Name: Christophe Bertrand | Enterprise Strategy Group
Uploaded: 2024-04-30T15:28:00.000Z
Duration: 15 min 54 s

Clips
More from Building Cyber Resilience on Trusted Data Protection Infrastructure

Christophe Bertrand

Principal Analyst

SiliconANGLE & theCUBE

play_circle_outline Ransomware as a top threat to the viability of organizations

play_circle_outline The shift from disaster recovery to cyber recovery and resilience

play_circle_outline The Unique Challenge of Ransomware Recovery: Reinventing Data Protection, AI, and Machine Intelligence in Backup

play_circle_outline The potential for AI and ML to improve data recovery and RPO/RTO posture

79:02

16:18

11:06

Info

Christophe Bertrand | Enterprise Strategy Group

Christophe Bertrand

Principal Analyst SiliconANGLE & theCUBE

>> We're back with Christophe Bertrand, who's the practice director at ESG, and we're going to talk about reinventing data protection, and the disruptive impacts that AI and machine intelligence have on backup and recovery with some new fresh survey data and research that ESG has done around cyber resilience and recovery. Christophe, welcome. It's good to see you back in studio. We were here in December.

Christophe Bertrand

>> Great to be here.

>> Thank you for having me.Looking well. Yeah, you bet. So you have some recent survey research that we're going to talk about here, but before we get deep into the data, I wonder if you could share with the audience the background of the research, the demographics, the methodology, just take it from there.

Christophe Bertrand

>> So typically what we do is we speak to hundreds of respondents. They're highly qualified IT professionals or cyber professionals depending on the topic. And typically it's North America, Western Europe, combination thereof. It depends on the nature of the research. So what it gives us, it's essentially great data from end users. Not only do we get sentiment data, but we can quantify what's going on, what the challenges are, what the opportunities are. So it's great data for both end users. What are your peers doing? What's going on? You're not alone typically is the reaction we get. But also for vendors, we can get a good sense for what's going on in terms of trends and that gives me the ability to do more research. Where I see smoke as I detect problems, I know there's a fire, so I go do more research on those topics.

>> Yeah, you guys have done this and you've built on it for years, and because you have such domain expertise, you can ask the right questions. Now, a key finding in this recent survey, and if you guys could cue the first slide, was that 89% of respondents rank ransomware as a top five threat to the viability of their organization. But even more striking when I saw this data is nearly two-thirds say that ransomware is one of the top three threats to the viability of the organization. So this tells us that firms really aren't prepared.

Christophe Bertrand

>> Right. And by the way, the question we asked was really around the viability of the business, not about IT or IT infrastructure or anything like that. It was truly about the business. So the numbers are staggering. Why? Because this is a business issue. We're talking about existential type of problems here, extinction level type of problems with ransomware attacks. So what that changes is also the nature of other people who get involved. So here's the good news for IT professionals. They will have the ear of their executives, they will have access to more budget, they will have access to more resources. Here's the bad news, it's a business issue. It means you have new processes, new teams, and you have to really think through the technology. I'm going to say one more time, but here the nature of the problem has changed and that's really what's at stake here. You are critical in IT now in your ability to help de-risk the business. And that's a big deal.

>> So it's not just a technology case, it's not just about hardware failing. It's a whole house type of organizational business case that maybe IT people aren't used to making.

Christophe Bertrand

>> Absolutely. And it's also forcing some conversations across different groups that maybe were not really talking before. Let me put it this way. The minute the security guy calls the backup and recovery guy to say, "I've got a problem, can you help me recover?" Well, guess what? It's too late. You should have been talking before.

>> Okay. Nearly 70% of the respondents in this survey said that recovery from a ransomware attack is fundamentally different than recovering from other types of disasters like fires or floods. We have a graphic on this, it's a pie chart. Why Christophe, do people think this and what are some of the salient differences in your view?

