February 2020 Summaries

This series of articles demonstrates a simplified example of near real-time Ad Campaign reporting using Kafka, Aerospike's edge-to-core data pipeline technology, and Apollo GraphQL. The solution captures ad events in an Aerospike "edge" database, aggregates them into Campaign Key Performance Indicators (KPIs), and displays the KPIs in a user interface using GraphQL to retrieve data stored in an Aerospike Cluster. The companion code is available on GitHub, with each back-end service written in Javascript and Node.js.

Microservices architecture is a modern approach to building large complex systems by delivering high level business functionality through loosely coupled microservices that are independently developed and deployed. Aerospike, a distributed key-value database, supports this architecture by enabling control of data, design of data, and coordination and consistency across services. It allows for separate namespaces or sets to store and control data, provides transactional guarantees for single record requests, and facilitates event based coordination between services. Aerospike's Cross Data Replication (XDR) also enables a multi-cluster event bus.

This series of articles discusses a simplified example of near real-time Ad Campaign reporting using Kafka, Aerospike's edge-to-core data pipeline technology, and Apollo GraphQL. Part 1 covered the real-time capture of Ad events via an Aerospike edge datastore and Kafka messaging. In this second part, aggregation and reduction of Ad events are leveraged using Aerospike Complex Data Types (CDTs) to create actionable Ad Campaign Key Performance Indicators (KPIs). The values are stored in a data cube implemented as a Document or CDT in Aerospike. Part 3 will describe how an Ad Campaign user interface displays those KPIs using GraphQL and retrieves data stored in an Aerospike Cluster.