Deep Link Analysis at Scale

 

Use Case

Need for deep link analysis within Call Detail Records (CDR) to track and uncover the social interactions of people to discover new intelligence and how they interact within the community.

Problem: Relational databases or Big Data platforms may not be able to even load the data or return link analysis results when querying large scale, highly connected data.

 

Objectivity’s Non-Optimized Results Summary

Data Volume: 10.025 billion Call Detail Records

The CDR data included: Approx 5 million Persons, approx 20 million Phones, 10 Billion Calls

Ingest Speed: Distributing ingest with simultaneously massive parallel processing (MPP) and building the graph (connecting the data) in 4.5 hours.

  • 10.025 billion CDRs in 270 minutes (37 million CDRs per minute)
  • Linear performance observed from start to data ingest completion
  • Objectivity builds the graph upon ingest. This allows for:
    • Significant improvements in query performance
    • No need to separately index or connect the data since Objectivity performs this upon initial load
    • Allows users to use and query immediately upon load
  • Objectivity completed deep link analysis queries in seconds on the 10.025 billion CDRs. For comparison, the customer used relational databases and Big Data/open source tools and could not even load the 10.025 billion CDRs nor even return link analysis results on a smaller set of 1 billion CDRs.
  • Note: Initial results were performed in a standard cloud environment. These are preliminary results only. Performance can be optimized by scaling up and out.

 

Why Objectivity?

Objectivity is the only solution that combines the strengths of RDBMS, Big Data platforms, and graph querying capabilities into a single platform for performance at massive speed and scale.

  • Objectivity can do deep link analysis (graph), complex queries at scale (beyond billions of connections). These very deep navigational queries executing across datasets in excess of 10 billion nodes are some of the most difficult queries for any database platform to perform.
  • Objectivity can also perform relational data queries at speed and scale
  • Objectivity is proven to handle the scale required of operational systems with being able to perform queries across datasets in excess of 10 billion nodes