Christophe Bertrand

>> So I like to say that disaster recovery is dead. Now we're really in the era of cyber recovery and really cyber resilience. So it's not just marketing buzz or vendors making pivots, and messaging, and PR. This is real. What's happening is that traditional disasters are not traditional anymore. You could see the hurricane come, you could shut things down and fail over to a secondary center or to the cloud, whatever the case may be. But now things are different. You don't know what's going to hit you, you don't know how it's going to hit you. You don't know when. Could be bad, could be really bad, or could be benign. So the processes have to change, the teams have to change. And what's going on is, I believe today what we're seeing is that because there's so much data, so much complexity in IT environments to begin with, what we would call technical debt, and teams that are not aligned to face these new challenges. Well, people run around with their hair on fire in some cases, and it's a big disruption. So what's the net-net? We talked about it being a business existential issue. Let me sort of bring it back to data protection, back to recovery. The traditional metrics such as RPO or RTO used to be pretty well managed. We got to a good place and now they're just out the window. Because you have to rethink the recovery process, you have to be more preemptive, you have to do a lot of different things that you didn't have to do before. So the combination of cybersecurity, traditional disaster recovery, data management is happening and that's happening against a background or backdrop of a lot of data being generated around AI and a lot of investments going in all new areas of technology. So people are running behind this train constantly. So change is needed and technology can help, but the processes and the people will have to evolve as well.

>> I can't tell you how many times I heard during the pandemic, I'm sure you did too, that we were way too focused on disaster recovery versus just our overall business resilience. So that was a catalyst and people are clearly still catching up based on that survey data. Now you've got the AI wave comes in, so we want to talk about AI and data protection. People don't often think about those two topics going together, but should they, and if so, why?

Christophe Bertrand

>> Right. So I think there are two facets to this. Two sides of the coin, there's the creation of data that AI is generating. So think of it as a new workload for lack of a better term. So as if we didn't have enough data and we're generating even more data, trained data, model data, whether for analytics, for gen AI, you name it, and we've looked at that and it's starting to become significant. Before we know it, it's going to be very, very significant, certainly in the petabytes range already. So you've got to protect that. The reality is it's not really quite happening. Now the second aspect is how do I use ML and AI in the context of easing my data protection, my ransomware recovery? So it's the use case of AI and ML applied to data protection and backup and recovery. So those are very different conversations. Of course, they are intertwined to a large extent, but they're different. Protect the generation of new data is number one. And then secondly is generating better data protection with gen AI, with AI and ML, the latter to really reduce some of those mundane tasks. The former, people are talking a lot about synthetic data, the building digital twins, things of that nature. There's a real IP in there. You don't want to lose that. You certainly don't want that to get stolen.

>> All right, so this next data point is something that I'd like you to address. Guys, maybe bring up the AI data backup gap slide. Only about a third of respondents say they regularly back up more than half of their AI generated data. So Christophe, what does this tell you? Why is it important to protect data that AI generates?

Christophe Bertrand

>> Well, it's very simple. I like to follow the money. Is it worth money to you or not? And if it's not, then forget about it, don't worry about it. It's unlikely not to be important. Why? Because you've just spent a lot of money and great skill sets on generating new models and new data to train more models. So this data you're generating with AI is key to you. It's important to your business. You're doing this for a reason. Why not protect it? Now, it's typical when there's a new workload or a new process or a new use case. Sometimes it takes time for backup and recovery processes to catch up. So I think we're seeing this at play here, but with the speed at which AI is evolving, absolutely ridiculous. I think we're past the big high of last year. It was all buzzy around AI, but it was PR. Well, now people are looking at actual use cases. They're actually pushing those out. They're actually investing. So now they have to follow up with the backup piece. They have to protect those investments. I look at it this way, whether you're using AI for gen AI to supplement a business process or customer experience or whether it's for analytics to better understand your business, for example, build new products, it's actually pure IP and that is worth a lot of money to use. Data has become the business in many ways here.

>> Well, and images are becoming the thing. And actually, AI is getting pretty good at building images. And so that's going to drive a lot of data. And you want to be able to, for instance, I would presume, say, "Okay, this image was actually created without using external IP. It was done by our own IP." And so actually having a trail of what actually happened, an audit trail of whether it's the prompt or what the source of that data is going to be critical and it's discoverable.

Christophe Bertrand

>> Exactly. So you get into this question of what I would call data management on steroids now, because you have to make sure the data you use, first of all is usable. You can't use private data just like that. It has to be masked, it has to be anonymized. So there is a compliance and really a governance conversation leading to also this notion of do I have quality data that I'm using in my models? So this idea of quality of data is becoming very key. But here's the thing, actually, I have a lot of this data already in-house. Not just production, but in your backups, in your archives. So what can you use? What should you use? How do you reuse it? These are very important questions. And then as you produce results and new value out of these data assets, how do you protect those assets?

>> I wonder if it's got to be becoming a board-level discussion. It's got to be bubbling up just like cyber was. Data protection has become a fundamental component. We talked about this in December. It's no longer an adjacency. Now you've got this AI data protection, whole new set of processes. It's got to be bubbling to the top, don't you think?

