As enterprises consider how they might take advantage of adopting a hybrid cloud approach, AI may be poised to accelerate and shake up the possibilities, says Ruchir Puri, chief scientist with IBM Research. Known for his work with IBM Watson, Puri discussed with InformationWeek how IBM is approaching the digital era, hybrid cloud, and AI. This includes evolving AI technology born out of Watson and teaching AI to essentially speak code.
Puri’s team works on AI technology to assist with data migration from legacy systems and languages such as COBOL to the cloud, and he says AI augmentation could boost productivity through automation of IT.
How did AI become part of IBM’s plans for the cloud?
“Hybrid cloud and AI are two of the most important technologies as we move forward. Their intersection is one of the most important intersections as well. AI has been applied a lot to language, domain, and speech. If you look at information technology, hybrid cloud is the destination where we are going. We’re coming from a very legacy environment where governments and Fortune 500 companies have these legacy systems, and everybody is trying to transition. This particular pandemic crisis we are in has accelerated that journey through digital transformation. What is really needed from an AI point of view is, how can AI help transform my IT in particular?
“Every corporation is struggling with thousands of applications and they’re struggling with modernizing their real estate. Not only from an infrastructure point of view to cloud, but they are also going from a platform point of view and from a code point of view to cloud as well, which involves application modernization.
“Enterprises might have a monolithic, big piece of code with certain libraries that are stitched in and installed in a traditional environment, versus going to an agile environment where changes are constant -- sometimes hundreds of changes are made in a day. Their agility and availability are key parts of that hybrid cloud proposition and cost efficiency. Now we have to refactor code into microservices.
“Reimagining IT operations with AI -- applications need to be available. The cost of some applications being down can be as much as $300,000 per minute. They can be that expensive for the corporations. More importantly, as we deploy more and more applications in the cloud, they can be operated much more efficiently with automation with AI. Think of AI as an assistant that is observing applications all the time. It knows the history of those applications in terms of incidents that have occurred. It knows the topology of those applications and it is also observing the health metrics of those applications. It can coordinate it to the context you are in and tell you ahead of time if a problem was going to occur. In a more aggressive scenario, it can not only predict the problem but proactively fix the problem.
“Another area where we have rolled out a breakthrough capability is Watson AIOps. As we continue to expand on it, we are linking it to the DevOps process and the upfront changes being made. Changes are many times causes of incidents. Predicting whether a change is risky or not ahead of time and assessing the risk of that change is vital.
“The ability to understand machine languages -- such as Java, Python, and Go -- just like we understand human languages is what we call AI for code, which is the foundation of all this. It feeds into application areas such as AI for application modernization, AI for security, reimagining IT operations with AI, AI for compliance, and AI for developer experience to help developers with performance issues. To be able to do code reviews and transliterate language for them automatically. That intersection of AI and software engineering is going to fundamentally transform the journey of enterprise clients to an accelerated path, unleash innovation, and the value that exists in that hybrid cloud environment.
What presumptions do enterprises have about the cloud and hybrid? How do you bridge those presumptions with the capabilities you are introducing?
“Everybody buys into the value proposition, but the journey is more complicated. Enterprises have data in various places; it really all starts with data. Data sometimes cannot move from the places it is in for privacy reasons, for regulatory reasons. A lot of the industries we work with are regulated. Finance, insurance, and health care. The first barrier they encounter is they buy into the value of cloud but have these constraints. This is exactly where hybrid comes into play. It’s not just about a journey to a public cloud. It is really about a journey to a hybrid environment with a single pane of glass. A single experience that exists across your applications wherever they might be, and they might even span multiple clouds or span private environments to a public cloud.
“You can have these applications in any environment, but they can be managed by a platform that is open by definition and construction. This is precisely why IBM invested in Red Hat because of its openness.
“What enterprises also encounter is if they choose to move only 20-25% of their workloads, they are stuck with 3,000 applications and don’t know what to do with them. This is the struggle enterprises across the board are having and precisely why we introduced the accelerators we’ve launched.
What are some key drivers behind your work?
“The recent incarnation of AI technology really started with perceptual modalities. Languages -- text, speech, spoken language, visual expression and images. It really started with, ‘what is the language of machines?’ I would say it’s code and code can be expressed in very similar ways to human languages.
“The ‘aha’ moment came when we started applying technologies we had been working on in the language side to code to be able to detect problems and to be able to translate code as well. Just like there is AI for human language, there is AI for machine language.
Are there boundaries that you want to move beyond in order to realize the full potential of hybrid cloud? Where can you go from here?
“In hybrid cloud, the key benefits are agility, availability, resiliency, productivity, and cost efficiency. Where we are going is about automation with AI. Developers spend 60-65% of their time testing code, debugging code, reviewing code, taking it through the CI/CD (continuous integration/continuous delivery) pipeline, and struggling to know if the code is risky. How do they know if changes are fine? If something goes wrong, everything goes into a lockdown mode. AI can be the assistant you want reviewing your code and explaining why code is risky. I think the breakthroughs are going to occur as we move forward with AI being applied to so many manual tasks that today are labor intensive and error prone.
How will this affect the pace of change? Will we see an exponential evolution even though some organizations are still trying to adopt, adapt, and deploy in digital?
“Part of the future is here today, which is the hybrid cloud platform in terms of that single pane of glass. The future really is about that journey and that is not just a month’s journey. Many enterprises are on a multiyear journey. The urgency of that journey is only getting accelerated. We’re crushing that five-year journey into a two-year journey. What is really needed is that extra push and breakthrough with AI.
What will you focus on next?
“AI will transform and even disrupt software engineering itself. When people say software is eating the world, I would say AI is eating software. That intersection is tremendously exciting from a science point of view and has massive ramifications for the world in a future where AI will be observing software, writing software, and automating software.”
For more content on digital transformation, hybrid cloud, AI, and DevOps, follow up with these stories: