5 tips to ensure cloud data quality

Choose neutral platforms and read providers' fine-print. These are steps companies should take before leaping to cloud, say data management vendors.

Choose a neutral platform and read the fine-print in the cloud provider's SLAs (service level agreements) to maintain data integrity in the cloud, say industry voices.

The advent of offsite data residing on the cloud has raised concerns about data management and integrity. Earlier this year, the U.S. government held a two-day meeting discussing topics such as data management in the cloud, and the implications of data privacy across geographic borders.

When contacted, data management vendors offered several tips on how companies should handle a move to the cloud.

1. Don't take your problems to the cloud.
Suganthi Shivkumar, South Asia managing director for Informatica, said: "The cloud is just another medium [and is] driven by economics. It will not radically change data quality."

That is, companies should not take the move to the cloud as an overhaul because problems from badly-kept data will still exist in the cloud, said Shivkumar.

Companies should first ensure data is complete and accurate, and has passed through all traditional data hygiene checks, before preparing the move to the cloud, she said.

2. Choose a neutral platform to avoid lock-in.
Ashok Munirathinam, solutions and services manager at Sybase Singapore, said companies should make data available on various environments to support the myriad applications on which enterprises depend.

Off-the-shelf applications often come with proprietary databases and platforms, said Munirathinam. Similar to an on-premise set up, enterprises should choose a cloud infrastructure provider that allows a heterogeneous operating environment capable of running operating systems such as Linux, Windows and Solaris, for example, he added.

Developers have also noted that a neutral platform will help data move between online repositories, should a user choose to switch cloud providers in future.

Shivkumar added that companies should select a platform-neutral central repository so that users can access company data from different platforms.

3. Check if cloud provider has data management SLAs.
A cloud provider should spell out its data management practices in its SLA, said Shivkumar. "Companies should be more vigilant and ask: 'Is the provider able to conform to my best practices?'"

A cloud provider that is able to provide data updates in real-time will help maintain data "freshness", she noted.

Munirathinam said cloud availability is also important. The ability to perform cluster and remote site recoverability in seconds will help reflect updates and allow users to access data, in the event of an outage.

4. Maintain best practices and control in the cloud.
"Whatever best practices the company has, should stay, even when data goes to the cloud," said Shivkumar.

When they audit cloud providers, businesses should measure how much control they are losing over the data-cleansing process, in order to maintain flexibility, she said.

Marrying the organization's existing best practices with the tools offered by a cloud provider, will give a company flexible management of its data, she noted. "Flexibility will be delivered through a holistic data integration plan," she said.

Munirathinam said enterprises should choose a cloud that carries scalable tools, to provide the elasticity touted by cloud providers. Database management, as a "key component" of the cloud, should also be scalable, he said.

5. Learn from other companies that have gone to the cloud.
Shivkumar said network issues in parts of Asia will likely hold back some of its countries from moving to the cloud as quickly as counterparts in the West. By that token, "we can learn lessons from Western early adopters", she said. "Adopt best practices, and learn from [others'] mistakes to leapfrog the competition."

Munirathinam said companies can choose two different approaches to data migration--to choose a similar platform on the cloud as the on-premise platform, or synchronize and consolidate data onto a single platform, which she said, typically requires a longer migration path and more investment.