We visited with Dan Maloney, CEO of Perspica, Inc., about the emergence of “bi-modal IT,” which calls for a reorganization of IT in order to meet the burgeoning digital requirements that enterprise IT is facing.
Maloney provided insight into the greater need today than ever before for visibility into an organization’s application stack and how new approaches must be employed to provide intelligence across hybrid-cloud environments to keep applications running at desired service levels. This new reality led Perspica to create Incident Replay.
Perspica’s Incident Replay is a subscription-based SaaS solution that delivers application infrastructure intelligence across virtual, physical and hybrid-cloud environments. It is an automated, self-learning solution that analyzes and provides insight into ever-changing applications and network topologies.
ADM: This idea of bi-modal IT is gaining steam. Walk us through what that looks like and why it’s a challenge.
Maloney: Gartner coined the phrase, defining bi-modal IT as “the practice of managing two separate, coherent modes of IT delivery, one focused on stability and the other on agility.” This includes TechOps and DevOps, respectively. Gartner predicts that 75 percent of IT organizations will be bi-modal in some way by 2017. The complexity of this new model means that both IT leaders and application owners need to understand what’s happening throughout the entire application infrastructure stack at any given time.
ADM: It seems that visibility is a key issue then, right?
Maloney: Absolutely. Applications are running on complex and dynamic infrastructures, with underlying resources constantly changing to meet these applications’ performance requirements. Our Incident Replay is a time machine for performance data, logs and topology, and it lets you visualize all layers of your application stack in one place. It’s like an instant replay when you’re watching sports.
Incident Replay enables you to identify the root cause of an outage or performance degradation at any point in time and improve mean time to repair. Without these capabilities, conducting definitive post-mortem analysis is an expensive, manual process – if it can be done at all.
ADM: Aren’t security tools and all those carefully set up policies supposed to send out alerts if there’s a performance issue?
Maloney: Today’s IT operations tools trigger numerous alarms, but they have no inherent capability to distinguish between the critical, service-impacting events and false positives. So, the problem is that people get “alarm fatigue.”
Perspica’s machine learning engine analyzes performance metrics collected from various sources including logs, availability and configurations to create a behavior profile for all the components of your infrastructure. Through this analysis, Perspica lets you know exactly what needs your attention by providing you with a solution that includes smarter alarms, automated root cause analysis and actionable recommendations on how to fix the problem.
ADM: We hear a lot about machine learning these days. What does it look like in your solution?
Maloney: That’s a great question! Today’s complex IT environments generate up to millions of events and metrics every second. Perspica ingests and understands this high volume, velocity and variety of big data and analyzes it in real time by running five concurrent machine learning engines.
We call them our Fab Five Analytics: topology, behavior, anomaly, problem and predictive. We’re able to cross-correlate your infrastructure’s behavior and create a baseline of normal behavior. If the application or infrastructure is out of the norm, even before an outage has occurred, we will notify you that something needs your attention.
This level of machine learning makes Perspica the only solution that provides past, present and predictive analytics. The key to preventing outages is to predict issues before they become problems, but traditional monitoring tools trigger alerts only after a problem has already occurred. Equipping users with the ability to replay and analyze past incidents, along with state-of-the-art behavior profiling and anomaly detection, Perspica improves current performance while predicting and preventing future costly outages and downtime.