Innovations in Cloud Infrastructure: AI Integration and Data Security

Pradeep Vincent, Senior Vice President and Chief Technical Architect of Oracle Cloud Infrastructure (OCI), recently discussed Oracle’s strategies for cloud computing, artificial intelligence (AI), and data security. With expertise in Oracle's cloud architecture and advanced technologies, Vincent provided insights into how the company addresses industry challenges and delivers innovative solutions to its customers.

Leveraging Generative AI Strategy for Industry Challenges

When it comes to GenAI, our cloud really encompasses infrastructure, platform and applications. But at the end of the day or industry specific applications, our customers are looking for us to solve the infrastructure and the integration behind the scenes and that's exactly what we're doing. We have many AI features already shipped as part of our portfolio of apps, industry apps as well as enterprise apps.

Behind the scenes they essentially use OCI GenAI service and our super clusters and GPUs. All of that. But at the end of the day, what customers want to see is not all the complexity. They just want to see the end result, nicely integrated for them and delivered. That's true in industry applications as well. There are specific examples we can point out, we have an example in fusion apps, we have Gen AI base functionality that can summarize the resumes and help pick resumes and create job postings as an example.

And there are many industries specific examples as well such as AI assist for scheduling, automatic schedule generation for project planning purposes. But there are lots of it so it's a little hard to point out everything but the point is that our goal is essentially to bring embed AI functionality in all our vertical apps.

Innovations in Database Management

GenAI is a building block. It's not necessarily the end result. What we want to do is essentially create powerful building blocks so that customers can leverage it how they see fit in order to solve their own use cases. And the select AI functionality in database is one such integration that's specific to database.

It is really a part of a comprehensive strategy where we are pushing GenAI features to all our products so that it brings customers to choice and the flexibility at the OCI level to use whatever they want in the context of GenAI to pull that together.

Balancing Innovation and Stability in Technical Architecture

One of the best examples in this case is Gen AI, because we see a massive rate of innovation that's going on in this space. But at the same time, customers when they come to us, they want and expect stability and availability. And sometimes they can be in conflict if they keep changing the hardware very frequently and introduce nascent features it may not be aligned with our goal to have a stable platform.

We essentially have multiple strategies to offer customers a stable platform, but also innovate that the clip that we need to and as an example with GenAI, we have a stable set of hardware and a fleet, but then we also have innovation labs that we do in close partnership with our customers and partners. There we actually lean forward and go to the bleeding edge and take risks and innovate. And then once we are happy with the stability, we bring it back. The key difference is that this is not just a test. It's hard to test very large cluster of GPUs in isolation because the scale is too big and the complexity of the workload is what drives many of these bugs and issues with these clusters. So we really need to experiment with large clusters almost in the context of production like workloads. So we need to do that in conjunction with our real work workloads internally or with partner.

Prioritizing Security and Privacy in AI and ML Integration

It’s a huge focus area for us. I think having an awesome infrastructure models and apps is not sufficient. We fundamentally believe that security and privacy in the context of GenAI or otherwise, it's actually quite important for Oracle customers. Our goal with security in general is to take the approach of embedding the security functionality by default into the products, not as an optional or add on pack, but literally into the product and then turn it on by default. Customers essentially should not have to think or do anything to turn on the security features. You should just get security features out of the box.

And that's exactly what we're doing. And that's also what we're doing with GenAI. So as an example with OCI generative AI service, the customers might use their data, potentially sensitive data to fine tune their models or as part of their inference stack, we do not use any of the data outside of the customers. In fact, there is really no way at all for them to do.

We actually use private clusters of GPUs behind the scenes for OCI generative AI service, so that every customer 's OCI GenAI stack is completely independent and distinct from other customers OCI GenAI. So we go all the way to essentially separate all the data and ensure that you know there is no sensitive data that leaks out of their environment.

Navigating Security in Hybrid and Multi-Cloud Environments

Two things I would call out. One is security in the context of unified cloud. Unified Cloud is essentially our strategy to have a cohesive layer across apps, industry apps, enterprise apps and infrastructure.

And one of the reasons we do that and we want to do that and we believe it's massively important for customers is because of the security and compliance aspects, we want to essentially have a single boundary and single set of controls that cuts across apps and infrastructure. Internally, that's essentially how we are gearing things up. As apps run on top of OCI, we are leveraging common set of security controls across them, common security standards, fuel and posture that cut across all of them, so they essentially get the same benefits. A very good example of that is essentially dedicated regions as well as our government regions, national security regions.

We came up with those for very good security and compliance goals that the customers are trying to achieve in that region. But that's not just for infrastructure, it's also for apps. And the customers want it for apps for exact reason why they want it for infrastructure. So it's a common set of security and compliance goals that cut across IaaS and apps that essentially allow us to deliver the value to the customer. 

The second thing is what I mentioned in the context of GenAI security as well is the way we look at security, it should be on by default. It's not an optional  add-on pack. And it's always there by default secure. And what we found is that in cloud, many times for convenience reasons, it's easy to ship the default port as an easy, not so secure mode and but then that results in more security challenges. Public objects or bucket is a frequent way by which customers leak their sensitive data. By default, objects or bugs should be private, by default your network should be private, no Internet access. If you really want it, then you can essentially do that to make it public, we have secure ways to do that. But then by default it's not public. So essentially security baked into the product, security on by default.

Supporting Data Privacy and Compliance

That's exactly the reason why we do distributed cloud because we know that and it's not just India, we see this across Europe and APAC in general where every country wants to have data that's in the country or the region and be aligned with the compliance goals of that particular country or region. So that's exactly why we had we came up with dedicated regions.

And I think that aligns with how customers would want to do that in India as well. So we have public regions in India that's already physically in India. And in addition, with other dedicated regions we have, customers can essentially buy a dedicated region from us and we can actually put that in whatever data center they want. Then we can actually put it in their data center.

Growth and Influence in the Indian Market

OCI has been growing at a good clip. We've been launching many products, and the product velocity has been really good. I think our focus has really been on solving some of the hard customer problems that other cloud providers have a hard time solving. And growing effort and that way we deliver the value which indirectly delivers growth. We focus on customers and customer problems and growth will come with it.

Presence in Key Sectors

Healthcare is a big focus area for us and it's a very strategic industry and vertical for Oracle. I think there's lots of innovations that needs to happen both in terms of pure integration, and advancement in terms of the software stack in the healthcare. Two important things I would call out. One is how we bring AI and generative AI technology to the verticals including healthcare and manufacturing.

And the second one is how do we leverage the distributed cloud in the context of verticals. We've had good conversation with bunch of customers in the context of healthcare and like some of the other verticals as well. Having this notion of a dedicated region, hosting specific vertical apps is extremely compelling for initial stage. So that's a big innovation that we plan on driving across verticals including medical.