Dynamic/Flexible Schema
The Challenge
Soon, most Enterprise data will not be transactional, and will not be stored in a traditional relational database. It will be unstructured user data – Web click-streams, Tweets, blog posts, Facebook updates, picture uploads, documents, customer product reviews, metadata, health records, audio, video as well as machine generated data.
The number one problem facing enterprises today is how to store and structure all this data and leverage it, not simply where to put it. One of the defining characteristics of non-transactional data is that it typically has no structure. It does not fit into your defined schemas. Enterprises typically require a database where your schema needs to be defined before you start to capture and leverage data. What if you need to capture unstructured data? You’ll need a system where the schema can change as your data types do. Cassandra is one such database.
The Solution
Data changes over time. As originally described in the Google BigTable paper, Cassandra offers the organization of a traditional RDBMS table layout combined with the flexibility and power of no stringent structure requirements. This allows you to store your data as you need to without performance penalty for changes as your needs evolve. These changes are transparent to you and the Cassandra users. The Cassandra schema-less data model is much easier to adapt to rapidly changing requirements than a highly structured, rigidly enforced schema.
The Cassandra data model is designed for distributed data on a very large scale. It enjoys advantages in performance, availability, and operational manageability. Some newcomers to Cassandra approach the data model with analogies to relational models – for instance, equating column families with tables, keyspaces with databases, and so on. In Cassandra, denormalization is the norm. A standard and very efficient way of working with the Cassandra data model is to create one column family for each expected type of query. With this approach, data is denormalized and structured so that one or multiple rows in a single column family are used to answer each query.
