Data Assurance in the Age of AI

dezembro 21, 2024

2

What do you envision when you think of data in motion? It’s a beautiful, abstract idea.

Vinay Prabhu, Director of Product Management at Graphiant is our latest guest on TeleGeography Explains the Internet. He recently joined us to discuss the intricacies of moving data around the globe.

Vinay outlines the reasons you should be thinking about your data in motion and what Graphiant is doing to ensure its safe passage, particularly as AI becomes more prevalent.

The purpose of networks, of course, is to move data around the world, but we often think about data and applications as what we do with them at the edge or how they’re stored in data centers. Vinay helps us consider what happens to data between those two points, which is crucial. More and more, it’s a problem enterprises must solve.

We’ll share a taste of our discussion here and link the whole episode below.

Greg Bryan: So if you’re talking about getting your data from whatever starting point to to whatever end point, how do you at Graphiant define data assurance?

Vinay Prabhu: A really good question, right? So we see data in three phases of its life cycle. So there’s Data at Rest sitting in databases; there’s Data in Processing that’s being worked between application servers and so on and so forth.

And then there’s the data-in-motion piece where the data really moves from your private space into this ether of the internet, going into the public domain and then hopefully reaching the destination it’s supposed to reach.

Graphiant really wants to focus on the data-in-motion piece of it. The network should be participating in not only the data mobility piece of it, but the privacy and compliance piece as well.

Greg: And that is where data is most vulnerable, right?

I have a good friend in the industry, Jason Gintert from Nitel, who has a blog that he’s had for a long time called Bits in Flight, right? So bits in flight are where we’re focused on the network side, but also where a lot of things can happen, right?

Vinay: Exactly. So I like to call it the data information highway.

You get on the on-ramp, get onto this information highway of sorts, and then get off the off-ramp. And it’s on this highway or the information highway where you’re most so exposed and vulnerable.

So having visibility and control over that highway and getting your dedicated lane over there is fairly important, especially in this AI-driven world where data is getting highly decentralized.

It’s no longer just your data. It’s data that you’re distributing with LLMs and GPUs, B2B in a B2B kind of format. So, I like to say that AI is a data-hungry monster. The more you feed it, the more you can get out of it. But that also exposes you to a lot more vulnerabilities and privacy risks as well.

It’s no longer just your data. It’s data that you’re distributing with LLMs and GPUs, B2B in a B2B kind of format. So, I like to say that AI is a data-hungry monster. The more you feed it, the more you can get out of it. But that also exposes you to a lot more vulnerabilities and privacy risks as well.

Greg: Yeah, absolutely. So, you know, of course, everybody listening to this show has most likely been involved somewhere in the telecom industry. And, first to mind is going to be security issues like deep packet inspection and things like that. But what do you see a Graphiant as some of the main specific drivers for being concerned at the enterprise level or otherwise with data assurance in this current environment?

Vinay: Sure. So I see three big factors, right? Data mobility is important. We need the any-to-any connectivity today more than ever. It’s businesses talking to businesses, machines talking to machines, users talking to machines as well.

So we need to maintain that flexibility to allow mobility of data. But we also need to ensure privacy of the data. So we need to know that it’s reached the destination it’s supposed to. It’s encrypted. Nobody else is snooping on it.

We know the harvest now, decrypt later attacks prevalent now. Infrastructure is at threat. So the infrastructure is getting pressurized as well with this large amount of data decentralization that’s happening.

So privacy is a core concern, and then compliance, staying compliant. Keeping up with the regulations is a big challenge for any enterprise that’s really moving data in bulk today over the infrastructure.

Greg: Right. Like GDPR, for example, and things like that.

Vinay: Exactly. The GDPRs, HIPAAs of the world. So the PCI compliance regulations both form a governance and a government-centric compliance regulation versus an enterprise compliance regulation like PCI.

So you have a lot of compliance-oriented needs, which are getting even harder to maintain and monitor with the way data is moving in this world today.

Listen to the full episode below.