Why are more enterprises running data science programs in the cloud?

A solution sheet from Immuta outlines the benefits of running data science programs in the cloud

A solution sheet from Immuta outlines the benefits of running data science programs in the cloud

It is evident that lifting and moving data science operations to the cloud can be expensive and risky. Nevertheless, Immuta’s latest release enables companies to save more than 60% of cloud infrastructure costs when running data science programs in the cloud.

Running data science programs in the cloud

More companies are now utilising the power of the cloud in order to accelerate their data science programs. Indeed, Immuta insists that this is due to the elastic nature of the cloud.

The cloud also provides the opportunity for companies to experiment with new storage solutions, compute services, and machine learning. In turn, these frameworks are reportedly driving better business outcomes.

As major cloud players are rapidly introducing new innovations, more organisations are choosing to train their models in the cloud. Research Vice President at 451 Research Matt Aslett insisted that “we are seeing a growing trend towards cloud adoption for analytics workloads,” particularly within data driven organisations.

Although the cloud is undoubtedly useful, there are reportedly three key areas where enterprises are struggling. Indeed, Aslett added that “privacy and security issues continue to be of concern” for organisations.

Challenges to cloud-based data science

Time is also a prevalent issue as users move and copy the data required for training models from on-premises to the cloud. As a result, compliance concerns and decisions on where that data should land can be complex.

There is also an element of risk due to the way in which some companies build custom apps around each on-premises data source to uphold legal and compliance controls. While legacy apps are not moving to the cloud, the data that must move ultimately has “zero to minimal controls.”

Finally, users should execute data policy enforcement manually by creating “anonymised” copies of data sets, across various user types. However, more user and policy combinations means users must make, store, and manage more copies in the cloud.

In turn, this can potentially raise the costs for training artificial intelligence models by ten times. Overall, this eliminates the perceived benefits of utilising the cloud for data science programs.

In order to solve these issues, Immuta enforces dramatic policy controls on databases, Hadoop, and file systems. This consistent control plane allows data scientists, legal & compliance, and data owners to work together in a “frictionless, symbiotic environment.”

Looking to enhance your data strategy? Check out the Top 10 Companies Unifying Data in 2019