From weeks to hours: How AI slashes the time to optimize in the cloud

Presented by Cloudreach

Cloud modernization isn’t a destination — it’s a journey. More companies than ever have migrated to the cloud, stemming from the pandemic-led uptick in cloud adoption as companies pivoted to remain relevant and competitive in their markets, to the urgent need to leverage new technologies, new ways of adding value for customers, and new revenue-generating streams.

And that’s where too many organizations have simply stopped.

“Cloud is not a one-and-done type of move,” says Jonny James, senior technical product manager at Cloudreach. “People often think that if you just take that first step, then you’ll be able to realize all those great benefits that you get sold on — better performance, lower cost, better agility, hyper scalability and so on. But you need to do continuous follow-on work from there.”

Unfortunately, companies are running modernization projects as if they were traditional lift-and-shift migration projects, with a finite start and finish. The focus is on pushing on-premise infrastructure into the cloud, rather than re-architecting applications. The migration wave begins when the first server is pushed over, and ends as soon as the last application is migrated to the data center, all on-premise workloads are shut down, the data center provider contracts are ended, and that portion of IT history is left behind.

There’s a real and heavy cost in this short-sightedness: organizations are very unlikely to generate all the benefits they believed they would see from migrating to the cloud.

To really deliver on the promise of the cloud, from democratized data and technology access to setting the in-house team loose to deliver value instead of admin tasks, modernization work goes far beyond just picking up a data center and dropping it into the cloud.

Kicking off the modernization journey

Identifying modernization targets means evaluating a cloud estate for modernization opportunities that are high impact, low effort, reduce cost, increase performance and security posture and improve the ability to innovate. But you have to start to walk before you can run, James says.

“Everybody wants to immediately leverage serverless and containers and implement new technology for understandable reasons,” he says. “But coming to grips with the basics is where you have to start.”

For instance, a good first step is ditching compute instances for managed services which contracts out the responsibility for the underlying infrastructure maintenance work to the hyperscalers, which in turn  frees up internal teams.

That’s because running a database with a managed database service from one of the CSPs, instead of running it on a compute node, means that the database administrator, who was spending upwards of 65% of their time maintaining it, can now take on more databases, manage more at the same time, and get more done. That realizes some hefty productivity gains — or they can switch to value-generating work for the organization.

And from there you can start to evolve the maturity of the types of technologies leveraged within the estate, to see what can be containerized, where functions can be leveraged as a service, add serverless technologies and so on.

“Lean on the PaaS solutions that the CSPs provide — let them handle all the low-level infrastructure maintenance, and ultimately that’s going to act as a stepping stone for you to start leveraging more advanced technologies,” James says. “And then your in-house teams can work out how they can use AI or ML when it comes to driving value for your organization.”

Modernizing without technical debt

According to Gartner, by 2025, new technical debt will have snowballed on top of existing debt, becoming 40% of a company’s current IT budget. That’s the kind of debt that diseases growth and stalls innovation. A company can get out from under it several ways — and it should include adopting that continuous modernization mindset, and focusing on strategy from an application-centric perspective instead of component-specific.

In contrast, the traditional application modernization mindset means that companies are mostly drawn to cost-driven, one-off projects that are often difficult to green-light and have little focus on joining the dots between business value and technological capability.

“Everyone wants to look down their resource sheets and look at the components that cost the most money,” Jones says. “That’s an understandable way to approach the problem. But there’s also a line that should be drawn between technology and business function. That will encourage leaders to help assess an application at a more holistic level.”

These kinds of one-off projects can help address narrow modernization needs but, without a view into the overall landscape, technical debt will still continue to rise. Those costs will trigger yet another one-and-done project that can’t take context into account, and means that technical debt will continue to compound.

This big-project, one-time-effort cloud modernization belief system needs to evolve. That means finding a way to evaluate a cloud estate for continuous modernization opportunities — the high impact, low-effort targets that cut the technical debt principal, reduce cost overall, increase performance and security posture and improve the ability to innovate.

It’s back to learning how to walk before you can start to sprint. That’s where AI and machine learning need to come in.

Modernizing smarter with AI and machine learning

Products that incorporate AI and machine learning can offer intelligent recommendations that are tuned to a business’ own specific modernization strategy, James says. They continually analyze the cloud infrastructure to find the best, most cost-effective opportunities to modernize the cloud, and ensure modernization projects are accurate and thorough.

For instance, it’s essential to understand application dependencies as you modernize. A three-tier application is spread across a database, an application layer, and a web layer, but there’s so much touching that application upstream and downstream. An AI/ML solution can look at the typical services that interact with a LAMP stack application, for example. It can build that into a data set in the back end that an analytics engine can detect and pick up on when it maps applications, which helps make migrating and modernizing far smarter – and ensures a company sticks with its application-centric view of all the components that sit in its app stack.

These kinds of analyses, done manually, can take an organization weeks or months to complete, and as James says, it doesn’t always surface the kind of bang-for-your-buck opportunities that can start to compound bigger wins over time. This real-time visibility and continuous monitoring slashes the time it takes to truly realize the potential of the cloud. 

And the great thing about automated AI and ML solutions is that the product just gets smarter over time, as data feeds and continuously optimizes algorithms. But the strength of the product’s data sets makes the difference. Cloudreach’s own Sunstone product benefits from a decade plus of helping companies across industries migrate into the cloud.

“With all the different servers we’ve touched, all the different organizations we’ve helped to move to the cloud, you can start to infer patterns of what types of companies are going to adopt which kind of modernization and migration approaches,” he says. “You can then use that type of data to feed into these recommendation engines. The combination of AI, ML, and then big data sets, is the holy trinity of really starting to kick up to the next level of cloud innovation.”

Dig deeper. Learn more here about what comes after cloud migration, and realizing the true potential of the cloud.

Source: Read Full Article