BigPanda’s Open Box Machine Learning
Machine learning you can trust
The modern IT environment is noisy, complex, and fast-moving – often overwhelming IT Ops, NOC and DevOps teams that need to handle hundreds of thousands of alerts a day, detect incidents in real-time and surface their root cause by understanding the dependencies between the different elements in their hybrid cloud IT stack.
Artificial intelligence and machine learning can help as they are exceptionally good at processing enormous volumes of complex data in real-time and surfacing actionable insights. But many machine learning tools suffer from a “Black Box” approach, where teams cannot see or control the machine learning logic, leading to trust, adoption, and reliability issues.
BigPanda’s machine learning logic you can see, understand and control
Open Box Machine Learning provides unparalleled event correlation and root cause analysis capabilities due to a unique open approach.
With BigPanda Open Box Machine Learning, the machine learning logic is explained to IT Operations teams in plain English. Teams can then edit this logic to add situational and tribal knowledge to strengthen it on their own without requiring expert data scientists. From there, teams can test it and run what-if experiments on real live production data to make sure their changes work as intended, before deploying them – promoting higher trust and adoption of the machine learning throughout the organization. This leads to exceptional event correlation results and rapid root cause detection, reducing operational risk, and increasing service performance and availability.
Main features of BigPanda’s Open Box Machine Learning
Automatic event correlation leveraging machine learning
BigPanda’s Open Box Machine Learning engine automatically correlates related alerts into high-level incidents using time, topology, context and alert types.
BigPanda ingests the raw event data from monitoring, topology and change systems through its Open Integration Hub. This data is first normalized into standard tags and then enriched with configuration information, operational categories and other metadata.
The normalized, enriched data is then correlated by BigPanda’s Open Box Machine Learning, which merges the events into alerts and clusters the alerts into high-level, actionable incidents by evaluating their properties against patterns in four dimensions: time, topology (i.e., datacenter, rack, cluster), context (i.e., criticality, team, customer impact) and alert types (i.e., network, storage, application). As new alerts are received, the Open Box Machine Learning evaluates all matching patterns, and determines whether to update an existing incident, or to create a new one.
With Open Box Machine Learning, BigPanda can effectively and accurately correlate alerts to dramatically reduce monitoring noise by over 95%, in real-time.
Explainable root cause determination
BigPanda’s Open Box Machine Learning provides advanced, automated root cause analysis capabilities that explain their logic to the user in plain language.
Infrastructure and application related root cause: When BigPanda’s Open Box Machine Learning engine correlates related alerts into high-level incidents, it verbalizes the relationship between these alerts in the title of the incident, to help users identify the probable root cause. As the correlation evolves and the logic becomes more precise, the title is updated in real-time, and ensures operators retain ongoing situational awareness.
Change related root cause: BigPanda aggregates change data from different change feeds and tools, including continuous integration/continuous deployment, change management and auditing. It then uses Open Box Machine Learning to analyze the change data against incident data and identifies which of these changes likely caused the incident by matching patterns between the incident and the change – based on time, domain, topology, geography, context and more.
Open Box Machine Learning control panel for pattern editing and testing
BigPanda’s Machine Learning Engine generates correlation patterns automatically based on historical user data. Upon the integration of a monitoring tool BigPanda’s Open Box Machine Learning engine starts to review the data being collected. Over time, new auto-generated patterns are created and displayed in the Open Box Machine Learning control panel as a suggestion.
Once the Machine Learning Engine suggests a pattern, administrators can decide to activate it, reject it, or further customize it within the control panel using the pattern editor. Teams can customize the correlation pattern definitions based on their environment, to increase the effectiveness of the correlation, without the need for expert data scientists.
The Control Panel also provides a pre-production tester, so teams can run the edited logic on real live data without having to go into production.
BigPanda’s Open Box Machine Learning provides significant advantages to IT Operations teams
Advanced proprietary machine learning models
Based on advanced techniques such as word embedding, clustering and convolutional neural networks, BigPanda’s mix of proprietary supervised and unsupervised machine learning algorithms were developed by world-class leading experts, and leverage the company’s extensive experience with some of the world’s biggest and most complex IT environments.
Transparency – no data scientists needed
BigPanda’s Open Box Machine Learning technology creates unparalleled transparency into its machine learning logic by displaying the different parameters associated with each of its machine learning-generated correlation patterns in plain English inside the BigPanda administration console, so that users can easily understand how it works, even if they have no expertise in AI or Machine Learning. BigPanda’s Open Box Machine Learning models are designed to be trained by incoming data automatically, over time, without intervention from a data scientist. This unique explainability helps users trust the results generated by BigPanda, driving the adoption of artificial intelligence operations and a rapid return on investment.
Full control and testability
BigPanda’s Open Box Machine Learning technology offers unprecedented control by letting users incorporate their tribal knowledge and business context into its machine learning logic. Users can edit BigPanda’s machine learning-generated correlation patterns with just two or three clicks, and then test the pattern against data collected in different windows of time. Once satisfied with the results of their edits, they can deploy the new or edited pattern to production.
Unrivaled time to value
BigPanda Open Box Machine Learning creates correlation and automation logic that goes to work out of the box at a 70% correlation rate. Over the next 6-8 weeks, BigPanda’s machine learning-driven logic achieves a 95% correlation rate. New logic is created on an ongoing basis to continually increase operational efficiencies.
Adapts to evolving needs
BigPanda’s machine learning models continuously and automatically learn and adapt to ever-changing IT environments, so they are always in sync with evolving business and technical needs. BigPanda’s Open Box Machine Learning regularly suggests new correlation logic as new infrastructure elements, apps, services and tools are added to environments and new data is ingested and correlated. This makes BigPanda the perfect match for organizations that are modernizing their applications.