LinkedIn is a members-first company and we strive to maintain our members trust by enforcing security and privacy controls across our data systems. This is executed, in part, through procedures for controlling data access, deletion, anonymization, pseudonymization, and obfuscating Personally Identifiable Information (PII) across all of our data systems. We are continually improving how we protect member data through monitoring, auditing, alerting, reporting, and policy enforcement. Our best-in-class architecture is designed to capture, track, annotate, and aggregate metadata so we know exactly what is happening with the data that we store and analyze.
LinkedIn’s data ecosystem consists of online and offline distributed platforms. LinkedIn uses Kafka, Espresso, Hadoop, and MySQL to process data at a large scale. To ensure we are handling data responsibly, we utilize the Compliance Monitoring (CMON) Platform to scan all of our systems for issues. CMON is a fundamental platform to help data system, dataset, and application owners along with security, legal, and executives meet our commitment to our members.
The CMON platform leverages an end-to-end pipeline across our systems which detects violations and surfaces insights, metrics, and tickets. It also provides a unified view across metadata attached to people, schemas, data sets, owners, purge policies, access patterns/controls, and member PII and settings so that all system behaviors operate as expected.
Our approach is a unique effort to operationalize unprecedented visibility into how the LinkedIn ecosystem operates. Integration with the CMON platform is a mandatory component of all engineering efforts on the horizon which serves to improve the craftsmanship of our products. Since launch, we are proud to have surfaced and resolved issues that required attention, and as we improve our detection models we will be able to dry-run scenarios and forecast violations enabling teams to be well-informed and prevent issues.