The introduction of the Machine Data Lake represents Splunk’s answer to AI’s insatiable appetite for machine data. This purpose-built solution is designed to handle massive scale while maintaining the performance and accessibility required for AI workloads in Splunk. While at Cisco Live EMEA, I had an opportunity to chat with Greg Ainslie-Malik about Machine Data Lake and the thought process of how it complements Splunk in AI use cases. Stay tuned for more to come in this area…

Transcript

Auto-generated captions, lightly cleaned. Speakers are not separately labeled.

Hey everybody, we’re out here at Cisco Live in Amsterdam and I ran into Greg here and we’ve been talking about machine data lake and what it is, what it does. Can you give us some detail on the thinking there? >> Absolutely. So maybe not so much what it is or what it does, but we can talk a bit about the thought process behind why we need it, particularly when we look at AI. So if we think about things like the AI toolkit that we’ve had for many years, one of the challenges we have is that if you want to train a machine learning model or an AI model, you need huge volumes of data, lots of historic data, lots of context and Splunk is super optimized for the needle in the haystack searches. But that’s not so good at scooping up huge volumes of data. So part of the thinking behind the machine data lake is how can we have something that allows us to extract as much data as we need to train our models without hitting that search bottleneck that we have today. Like I say, super useful for needling in the haststack, not so useful for large scale model training. >> So that’s part of the thought process behind what we’re trying to do. >> So it just opens up those use cases for machine learning training models. What’s normal? What’s abnormal based on much larger data sets, much faster, more performant. Is that accurate? >> Exactly. So, if you think about the typical things we love to do at Splunk, anomaly detection is a big one. And understanding the baseline of normal behavior is absolutely critical to finding the unusual points. So, the more we know about what normal looks like, the better we can find the abnormal points. So we definitely need those huge volumes to get that baseline and get it set properly. >> Yeah, that’s some real interesting information and use cases. Now if I want to go learn some more, this is kind of emerging technology, some new stuff with Splunk. Where can I go find some more information? >> So right now I just keep an eye out for announcements. Uh this is something we’re actively working on. So as long as you love Splunk and Cisco, just keep an eye on the news from us. >> All right, awesome. Thanks so much. Appreciate your time. All right.