that we reduce the revenue spent on technology.”