We provide services in evaluating and selecting appropriate big data store that is suitable to manage our client’s data variety, data volume and data velocity.
IBM defines big data using three characteristics: volume, variety, and velocity (Zikopoulos and Eaton, 2011). Accordingly, big data can be defined as the data -with characteristics such as volume, velocity or variety- that cannot be effectively managed using traditional Relational Data Base Management Systems (RDBMS) and that often requires horizontal scalability for efficient processing.
Following table compares major big data store types in their performance of storing and accessing big data, in their flexibility of handling data variety, in their flexibility of handling data types such as nested data, in their horizontal scalability, in their complexity of operations, in their applications and in their query ability:
Finally, following visual guide uses Brewer’s CAP Theorem to group different big data store options available today:
Given different characteristics of available big data store options, it is very important at the big data store selection stage to understand the needs of the application that will use these big data stores. For instance, if the application needs that client always have the same view of the data and that query ability of the data store is similar to traditional RDBMS, document-oriented stores such as MongoDB could be the suitable big data store option. Moreover, big data store options can be used in combination or with traditional RDBMS to satisfy different needs of the application. For instance, archived data can be stored in big data stores while using traditional RDBMS to store recent transactions.
Contact us for more information. See below for more details.