Creativa - Fotolia

Cloud-based data storage speeds enterprise file operations

When enterprise workers compete for file access, the last operation usually wins. With cloud-based data storage, performance gains and file locking render the competition moot.

Cloud-based data storage is not just for CFOs and CIOs looking to cut costs. It has become a powerful tool for developers working to improve shared access to and version control of files for geographically dispersed employees working collaboratively.

"Most people still think of the cloud as a warehouse for data that isn't being used, instead of the data that is being used," said Barry Phillips, vice president of product management at Panzura, a Campbell, Calif., supplier of cloud data management products. "With the performance levels available today, you should use the cloud for all data, not just old data."

The trend is catching on. Spending on cloud-based data storage is forecast to double in the two-year 2016-to-2017 period, according to a survey of IT professionals published by 451 Research in January 2016. The survey also noted that spending on cloud-based data storage during the same period is expected to rise from 8% to 17% of overall enterprise storage expenditures while, at the same time, spending on on-premises storage is forecast to fall from 70% in 2015  to just 58% in 2017.

For developers, the challenge is moving from discrete islands of storage at each corporate location, branch office or retail store to a cloud-based data storage model based on the premise that consolidation of assets improves performance and simplifies administration. In moving data assets to the cloud, the key task for developers and architects is maintaining or improving performance. "You want to run applications in the cloud in the same manner in which you would run them on premises," Phillips said.

For Nelson, a global architectural design firm based in Philadelphia, a legacy of slow file performance made improvement easy. "Anyone working on projects in AutoCAD, Adobe Photoshop and Illustrator, or Microsoft Office from a central share was very dissatisfied and unhappy," said Markus Weidner, director of technology at Nelson. "These complaints were consistent and long-standing."

Use the cloud for all data, not just old data.
Barry Phillipsproduct manager, Panzura

To solve the performance issue and ensure that users were not attempting to make changes to the same files simultaneously, enterprises are turning to cloud-based data storage providers, such as Amazon S3 or Microsoft Azure Storage, fronted by a management controller appliance installed at each location.

Those appliances, from companies that include Panzura and Nasuni, typically employ solid state drive (SSD) caching technology to ensure fast file access. Both companies also offer interactive management consoles for configuration, oversight, and defining policies. Another provider, Cleversafe, acquired by IBM in 2015, is targeting customers that deal with vast volumes of data at the petabyte scale.

Scalability concern abates

Traditional on-premises storage requires spending on expansion as the volume of data requiring retention grows. Keeping ahead of demand means always maintaining and paying for idle headroom. Not so with a cloud-based data storage model that can flex instantaneously on a pay-as-you-go basis.

Nevertheless, planning is essential, according to John Capello, vice president of product strategy at Nasuni, a Natick, Mass., developer of cloud-based data storage products. One concern is spinning up cloud storage resources during testing and migration, then neglecting to spin those resources down at day's end or when no longer needed, Capello said.

"Failure to de-provision resources can result in significant unnecessary charges," he said. "Developers need to have a tidiness to their personality and remember to sweep up at the end of the day." He added that "developers should be comfortable doing their own environmental administration. It is a great skill to have."

So far away, yet so near

Transitioning from on-premises to cloud-based data storage is a straightforward process that should require little to nothing in the way of rewriting application code, according to vendors and users.

Weidner noted that once the controllers were installed and assets copied to the cloud, more than 600 employees worldwide were pointed to what appears as a single mapped network drive. "Employees see their files as being on premises even though they actually reside on Amazon S3," he said, adding that both file read/write performance and user satisfaction have increased.

The performance benefits can stretch believability but are very real, said Phillips. One of Panzura's customers is a developer of video games that distributes large files across multiple locations. "This company was able to cut transfer times from 30 hours to 10 minutes," he said. "It's just crazy."

Next Steps

At what point should file storage give way to object storage?

Which performs better, Amazon S3 or Microsoft Azure Storage?

Tips for culling old, unused data

Dig Deeper on Managing cloud infrastructure