Sachin Jain, CIO and CISO of Gurgaon-based Evalueserve had a challenge at hand. The company had cross-functional teams that worked from different locations. These teams ended up working on the same document and creating multiple copies of it and storing them separately. This resulted in duplication of data and consumed around 30 to 40 percent of the overall storage space.

“We realised that our storage was not being utilised effciently and we wanted to address this issue immediately,” says Jain who decided to go for a data deduplication solution to reduce cost, control data growth, reduce storage requirements and improve performance.

Similar to Jain, many enterprises acknowledge the fact that as the amount of primary data within organisations continues to grow, the amount of duplicate occupies signifcant amounts of storage infrastructure. Redundant data quickly consumes storage resources at an alarming rate, driving up business costs through increased storage purchases and management. Within no time, storage consumption grows exponential and becomes extremely diffcult to control.

Benefits galore
With deduplication, only “unique” data is written during the backup process, which means that signifcantly less disk capacity is needed on the back-end to store changes. This offers a number of important business benefts such as reduced capex and opex costs.

Most importantly, organisations can reduce backup windows and improve restoration times by using disk versus tape. Recovery time objectives (RTOs) will improve because data can be recovered from disk. With more capacity available, IT may choose to increase the frequency of backups conducted during the day, improving recovery point objectives (RPOs).

“Data deduplication also facilitates server virtualisation deployments as it eliminates much of the disadvantages of server virtualisation projects,” says Surajit Sen, Director–Channels, Marketing and Alliances, NetApp India.

Virtual machine disk images contain highly redundant data and increase storage capacity requirements. Through server virtualisation, enterprises can reduce the amount of servers in their environments and through deduplication they can reduce the amount of storage.

Though data deduplication is an important technology that is quickly being embraced by users as they struggle with issues of data proliferation, enterprises must ensure that they make the right choice while selecting providers.

Making the right pick
It is expected that the amount of storage will be reduced with deduplication solutions. However, the results will vary depending on the technology one chooses and the type of data that is being duplicated. Marketing brochures that claim large deduplication ratios are simply an indication of what is possible in the best case scenario and may not be achievable in each and every environment.

“It is important for enterprises to quantify. Many storage providers offer estimation tools or tests that provide a more realistic expectation of what can be delivered,” says Adrian De Luca, Director, Storage Management and Data Protection, Asia Pacifc, Hitachi Data Systems.

There are a few problems that CIOs can encounter when they look at various deduplication solutions. One of the most critical among them appears in the scenario when data deduplication is applied utilising proprietary formats. Here data is written directly into the fle headers. These headers describe the manner in which data has been deduped and provide pointers to applications. These applications are assured that they will get the right access to the copy of the data required. According to industry analysts the system of pointers that results from these solutions can lead to some performance degradation.

“Our choice of the provider will only be finalised only after considering all aspects of design, including space savings, effciency, performance overhead, and resiliency,” says Vijay Kumar, Chief Manager - Information Technology of Bangalore-based The Himalaya Drug Company which is also looking at a data deduplication solution for their storage environment.

Jain from Evalueserve adds that the choice of a data deduplication solution should be based on how effectively the solution fts within the storage environment. “The solution should come from a reputed vendor who has experience in this technology,” he adds.

Data deduplication will become a critical enabling technology. Besides, it can be successfully paired with other emerging storage technologies including thin provisioning and virtualisation.

ashwani.mishra@9dot9.in