The unstructured data that big data analytics extracts intelligence from is often too ungainly for the traditional block storage technology that most cloud providers offer.
With many enterprise customers struggling to accommodate their existing internal IT infrastructure and traditional data warehouses, cloud providers have an opportunity. With cloud object storage services, providers not only are offering cost savings, but also are addressing technical challenges associated with big data management for their customers.
Cloud object storage: Adding structure to unstructured data
Collecting and storing large data sets is daunting for businesses of any size, but accessing, organizing and analyzing unstructured data offers a whole host of new challenges for companies. Using block storage technology for big data requires that enterprises build a complex data model in advance to organize the unstructured data. Object storage can eliminate that costly step by allowing companies to integrate metadata into their unstructured data, said Dave Bartoletti, senior analyst at Cambridge, Mass.-based Forrester Research Inc. "Cloud object storage is a much easier way to approach vast volumes of data that doesn't fit well in a structure, but should be stored together," he said.
Object storage technology stores data together with metadata, which gives security, authentication and context to companies' information and applications -- a particularly appealing feature for industries faced with compliance regulations and those with large quantities of rich media, said Adrian Herrera, senior director of marketing for Austin, Texas-based object storage software provider Caringo.
More on cloud object storage
Storing big data with object storage
HP offers object storage in the cloud, challenges Amazon
Cloud object storage: making scalability a reality
"In traditional cloud storage, [users] have very little information about the file stored," Herrera said. "But as data sets are increasing in size, the storage solution needs to be able to efficiently store customizable metadata with the object itself."
Access is just as crucial to big data as the ability to search for it, and object storage can offer inherent security benefits over traditional cloud storage, Herrera said. Caringo's object storage offerings allow customers to store data as immovable objects, and audits access of each object, helping customers meet compliance regulations.
As cloud adoption for highly regulated industries gains momentum, certain vertical markets -- like health care -- are seeing the value in object storage for large patient files including X-ray and magnetic resonance imaging files, he said.
While the large cloud players -- like Amazon, HP and IBM -- are touting object storage offerings alongside their traditional cloud storage offerings, object storage is another option but not a replacement for traditional block storage. Businesses are choosing block storage over object storage for fast access to everyday applications, but object storage is better suited for very large data sets, and can scale easily. "[Object storage] is another type of storage that cloud providers have to understand, and it's best suited for particular applications and processing needs," Forrester's Bartoletti said.
Moving data is half the battle
While cloud storage providers could be able to sway more customers to the cloud with object storage offerings, the challenge lies in moving big data from enterprise data warehouses to the cloud, said Sam Barnett, directing analyst for data center and cloud at Campbell, Calif.-based Infonetics Research Inc. "Even with really good network connectivity, many users struggle with how to move big data to a cloud provider," he said.
Because most customers don't have large Ethernet pipes connecting them to their cloud provider, many cloud service providers have started overcoming the data transfer obstacle by taking snapshots of customer data or replicating over time from the customer site. While different cloud providers are taking different approaches, hard problems remain with transmitting big data into the cloud, Barnett said.
Big data analytics is a great opportunity for cloud object storage providers, however, because data mining requires specialty software. For locating and getting fast access to data, object storage makes a lot of sense for business customers, Barnett said.