Machine learning applies adaptive algorithms to cluster patterns in your data and surfaces anomalies.
Today's Analytics Challenges
In the past, companies adopting machine learning have faced several hurdles to deliver real-time and predictive insights from large volumes of machine data (logs and metrics).
Multiple Data Types
Unstructured and semi-structured big data, often in the form of logs, is difficult to parse and organize without machine learning to uncover patterns, detect outliers and compare time periods.
Exponential Data Growth
Neither traditional analytic architectures nor humans were built to analyze the petabytes of machine data generated by today’s applications.
Inability to Adapt
Many tools do not accommodate learning from past behaviors and require manual adjustments to derive the accurate results.
Accelerate Modern Applications
With machine learning, you can parse and organize large volumes of unstructured and semi-structured machine data from complex, modern applications to:
- Prioritize application development efforts based on your users’ behavioral patterns, including seasonality and cyclicality.
- Rapidly surface operational issues based on multi-dimensional comparisons, pattern extraction and anomaly detection.
- Quickly identify threats that indicate elevated security risk and prioritize these threats without creating pre-defined policies or rules.
Machine Learning Techniques
Sumo Logic goes beyond simple search capability and min/max comparisons to deliver advanced analytics based upon:
- Clustering: Determines common patterns within unstructured data and surfaces the few exceptions in petabytes of data.
- Baselining: Identifies behavioral or statistical baselines in machine data and compares conditions across time periods to quickly identify issues as they occur in real time.
- Statistical Analysis: Determines statistical properties of numerical values to identify unexpected changes that may signal outliers.
- Regression Analysis: Overlaid on top of time series values to project where the values may go and predict future states.
Sumo Logic Platform for Machine Learning
Sumo Logic uses unsupervised, patented machine learning in LogReduce and LogCompare to help you organize unstructured data, identify patterns and detect outliers.
- LogReduce: LogReduce does the near impossible: put structure around unstructured data. It finds the inherent patterns in log file data, boils them down to the pattern that resembles the line of code (“printf” statement) and allows you to drill down into the most interesting data.
- LogCompare: LogCompare allows you to compare a section of your log message from one point in time with the same section at another point in time. It can help you gain insight into an application failure after a code change and even contrast what is occurring in a poorly performing server vs. its highly performing peer.
Sumo Logic’s cloud native, machine data analytics platform provides the machine learning, elasticity and scalability needed to surface insights in oceans of machine data.