Scalable Archiving with the Cassandra Archiver for CSS
S. Marsching · 2013
An archive for process-variable values is an important part of most supervisory control and data acquisition (SCADA) systems, because it allows operators to investigate past events, thus helping in identifying and resolving problems in the operation of the supervised facility. For large facilities like particle accelerators, there can be more than one hundred thousand process variables that have to be archived. When these process variables change at a rate of one Hertz or more, a single computer system can typically not handle the data processing and storage. The Cassandra Archiver has been developed in order to provide a simple to use, scalable data-archiving solution. It seamlessly plugs into Control System Studio (CSS), providing quick and simple access to all archived process variables. An Apache Cassandra database is used for storing the data, automatically distributing it over many nodes and providing high-availability features. This contribution depicts the architecture of the Cassandra Archiver and presents performance benchmarks outlining the scalability and comparing it to traditional archiving solutions based on relational databases.