Idempotence, Immutability, and Concurrency: Backing up Big Data

Abstract

Traditional database, and storage, approaches have a well-defined, well-understood, "enterprise-ready" approach to backup and restore. NoSQL, as a general category, encompasses a wide variety of approaches for data storage, persistence, and database configuration. This has resulted in increased complexity when planning and implementing a backup and restore strategy. As DevOps practitioners, understanding application semantics ensure that our approach to backups aligns with the application, and business, needs.

In this session, Basho will discuss the core concepts/reasons why backup & restore strategies are considered, explain data persistence in the context of "dynamo-like" (i.e. Riak & Cassandra) solutions, and discuss ways to achieve the goals of backup and restoration as datasets continue to scale rapidly.