By Aaron Lazenby
The emergence of enterprise cloud computing is no longer a matter of debate—and the numbers bear that out. Research firm IDC found that companies are acquiring cloud services at more than six times the rate of IT spending, while Deloitte predicts that IT-as-a-service will account for more than US$500 billion in economic activity by 2018.
As the leader of one of the world’s largest cloud development organizations, Oracle President of Product Development Thomas Kurian has an enviable view of the cloud economy. And whether in private conversation or on the public stage at events (such as Oracle OpenWorld), Kurian makes a strong pitch for Oracle’s position in the marketplace.
“We’re the only cloud vendor that provides infrastructure as a service [IaaS], platform as a service [PaaS], application software as a service [SaaS], and data as a service [DaaS]”,” says Kurian. “We have an unmatched breadth of functionality and integration across our platform and application portfolio. And we can run that cloud stack in any data center. This affords us some very unique capabilities as we look into the future.”
In the days after Oracle OpenWorld 2017, Kurian spoke to Profit about some of those unique capabilities—including the impact of emerging technologies such as artificial intelligence (AI) and the Internet of Things (IoT) on enterprise cloud computing.
Profit: Oracle is ten years into its cloud journey. What have we learned from customers about our cloud portfolio?
Kurian: Customers today understand the value of cloud computing—how much money you can save, how fast you can deploy workloads, and how easy it is to run and use applications without having to manage the systems themselves. All of that is clear at this point.
Now, customers are making requests for novel uses of the technology. I’ll give you an example. We are currently working on a project with an oil and gas exploration company. They are moving hundreds of terabytes of data to Oracle Cloud as the foundation of their exploration and science algorithms. That critical part of their business uses Oracle Cloud Infrastructure to run the computations, and they’re using our analytics and Oracle Big Data Cloud Service as a platform to manage the large data assets that they’re loading into the cloud.
We’re also working with a grocery retailer who wanted to connect our software-as-a-service applications to their mobile user base. They wanted to use a bot to make that connection. So the bot and the user interface run on our platform as a service connecting up our SaaS applications.
Cloud architecture must be engineered with future use cases in mind. Otherwise, you’re going to be trapped in a cloud that is not keeping up with the market—and that would undermine one of the core benefits of cloud computing.”
There are hundreds of such examples: supply chain companies using our IoT platform as well as customers activating the AI technology inside our applications and then using that as part of a platform to do trending. These use cases largely reflect what an IT footprint looks like within a traditional corporate enterprise. IT needs the whole stack, but they can’t say, “Hey, we’ve got to go to four different clouds to do that.”
Profit: How does Oracle’s cloud portfolio support customers’ need for the flexibility to adopt new technologies as they emerge?
Kurian: Well, take AI and machine learning as an example. These are areas of enormous importance as we go forward, and our strategy is threefold. First, we’ve put very fast computers in Oracle’s IaaS, based on a technology called graphical processing unit [GPU] that runs machine learning algorithms with extremely high performance. So the computation is lightening fast. Second, we’re building out a PaaS so data scientists and experts can build applications on top of their own algorithms or off-the-shelf algorithms.
Third, we have taken a specific set of AI algorithms and integrated that into our SaaS applications. There are a number of business processes these AI algorithms could help with. For example, an AI algorithm could tell a marketer the best message to send to a lead. It could tell sales reps which of the many leads they should focus on closing or what’s the best product to recommend. It could tell a customer service representative what to say in response to a customer’s query. All of these are scenarios where human beings write very sophisticated, complicated rules to try and drive what a person should do. In our view, algorithms can do it a lot better than just human-encoded rules.
But the overall cloud architecture must be engineered with future use cases in mind. Otherwise, you’re going to be trapped in a cloud that is not keeping up with the market—and that would undermine one of the core benefits of cloud computing.
Profit: How is Oracle uniquely qualified to help companies introduce emerging technologies such as AI and machine learning into their business processes?
Kurian: We understand the application’s function domain in great detail, so we can tailor unique algorithms for specific domains. For example, traceability is important with an AI algorithm that’s assisting a human with workflow tasks—you must have clarity on why the algorithm recommended or rejected something. This is just a fact we know because of our work with our customers.
It’s also very important to optimize algorithms for the domain, so that you avoid statistical skew and selection bias. For example, algorithms on ecommerce recommendations are relatively simple to do with a large number of shoppers and an assortment of products. But what if you have a really small number of customers and a small assortment of products? You have to be careful to avoid statistical skew in the recommendation that the software is making.
IT needs the whole stack, but they can’t say, ‘Hey, we’ve got to go to four different clouds to do that.’”
It gets even more complicated when you introduce a new product, because the state of the art in algorithms is looking at what an individual and people like that individual are buying, and then making a recommendation based on that information. But nobody has bought the new product yet, so how do you introduce novelty into the algorithm? Our experience in enterprise workloads and enterprise data helps us understand and create machine learning algorithms very differently.
Profit: What role does AI and machine learning play in the area of security?
Kurian: The fundamental heart of a security algorithm is an understanding of what people are doing what operations on what systems in the organization and whether any of those operations are malicious. Our technology uses AI, particularly machine learning and clustering, to understand the patterns by modeling systems. What are the relationships between systems? Who is accessing these systems? Why are they accessing these systems? Our AI allows us to very quickly compare anticipated behavior against a user’s actual activity within the system to flag whether the user is nefarious or authorized.
We also applied AI and machine learning to the Oracle Autonomous Database solutions, which Larry Ellison announced at Oracle OpenWorld. There are many mundane tasks required to operate a piece of software. There’s a certain tedium associated with these tasks, and staff often does not perform them consistently. They don’t really add competitive advantage, so they are not a priority to senior management. But if you don’t make these mundane tasks a priority, you can introduce significant security vulnerabilities to the system.
So instead of having a human do all of those tasks for a database—install, configure, patch, upgrade, back it up, and encrypt it—the software does this automatically on behalf of the user. This gives you greater predictability and accuracy in the way systems are configured, at a substantially lower cost.
Profit: Aside from AI and machine learning, what other emerging technologies do you see having the greatest potential to radically change business in the decade ahead?
Kurian: Three of the most important new technologies we’re looking at are chatbots, IoT, and blockchain. The chatbot makes messaging a client to any application, essentially becoming the new browser. It offloads the need for human beings to be involved in very mundane tasks—such as answering basic questions—by having a digital assistant do the job. I believe chatbots will transform the human interface.
IoT is a way to improve business process efficiencies by instrumenting devices with sensors, and it delivers the ability to respond when sensors send information up to the cloud-based system that’s monitoring them. When you are able to monitor logistics and assets in that way, you can make business process more efficient. This will transform supply chains, for example, by digitizing the flow of data from the physical world, which allows algorithms to interact with physical space.
Blockchain is enabling any business process that’s highly centralized in a company to be decentralized or distributed in a fundamentally different way. For example, intracompany transactions could be totally transformed. Today, when two different divisions of a company reconcile their financial positions, they have to do a lot of manual reconciliation of intracompany transactions. With blockchain, you can use the [open source project] Hyperledger as a mechanism to guarantee integrity of data, validating data that may be subject to regulatory or market scrutiny.
Each of these new technologies will transform the modern business in its own way. Our job is to make sure our customers have the platform they need to adopt these technologies as they see fit. That is why I believe Oracle Cloud makes the strongest case for what enterprise IT will look like in the future.
Photograph by Katelyn Tucker/Orange Photography