New DarkShield Version Finds and Masks More PII Data at Risk
V5 Updates Engines, Insights and UIs while Protecting More Semi- and Unstructured Sources
January 10, 2024 — Melbourne, FL — Innovative Routines International (IRI), Inc., the 45-year pioneer in big data management and data-centric security software, has announced a major new version of its DarkShield software for searching and masking personally identifiable information (PII) and other sensitive data in structured, semi-structured, and unstructured sources on-premise and in the cloud. DarkShield V5 expands significantly on the features in V4 announced at the end of 2020, particularly via enhancements to the IRI Workbench GUI for DarkShield.
According to IRI software engineer Devon Kozenieski, “DarkShield 5 is the most significant upgrade in terms of capability, usability, and performance since the product’s inception. We paid particular attention to the user experience and the need to easily apply consistent masking rules.” Specifically in this release:
- DarkShield API functionality not previously front-ended in the IRI Workbench GUI now is, and Workbench uses V4 and new V5 (NoSQL and RDB) API engines for all search and masking jobs.
- New front-ended API options support: multiple cloud stores, advanced file configurations, new file types (including Parquet and DICOM images), and file filters and matchers to reduce search time.
- DarkShield source and target files can now be in one or more cloud stores: Azure Blob, GCP, OneDrive, and SharePoint Online (in addition to AWS S3 buckets which were already supported)
- Modernized data classification infrastructure in the IRI Workbench GUI for DarkShield supports data class groups and sensitivity labels, plus out of the box search matchers and masking rules
- Fuzzy matching to values in a dictionary (set) file can leverage Levenstein distance and other non-exact-match algorithms to broaden the reach of lookup-based value searches
- CLI execution of Workbench GUI search/mask configurations is supported, and another new Workbench wizard will export DarkShield job parameters to JSON contexts that can be sent directly to the DarkShield API for embedded use in external applications.
- Dark data file configuration and connection profiles have been added to allow their re-use
- Post-execution HTML5 graphs and “heat-mapped” sensitive data discovery and masking charts have been added to rank and review amounts and types of PII found in each silo and source, the names of each source, and those successfully (or unsuccessfully) masked, or just searched
- Built-in machine-learning for training OpenNLP NER models has been expanded to also cover PyTorch and TensorFlow NER model improvement, increasing the likelihood of finding PII or other sensitive data within their grammatical contexts (whether in English or other languages)
- Separate new Workbench views now also provide for: a preview of masking results on data found by the aforementioned NER models; a look at DarkShield API preferences and controls to manage the API server; aggregated audit log data; and, a Gantt chart of recently elapsed jobs.