Christophe Bertrand

>> Just like ransomware or cyber protection or resilience is an executive or board-level issue, the same thing is happening here. I think what we're seeing though, it's early stages. So there are many, many challenges. That means there's also an ecosystem that's forming. And in order to be successful, you really need to think through your processes, but also the investments you're making in technology, the type of vendors and partners you want to work with. And of course the skill sets you need to build to get to the point where you can safely generate these models if you want to use them in your business. You mentioned synthetic data. That's another topic. I think there'll be more in the future about that. At the end of the day, you are looking to make for quality in the data so that the models do not hallucinate, so that the data you generate is compliant, and so that you are responsible as an organization. So this notion of responsible AI is a very, very hot topic we're also researching in my practice.

>> Early days and a lot of unknowns. Now more than half of the organizations in your survey said that AI and ML will improve their ability to recover data after a ransomware attack. A large portion of the respondents also said it's going to help them improve their RPO and RTO posture. Guys, bring up the next graphic if you would. So better RPO RTO as it relates to recovering from ransomware attack. And only in the topic of gen AI, only 1% said gen AI will not help with improving the backup recovery process. So people are pretty positive about AI. What else did you learn from this data? And are you as optimistic?

Christophe Bertrand

>> I am as optimistic as they are. I think it's going to be absolutely key. I'd like to talk about this notion of generative data protection, really, where you can really interact and ask questions like, "Well, do I have the right recovery plans in place?" But tying this back to ransomware to can I actually automatically generate a plan in plain English, or at least in understandable terms that can be then executed by another part of the team. I'm talking about, again, this interaction of multiple teams. The automation, it's like the self-driving car. Do you really want to fully take your hands off the steering wheel, and do you need a steering wheel? Yes you do. It's not going to happen overnight. But there's plenty you can do to automate to do more. And the reason why you have to do this is because it's just too much data. Processes are very complex, and the nature of the people operating the environment has changed and they're not PhDs in backups or PhDs in ransomware. Those people don't really exist anymore if they ever did. So you have to have those tools. And I think gen AI and AI in general and ML in data protection is absolutely key. That's going to be, in my opinion, what makes a difference in the future between the vendors who win, the vendors who do not.

>> And it'll be an evolution. It's like full self-driving, right? Every day, we use automation capabilities when we drive. We're not full self-driving quite yet, but it's super helpful and I would imagine this is going to be a similar trajectory. All right, we've been talking all day about building cyber-resilient infrastructure. What's your take on Dell's announcements in that regard?

Christophe Bertrand

>> Also, first of all, when it comes to, we talked about AI too. I think they're clearly an established leader both in cyber resiliency with great solutions on this topic, and of course a key player and key partner when it comes to AI infrastructure, hardware and software, and really great solutions. So definitely not a big surprise to see that we're seeing this continued investment in the portfolio. Definitely great news for existing users of data domain. I think this is a couple of new platforms here that are going to change not only the performance characteristics of what they have in environment, but it's also about the reliability. I like also the fact that there's plenty of built-in security in the platform starting at the supply chain, really literally, supply chain of the actual system. So that's very powerful. And then there were other components to the announcement also around cloud and leveraging AI and cloud data protection. So I think when it comes down to cyber resilience, Dell has a very established reputation, thousands and thousands of customers to prove it. And this announcement only confirms their leadership. And I think I expect to see actually even more in the next few quarters because the problem's not going away. They're solving it really well. And there are plenty of customers who would love to be able to use easy to deploy solutions. .We

>> love having you as a guest because you always bring the perspective and the data. So thank you so much for spending some time with us.

Christophe Bertrand

>> Thank you for having me.

>> You're very welcome. And thank you for spending some time with us today to participate in Building Cyber Resilience on trusted Data Protection Infrastructure made possible by Dell Technologies and Intel. Remember, this episode and all its segments and the short takes are all available on demand at thecube.net with more coverage and written content on siliconangle.com. This is Dave Vellante, thanks for watching and we'll see you soon.

Christophe Bertrand | Enterprise Strategy Group

Christophe Bertrand

>> Great to be here.

Christophe Bertrand

>> So it's not just a technology case, it's not just about hardware failing. It's a whole house type of organizational business case that maybe IT people aren't used to making.

Christophe Bertrand

>> love having you as a guest because you always bring the perspective and the data. So thank you so much for spending some time with us.

Christophe Bertrand

>> Thank you for having me.