What AIOps Is—and Why Riverbed Is an AIOps Pioneer
The term AIOps encompasses the application of big data, machine learning, analytics, and automation to IT operations use cases in order to make sense of large quantities of mostly structured, specialized, cross-domain IT data. Most of those use cases would be very familiar to IT operations staff; as David Winikoff, Senior Director Product Management at Riverbed, puts it, “It’s about doing something that organizations have to deal with every day, but better. The AI part means we aim to tackle these problems with less effort—faster, easier, with less specialized skills required, and with less time expended.”
AIOps has become something of a trendy concept. But the Riverbed SteelCentral suite of products delivers on its promise—and has for years. “Riverbed’s toolkit has pieces that predate people knowing the terms ‘machine learning’ and ‘artificial intelligence,'” says Winikoff. “The buzzwords didn’t even exist when we started.”
A robust AIOps rollout consists of four dimensions:
- You’ll need data about the systems you’re monitoring—complete, granular data from multiple sources.
- You need to have knowledge of the context in which that data was collected. For example, while observing network packets, it’s possible to infer network delays for both sides of a conversation, but to do so requires context and understanding about the location of the observation point—is it closer to the user or the server?
- You need to be able to surface insights through visualizations and machine learning. These insights might be relationships between events that aren’t immediately obvious, like “VoIP calls drop when interface IF_02 is at 90% utilization,” or relevant contextual information, like “macOS users in Cambridge are experiencing slow Office 365 performance.”
- The systems in use by your customers might be able to use these insights to automatically remediate certain types of problems.
We can use these dimensions to see how Riverbed fits into the AIOps landscape.
Data and its context
Let’s start with data. Riverbed’s SteelCentral product suite has deep hooks into your customer’s infrastructure and collects the best, richest set of cross-domain data that exists in the industry. This is our core strength. With data from AppInternals, Aternity, SD-WAN, and AppResponse, Riverbed can deliver AIOps for networking, application, and operations teams.
But Riverbed’s tools don’t just provide data: they provide the context necessary to extract real meaning from it. This sets Riverbed apart in the field. “A lot of companies describe their machine learning as data agnostic,” says Winikoff, “which essentially means that you feed us a bunch of ones and zeroes and even if we don’t know what any of those numbers mean, we’ll tell you what correlates to what.”
Riverbed’s take is different. “We know that the data-agnostic approach is sub-optimal,” says Winikoff. “Our data, by contrast, has context, and that’s how we know that some data is more valuable—and more predictive—than other data. We’re not just doing an analysis of this big mass of stuff; we know which data to look at first, which data to look at when different types of problems or issues occur.”
Again, the breadth of Riverbed tools across various IT silos, from networking to applications, is important here. “We’ve got a portfolio of all of these different data sources, and we fundamentally know the meaning of not just what we’re measuring but how it was measured,” says Winikoff. “We have the domain knowledge to understand the interpretation of the data so that it’s not just numbers to us.”
With access to both the underlying data and the context in which that data was generated, Riverbed can offer unparalleled visibility into your customer’s infrastructure’s health. Riverbed’s data scientists experiment with dozens of algorithms on many combinations of our data, searching for approaches that yield worthwhile insights that help you resolve digital experience issues faster.
“What we love to find is the needle in the haystack,” says Winikoff. “We aim to find the early symptoms of a problem before it actually hits you—like being able to know from a weather report that a serious storm is coming even before the rain has started to fall. If things are starting to slow down, you want to be able to spot that and fix it even before your customer gets impacted.”
There are a number of use cases that can benefit from the insights Riverbed can deliver:
- Troubleshooting, which Riverbed’s insights makes easier and faster, even for those members of your team lacking guru-level skills
- Threat hunting, seeking out security and operational threats
- Using graphs and visualizations to better see connections and context, and cause and effect
This also reduces finger-pointing and can help to define a clear line where responsibilities are divided, in a way that is not the only objective, but also helpful, since you will be able to provide enough information to help other groups troubleshoot faster.
A healing touch
One of the newest and most exciting aspects of AIOps is automated remediation—when systems can “self-heal” with minimal or no human intervention. We talked earlier to Riverbed Product Evangelist Mike Marks about how this can work. With Riverbed, IT and our partners can save time by automating repetitive tasks. For instance, you can build a library of remediation actions for end-user device-based problems that can be executed either automatically or manually in response to common problems like hard drive failures, app or system crashes, low disk space, and the like. This can help you address reoccurring problems faster and refocus your resources to more complex issues that require their attention, increasing your efficiency, while at the same time offering a better service to your customers.
Our product/platform vision for the future
Riverbed has been building algorithmic approaches and domain-specific analytics into our products for many years. And techniques we’ve built over all that time are now coming together as we move into the future of AIOps.
What’s in store in the near term? “You’re going to be seeing the kind of analysis that we do today, but we’ll bring together the different insights we’re getting from each of our products into a common place,” says Winikoff. “Each of the crucial internal or outsourced audiences—the apps team, the ops team, and the networking team—will be able to better do their work based on that common analysis of the customer’s environment.”
For a deeper view of AIOps and how it’s going to change IT, check out the Riverbed white paper “What is AIOps? Four Dimensions of Artificial Intelligence for IT Operations.” To learn more about how Riverbed’s product offerings apply AIOps principles, read the Riverbed solution brief “Solutions for AIOps.”