Architecture

Components

Controller: Responsible to query the data and push it to the local Aurva database.

ML Engines:

• PII Analyser: To analyse textual information for semantic types (click here to know all the supported semantic types)
• OCR: To convert image and pdf files into textual information present in them to be analysed further by PII Analyser

Our scanning process is lightweight and designed to avoid imposing heavy loads on the database server. To ensure optimal performance and prevent any risk of overwhelming the database, we have made several key architectural decisions, outlined below.

Important Note: Aurva does not transmit actual business data to the management plane as shown in step 5 of the diagram. Only associated metadata is sent for reporting purposes — such as the name of the data asset, identified semantic types, table name, column name, and similar contextual information.

Infrastructure Pre-Requisites

CPU: 4 Cores
Memory: 16GB
Architecture: x86_64(AMD)
Operating System Support: Ubuntu

Please note that the CPU and Memory is the cumulative consumption by all the components.

If we proceed with the Kubernetes deployment, the total resource consumption of all the pods will be the same.

Permissions

AWS required Permissions

Architecture Discussion

Sampling

Structured databases