Testing the Suitability of Cassandra for Cloud Computing Environments Consistency, Availability and Partition Tolerance.
Felix Beyer, Arne Koschel, C. Schulz, Michael Schäfer, Irina Astrova, Stella Gatziu Grivas, Marc Schaaf, Alexander Reich · 2011
Abstract—Since relational database management systems (DBMSs) are ill-suited to cloud computing environments, multiple efforts are now underway to offer a viable alternative to relational DBMSs. These efforts have led to the rise of a new kind of DBMSs called NoSQL. One of the most visible products in this rise is Cassandra. Cassandra is a NoSQL DBMS, which can also be used as a clustered file system. Cassandra was claimed to be particularly well suited for cloud computing environments. Our goal in this paper was to confirm or deny that claim. Towards this goal, we conducted tests on Cassandra to determine what levels of consistency, availability and partition tolerance can be achieved and if these can be achieved without sacrificing performance. Keywords—Cloud computing, Cassandra, consistency, availability, partition tolerance, experiments.