To many companies, cloud technology means freedom. Instead of being in the business of running data centers, they can focus on delivering the apps that drive business value. But app portfolios are becoming larger and more diverse, requiring different cloud solutions. The result? A potential for greater complexity when it comes to managing apps and data in a multi-cloud environment.

So what does it take to tackle this multi-cloud management challenge? To find out, we sat down to talk with VMware thought-leader Purnima Padmanabhan.

As vice president of product management in the Cloud Management business unit of VMware, Purnima Padmanabhan helps companies orchestrate workloads across multiple cloud environments. By leveraging automation, artificial intelligence (AI), and machine learning (ML), Padmanabhan and her colleagues deliver solutions that streamline and simplify cloud management, reduce cost, and boost agility. As a result, companies can focus their attention on strategic business transformation and the innovation that creates value for their customers.

At university, you studied both computer engineering and classical music. Do you see connections between these two fields?

I studied Indian classical music, which involves a lot of improvisation—somewhat close to jazz. It’s highly creative, involving a mix of structured and unstructured elements. As a musician, I looked for ways to weave together notes to tell a story and form something beautiful.

Much of my work as a leader at VMware focuses on orchestration—bringing together different disciplines and elements to provide a complete solution that truly delights our customers.

There’s another sense in which my musical background is helpful. At VMware, we place a high value on diversity—not just ethnic or gender diversity, but diversity of experiences and thought processes. Diversity is critical for innovation. If all of us are manufactured from the same mold, we’re likely to come up with a limited set of solutions.

By contrast, when you bring together people with diverse backgrounds in technology, or music, or philosophy, you get different perspectives on how to solve difficult problems. Sometimes it means you can empathize with a customer problem in a way that lets you craft a better solution.

And what is your role at VMware?

I am responsible for cloud management products and solutions.

Our world today runs on software. To adapt to this reality, companies are reinventing themselves as software companies. At VMware, we believe that, in order for this incredible business transformation to continue at a rapid pace, companies must ensure that all of their software runs on a cloud. Leveraging cloud enables organizations to focus on business transformation rather than the bits and bytes of managing infrastructure.

But setting up and operating a private cloud can be challenging, involving many complex decisions around automation of critical services; management, optimization, and scaling of infrastructure and services; and maintaining consistent SLAs and security. This complexity is further compounded with the addition of public and managed clouds—requiring organizations to also secure, optimize, and broker services across clouds.

The role of my business unit is to make all that complexity invisible so that our customers don’t need to worry about delivering or managing these clouds. Instead, they can focus on what is critical: building their business software and applications.

We’re simplifying complexity further by AI and ML. If anything goes wrong in the cloud, AI and ML can quickly identify the root cause and remediate the problem. This should make it even easier for customers to simply focus on building and delivering business applications.

What is a hybrid cloud and how does VMware help customers manage a hybrid cloud environment?

Hyperscale public clouds offer many advantages—for example, supporting rapid scale-up of infrastructure and enabling the new blistering pace of innovation—without requiring an investment in on-premises infrastructure. But there are many reasons to have a private cloud. Some security and compliance use cases require private clouds. For example, in the European Union, companies may be penalized with expensive fines if regulators find that private customer information has been stored outside of private data centers.

Today, a business may choose to run its applications in a data center, at the edge, on a managed cloud, or in the public cloud. Hybrid cloud refers to this ability to mix and match cloud environments based on specific business needs and to deploy, operate, and manage workloads across all of these clouds seamlessly. VMware’s hybrid cloud approach helps customers look at their various pools of capacity across all of their clouds as one contiguous cloud that has consistent infrastructure and consistent operations.

In the past, it was very difficult and very expensive to create, manage, and maintain a hybrid cloud. Today VMware offers software solutions that significantly reduce the cost of maintaining a hybrid cloud by automating key operations and giving customers a single “pane of glass” through which they can view activity across multiple clouds—both private and public.

Ultimately, customers choose to combine public and private clouds not because they care about the cloud per se—that’s just a means to an end—but because this hybrid solution meets their application needs. Our promise is that we can deliver a management system that allows customers to work in whatever combination of cloud environments they choose.

How can VMware’s cloud management tools help companies run their businesses better?

Instead of talking about technology, we always start with the business problem we want to solve.

In our own case, VMware wanted to build applications faster so that we could innovate more quickly. We happen to be a software company, but every one of our customers has the same issue. In a sense, all companies are software companies, because software drives innovation and top-line growth whether you are a utility, a bank, or an airline.

