The open-source community now has access to Adobe’s one-stop-shop project for data processing needs. Adobe’s One-Stop Anomaly Shop (OSAS) is now available on GitHub. It was created to detect abnormalities in datasets with ease as well as improve the processing and format of security log data.
Adobe says that OSAS combines the security research done by the vendor and other open-source projects.
What it offers is an out-of-the-box system that can be used in processing, dataset experimentation and allows developers to find out ways that can shorten the path to find the best solution for detecting security threats.
Inside the project
The release leverages Hubble, a modular open-source security compliance framework. The framework provides on-demand profile-based auditing, as well as real-time security incident notification, alerts, and reports.
Security logs tend to be complicated and may not fit well with machine learning-based tools. It can create data sparsity and leads to problems that arise as a result of trying to make unstructured data structured (usable sets).
The command-line interface (CLI) toolset uses two processes on datasets to try and figure them out, according to the team behind the release.
More to come
The CLI toolset can be used to tag raw data with field types like multinomial, text, and numeric values, as well as label content based on set rules. With capabilities like that, it is easier to make sense of security logs.
In the second stage, the labels are used as input features for unsupervised (generic) or supervised (targeted) machine learning algorithms. There are only three standard options at present but the future will bring more.
Adobe also released the OSAS code in a full and has included a Docker version.