Many now are telling us that, though all cloud may be cooler, hybrid is here to stay. But why? And how does a company strike the right balance?

We hear so much conflicting chatter about cloud vs. on-premises — pros and cons regarding price, security, etc. Here’s a list of nitty-gritty tips for the confused on why hybrid works for most companies and how they can do it right, according to Manish Dasaur (pictured, right), managing director and North American lead for analytics and AI at Accenture PLC, and Anupam Singh (pictured), chief customer officer at Cloudera Inc.

Dasaur and Singh spoke with Dave Vellante, host of theCUBE, SiliconANGLE Media’s livestreaming studio, for a digital CUBE Conversation. They discussed what they’ve learned about cloud migration and hybrid best practices through working with clients. (* Disclosure below.)

Protips on price, performance, protection

Hybrid rule number one: Don’t randomly move applications to the cloud just because everyone else is doing it.

A company should have a worthwhile business goal in mind, further broken down to something manageable it can actually try and whose success they can measure.

“Start with a use case in mind and migrate the data sets and elements that are relevant to that use case, relevant to that value lever, relevant to that unlock that your trying to create. That, I think, is the way to prioritize it,” Dasaur said.

This will prevent overdoing it and also push the company to give it a go without spending forever planning. That’s important because waiting causes companies to incur technical legacy debt and increases cost and barrier to entry, Dasaur added. 

Hybrid rule number two: Once a use case is decided upon, weigh how on-prem and cloud technologies can facilitate it and ensure desired outcomes.

Workload analysis can help companies make the best decisions about where a workload should live, according to Singh. “If you have a new type of compute requirement, you go to public cloud,” he said.  

The cloud can’t be beat for highly advanced offerings like GPU computing and artificial-intelligence libraries. These can ratchet up app performance for a price not possible with an on-prem implementation. And that’s not all. 

“A lot of the cloud providers enable a set of native services for [data] ingestion, for processing, for modelling, for machine learning that organizations can really take advantage of,” Dasaur said. 

Hybrid rule number three: Put security first, whether on-prem or in the cloud.

Singh has seen clients move data to object storage buckets in cloud and lose all metadata and associated authorization and security policies.

“It’s important that that data, as it goes around the cloud, if it gets copied from on-prem to the cloud, should carry that quality with it,” he said. Cloudera has invested heavily in its Shared Data Experience to enable data collaboration across environments with no loss in security.

Hybrid rule number four: See about possibly hybridizing the same application — in parts and at certain times — if it makes sense.

Computing massive data is often much more cost effective in cloud. But suddenly dumping all data in cloud is probably not the way to go for many large companies. For example, 30 petabytes of data would take a year and a half to make properly aligned backup copies of, according to Singh.

One workable solution is to move just some data to cloud for processing when appropriate. For example, a data scientist might send data sets to the cloud for GPU computing, then pull them back on-prem and program results into a model.

“Sounds like science fiction, but the good news is you don’t need a Chevy to take all that data into public cloud,” Singh concluded.

Watch the complete video interview below, and be sure to check out more of SiliconANGLE’s and theCUBE’s CUBE Conversations. Disclosure: Cloudera Inc. sponsored this segment of theCUBE. Neither Cloudera nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE