Veeam exec illustrates how AI fits into backup, recovery

Generative AI might have all the buzz, but traditional AI also has the potential to make backup and recovery simpler and more secure for storage admins now.

While it's still early days for generative AI in backup and recovery, vendors have started adding features using the technology. In early December, Veeam became the latest data backup vendor to add an AI-powered assistant to its platform.

The vendor's update to the Veeam Data Platform -- which includes version 12.1 of the vendor's flagship Backup & Replication software -- also added more traditional AI to power its malware detection engine. Other updates to the platform include the addition of object storage backup, the Veeam App for ServiceNow integration, YARA content analysis that helps prevent reinfection with malware, and a new Threat Center that identifies risks and measures a security score.

CTO Danny Allan said Veeam has been incorporating AI into its products for several years now -- before the rise of generative AI. He pointed to Veeam Intelligent Diagnostics as an example.

"We could proactively predict when a customer was about to have an issue," Allan said.

In this Q&A, Allan details Veeam's new AI features, what works and what doesn't with AI in backup and recovery, and a look ahead to focus areas in 2024.

Editor's note: This Q&A has been edited for clarity and length.

How does the AI-powered malware detection engine work?

Danny Allan

Danny Allan: We've been doing security analysis for a while. But it was always done immediately after a backup was complete, or on demand on some periodic basis, or on recovery. The challenge with that approach is that you're looking at the full data set every time.

So we added the malware detection to the proxies, as the data was flowing from the source to the target. That gave us two benefits. First, it's only on the incremental data. You're only looking at the changed blocks. It's far, far faster to do the analysis if you're doing it on the proxy.

The second benefit is we know immediately -- we don't have to wait until the backup is complete. As soon as we see evidence of malware, we're aware of the fact that you've been hit by ransomware. So we don't wait until it completes.

This isn't generative AI, correct?

Allan: It's not generative AI. It is model-based AI or foundational AI. We're feeding in the blocks and we're tracking that over time to determine when things are changing in an anomalous way.

How does the AI-powered assistant work?

Allan: That is true generative AI. We fed three different sources of data into that. We fed in our knowledge base, we fed in our help center, and we fed in portions of our website. If you ask it questions, it will come back to you based on those models to say the answer to that question. It takes best practices and all the feeds from the systems and will answer any queries that the end user may have.

Technology for technology's sake is the wrong focus. If you go back three years, people would say, 'What are you doing with blockchain?' And then it was, 'What are you doing with Web3?' And now it's, 'What are you doing for AI?'

AI in general is making a lot of noise in the industry and in the market. How does Veeam cut through that noise and make an impression on customers with what you're doing with AI?

Allan: Technology for technology's sake is the wrong focus. If you go back three years, people would say, 'What are you doing with blockchain?' And then it was, 'What are you doing with Web3?' And now it's, 'What are you doing for AI?'

It should be, 'Are you making the product better?' The answer is, yes, we're using AI to make the product better.

We're giving you an assistant that helps you perform tasks quicker. We're helping the customer understand that they're about to have a problem before they have that problem.

It's more about outcomes than it is about technology for technology's sake.

What are the best uses for AI in backup and recovery?

Allan: The most dominant use of AI will be around detection within the context of ransomware.

Secondly, it's around assisting users to do things in a simpler way. Giving customers better intelligence about their environment is probably the second-most important way.

The third way is around forecasting. And that could be cost forecasting, which we do especially in our cloud products, but also in infrastructure forecasts. We use it now, for example, around capacity planning -- we can say, 'You're about to run out of storage two months from now, you need to look at it.'

What are some focus areas Veeam will be eyeing for the next year?

Allan: Security is going to continue to play a very big part. Ransomware is not going away; it's only going to get worse. I expect [ransom] payments to go up. Perhaps the number of payments will remain consistent, but I expect the amount that attackers are demanding will increase. That's going to cause us to look far more at zero-trust architectures and integrations with partners. Those partners will span the breadth of everything from key management systems to reporting systems to networking systems.

I expect to see growth in cloud. One of the drivers of that is the disruption in the industry with Broadcom acquiring VMware. I don't expect to see a mass migration overnight, but I've talked to a lot of customers that say now is a good time to evaluate what is happening in the cloud ecosystem. I expect hyperscale clouds to start getting more workloads than they had in the past.

Paul Crocetti is an executive editor at TechTarget Editorial. Since 2015, he has worked on TechTarget's Storage, Data Backup and Disaster Recovery sites.