Following on from Part 1, Ryan Noik (RN) sits down with Cisco’s Nabeel Rajab (NR), to gain a deeper insight into the newly released Hypershield.

In Part 2, Rajab explores the impact AI is having on datacentres, computing and security and explains how Hypershield is paving the way for an AI-infused future.

RN: It was mentioned recently by Jeetu Patel that the role of datacentres needs to be reimagined, in terms of how they are connected and operated, as well as how they are protected and scaled. Why is that?

NR: If you think about applications and where we came from with datacentres hosting those applications, you can see there's been a seismic shift in to where we are today.

As an example, consider: What is Instagram? What is what is an application?

Is it just the application that sits on your device? Or is it back-end servers that provide content to you, that store that content along with user information, that can process images? For example, there's all these filters that people apply to their images and a lot of security that goes around it. So as to what an application actually is, it can be difficult to define. Today, we have applications that span the entire globe. Yes, there's a front end, which is what we as users see, and what we can visit as a website, or we could download the user interface on our phone or tablet as an app.

These may have some Point of Sale (POS) systems for retailers, enabling them to have a mobile POS terminal. But on the back end, it's made up of so many different components. For high availability, you'd have redundant data centers that can serve the same sort of services for that application, so that if one goes down, you can failover to another data center. Now, if you think about these modern types of applications, they can span hundreds or thousands of different components, what we refer to as micro services, that build these applications.

These micro services are also tiny little pieces of the puzzle, tiny little pieces of the application. That is what applications are like today. And those little pieces can exist in the public cloud, they can exist on an on-prem private data center, they can exist as SAS services over the internet. When we take that into account, then we realize that we've had to adapt our data centers to be able to meet that scale and to be able to serve those components to the end user.

That means we've had to change the way we do data center design and operation maintenance. And, as the old adage goes, we're only still playing on the back end of the chessboard.

Where we are actually going with new tools, and with the future of applications, with the whole revolution that artificial intelligence is bringing, means that data centers are going to evolve by orders of magnitude going forward.

And knowing that things like AI are going to change data centers exponentially, and how we host applications exponentially going forward, then the question: becomes how do we adapt and ensure we are prepared for that with the right protection mechanisms, so that we can ensure security and reduce risk for all the things that we haven't even seen yet.

RN: Do you think AI is causing the Internet to evolve to its next stage, which is the putting pressure on data centers to likewise evolved to be able to deliver content services like they never have before, and evolving to the next stage of delivering information?

NR: I think that is exactly what we have taken into consideration. Because if you look at AI, you'll notice that you don’t have AI without data.

All of AI is built on data; you must be able to collect data, you must be able to ingest, structure and filter that data, to really cleanse that data. There's a whole new industry called data science and data engineering, specifically to support these AI use cases going forward. AI is fundamentally going to change the way data centers are built and operated. Because when we work with data, and there's an explosion of science and engineering around data, it means that the data centers and hosting environments must be able to adapt to support that industry.

And what we notice is that when we work with data, it is heavily resource intensive. So that's why we have companies like Nvidia, who are enjoying explosive growth at the moment. It is due to the fact that there are very specific types of equipment and components that are required to be able to process these large amounts of data for use cases that AI bring to the party. So yes, AI is changing the industry completely.

RN: Is that changing computing itself? Because if it's charging at the chip level of what's required, it seems that this trend may be changing the fundamentals of what computing is?

NR: it is, I would say it is definitely adding on to what is expected. I think that traditional ways of computing and delivering information and services to users and to the market, that I don't think is really going to change. It's the way we process all that data that's really being I would say, expanded on. And that's why we see things like GPUs, graphics processor units, because of the way that it parallelizes information processing, that we can do it so much quicker, for specific types of use cases.

AI is built on these machine learning models that require not necessarily, you know, sequential processing, but require a lot of processing. And that's why these graphics processor units that can parallelize all of what is fundamental to how we process data, specifically for AI models, for example.

And that's why these types of components that are now being added into the datacenter environment or the hosting environment, are adding such a strain, because they still need power, they still need protection, they still need all the fundamentals around datacenter design, such as cooling, for example.

However, the resources required are on this exponential growth curve. And it impacts organizations that are trying to adopt these methodologies and adopt AI within the organization or build AI tools for the market.

It exposes them to higher levels of risk because they are adopting methods that were not previously used before, and they're growing at an exponential rate. So that is exposing the attack surface. And this is what we're trying to do in terms of mitigating those risks, not just for now, but allowing us to be as flexible as we can in the future, to mitigate those risks and protect those environments for threats that we don't even necessarily know about yet. So that's the whole idea behind Cisco Hypershield.

RN: Do you foresee having AI infused security solutions as being essential moving forward?

NR: Yes, I think that is required now going forward, I think AI is so fundamental in so many things that we do. You know, I think we see these AI tools being increasingly adopted. This begs the question: why don't we do that in respect of security?

From the perspective of security, we have the tools, they are maturing, and we have the capability, because we have the data. Organizations have a lot of data. It's the ability to be able to monitor that data and use it effectively within that organization that's the important part. AI, I think, is, is a game changer for so many different reasons. And, one of the reasons is because it allows us to effectively use the data that is within our organizations.

So AI is certainly not going anywhere, I think it's it really is a fundamental shift in the way people are operating and going to operate going forward. And security is not immune to that. The cybersecurity space is going to fully leverage AI in the future because we can't scale without using it.

RN: Do you think that the introduction of Cisco Hypershield will help companies that are grappling with AI and what it means for their business? And do you think it will offer some peace of mind that it's not just bad actors that are using AI to proliferate threats but also good actors taking advantage on the technology to improve security as well?

NR: Yes, definitely. I think, you know, people should understand that there's always going to be a push and a pull a little bit of a, you know, a fight between, you know, good actors and bad actors. And, and I think, you know, whenever there's change, it'll always brings risk.

We understand what those risks are because we are doing more and more in the space to understand what effects these changes are having on the oarket, on the industry, and on users’ experience. And we're doing a lot of research in this space. At Cisco, we have access to a lot of data, because we firstly have a huge user base, we have a lot of feeds that come in to our analytics engines, especially in the security space. So as I mentioned earlier, we have our Telos security and threat agent intelligence agency, and we have hundreds of researchers, purely doing research in the data that they see coming from security events, and from, security environments around the globe, whether that consists of threats or vulnerabilities.

We are really optimizing the way we are using that data with technologies such as machine learning and AI. We have been using these tools for quite some time. So definitely any organization that is feeling a little bit uneasy, and thinks that, you know, this new world with all of these new age tools is really going to be a difficult space to be in. I'd like to give them some comfort that we've actually been doing this for the longest time and it's now maturing and we are starting to completely redevelop services, using what we've learned so far.

RN: What is the one takeaway that organisations and businesses should most consider from the launch of Hypershield?

NR: I think the key message that I would want to send to the market is that things are fundamentally changing. The application and data center space has significantly shifted and is going to significantly shift going forward. And I think, in a lot of negative ways, there are bad actors out there, that are already using AI, to attack these environments. There is a lot of risk that comes with this expansion.

What I'd like everyone to know is that we also have access to those types of tools, we also have access to AI, we will also have access to all the types of innovation using data that we can get out of organizations, for their benefit, or just generally across the globe. We are building tools to meet that need. We are building tools like Cisco Hypershield, with AI in mind, from the ground up completely reimagining the way we do protection for workloads and data centers.

We are thinking two steps ahead of what could happen in order to make to ensure that we have the right amount of protection and availability, and the right balance between those two, for the future.