EMC eyes cloud, big data convergence

Storage giant recognizes big data analytics possibly bigger than cloud as disruptive force and views integration of both as "enormous opportunity" for customers and EMC, notes company exec.

SINGAPORE--Cloud computing is a trend that is here to stay but big data analytics can possibly effect as much, if not more, transformation on how businesses utilize IT to improve performances.

This was the proposition put forward by David Webster, president of EMC Australia, New Zealand and Southeast Asia.

In an interview Tuesday with ZDNet Asia, Webster said cloud computing is creating increased demand for real-time business insights as data is being accessed and consumed through various endpoint devices anytime, anywhere.

This deluge of data reached 1.2 zettabytes in 2010 and is expected to increase to 35 zettabytes within the decade, said the executive, citing IDC findings. Additionally, 90 percent of data in today's digital universe is unstructured, meaning, information can come via various sources in the form of photos, voice recordings and e-mails, for example, he added.

Webster also pointed to the changing software landscape as another factor why opportunities for on-demand, big data analysis are growing.

He noted that traditional "large, monolithic software" with big licensing fees is giving way to "small, component-type apps" that adopt pay-as-you-use pricing models, making it easier for companies to deliver specific data insights to employees through their mobile devices in real-time.

With regard to the future of big data analytics, he said EMC's acquisition of Greenplum, which was ratified end-2010, represented a "change in the way people should view data".

Elaborating, he noted that Greenplum is an open system, analytics engine middleware that "sits on top of the hardware and just below the software". It leaves the reporting function to other products though, typically business intelligence software from SAS. This is different to other analytics offerings by competitors such as Oracle, IBM and Teradata, he noted.

"Big data analytics by these companies are boutique, black-box projects that require companies to buy the whole stack of supporting infrastructure to run analyses. Their method of analytics is still stuck in the traditional data warehousing mode, whereas Greenplum's technology is built for the virtual world," Webster stated.

Rethinking storage architecture
That said, the EMC executive suggested companies relook the way they architect their infrastructure environment in order to maximize the opportunities that exist where cloud computing converges with big data analytics.

He explained that this is necessary because of a fundamental difference in storage for enterprise business applications, compared with storage for big data applications.

Mission-critical applications such as ATM (automated teller machine) functions used by financial institutions, for example, will need supporting hardware infrastructure that ensures "Tier-1 availability and reliability", Webster said. Comparatively, for a video production house that processes thousands of images each day, the storage system will need to be "scaled out" and have a "massive file system architecture" that can hold petabyte-sized single files.

As such, Webster does not believe it is possible for a one-size-fits-all storage architecture that can support the needs of both sets of applications.