To innovate with software, you need to be agile. Practically, that means you need to build and run your software in the cloud, whether that’s public, private, or hybrid. If you’re not in the cloud, you’re losing valuable time figuring out how to set up machines, do virtualization, and leverage automation, while making sure your systems are secure, scalable, and resilient.

With our cloud solutions, we free customers to focus on the more interesting challenges of business transformation and creating the software that creates value for their business.

This same dynamic has played out in our own organization. We made the decision to build our cloud management software in the cloud. To make this work efficiently, we had to learn new skill sets, reorganize some teams, and hire some new talent.

With this combination of technology, people, and process, we gained the agility and velocity we needed. Before, we were releasing software updates every 6 to 12 months. Now, we can push new code daily because our development teams do not need to think about how or where their software is deployed. It’s all automated. They just write the code, build good apps, and check in the code. And then management software picks it up, deploys it onto a cloud, ensures it’s within appropriate security boundaries, and makes sure it can self-recover in case of failure.

How will AI and ML impact cloud computing in the near future—or are they already having an impact?

ML has been part of our product line for some time. It’s already deployed at scale by our customers to solve what would be hard problems. For example, we use ML to look through and correlate thousands of metrics and to automatically identify the few metrics that might point to the root cause of a problem. ML allows programs to examine millions of lines of logs, auto-categorize issues, and even suggest remediation actions. ML underpins our vision for a fully automated data center with the resilience to identify and recover from failures on its own. Today, most of our products use AI/ML to automatically process large data sets.

More recently, we have expanded our use of AI to test the application of reinforcement learning techniques to create auto-tuning infrastructure, as we announced with Project Magna in a tech preview at VMworld San Francisco this year. This is an automated system that constantly monitors certain key performance indicators (KPIs), such as web page loading speed or app response time. The AI can automatically tune the infrastructure software of a system to keep that KPI within a desired range. When the AI makes changes that have the desired result, it receives positive feedback—the reinforcement part of the learning curve—and keeps tuning the system to get better and better outcomes for the specified KPI.

The ideal cloud should be self-service, secure, scalable, resilient, and self-optimized, with a rich set of services. AI and ML are perfectly suited for solving the self-optimization part of the equation.

With our virtualization solutions, we have delivered a powerful infrastructure that has a rich set of capabilities that can be fine-tuned to meet customers’ business needs. But with all these controls, how does an organization ensure they have the perfect infrastructure settings to deliver optimal performance for all their apps? How do they ensure that tuning infrastructure for one type of workload does not compromise another? And how do they react and optimize based on workload changes? These are important questions, with an exponential set of potential answers.

Project Magna addresses these challenges. Leveraging the AI/ML techniques of reinforcement learning, Project Magna creates a model of the infrastructure based on learnings derived from many different customer environments. Leveraging this model, Project Magna combs through the potential set of answers and identifies the perfect infrastructure settings for given workloads, and as the workloads change, it changes the settings dynamically.

How does cloud management impact your clients’ businesses?

Whether I’m dealing with internal or external customers, I’m laser-focused on customer satisfaction and customer value. It doesn’t matter what a technology does unless it can deliver value to the end user.

For example, we had one large pharmaceutical client that needed more agility to accelerate its software development cycles. Due to stringent security protocols, it took this company six weeks to deploy a single virtual machine in the cloud as the security team performed a security audit and the networking team made sure firewall boundaries were set correctly.

We showed this client how to automate these tasks in a consistent and reliable process. Using automation, our cloud management software took the timeline for this 48-step process from six weeks down to six minutes—all while adhering to the same security standards and approval guidelines. The software enabled the client’s IT team to turn around requests from its internal clients—software developers—at a much higher rate, while also helping the company to become more agile and innovative.

What do you like most about working at VMware?

I’ve worked at many technology companies, both small start-ups and large firms. One thing that really stands out to me is that VMware has a very kind culture. This kindness provides an environment that nurtures talent and encourages teams to feel safe trying experiments. Ultimately, these experiments generate innovation that delivers value to customers.

What do you see as the most exciting developments—either at VMware or more generally in the fields of cloud computing and virtualization?

Various players including VMware are working on building a “self-driving” data center that would determine its own attributes automatically and deliver results with low human interaction. This data center would be able to self-tune and self-heal. It would run itself and manage itself. And if it ran out of options, it could call for human help by itself. There’s a huge amount of work being done in that area, with advances largely driven by AI and ML.