This last Tuesday EURA NOVA hosted an interesting Cassandra meetup. Several tech enthusiasts and Cassandra fans have met each other for two talks: I briefly introduced what are the challenges and potential solution when building a distributed graph layer on top of Cassandra, the second talk, presented by Ansar Rafique, was a general introduction to Cassandra and some performance comparisons against MySQL. Ansar is a PhD student at the department of Computer Science of the Katholieke Universiteit Leuven – Language technology and middleware taskforce (LANTAM) of the DistriNet research group. The talk presented the result of Ansar’s previous researches at Cinnober Financial Technology about factors influencing Read/Write operation in MySQL and Cassandra.
The performance evaluation of distributed storage solutions has always been a focus of interest at EURA NOVA. Back in 2009 one of our first technical report was about comparing Oracle RAC Share everything and MySQL Share nothing approaches [1]. In 2011 two Master thesis co-directed by EURA NOVA and the Faculty of Computer Engineering of the UCL [2,3] and the ULB [4] have studied the elastic qualification of distributed storages. Indeed, even if a solution offers an elastic feature (ability to add/removes nodes at run-time), the impact on the performance during the change in the topology could be quite different. Finally, one of the master thesis led to the publication of [5] in Cloud Computing 2011 in which we proposed an elastic model definition.
Interestingly, under the NoSQL movement pressure, very recent researches in the RDBMS side of the world, focused on the elastic evolution of relational storage. The two major evolutions are (1) the not so recent column-base storage model and (2) the implementation of the Data Fission, as described by El-Abadi [6] as the ability to support a full elastic behavior. We can cite Elastras [7], CloudSQL server from Microsoft Research or yet relationalCloud [8] from the MIT.
Sabri Skhiri
Twitter: @sskhiri
References
[1] Nam-Luc Tran, MySQL Cluster Design and Architecture Principles, EURA NOVA_whitepaper. October 2009.
[2] Thibault Dory, Comparing Scalable NOSQL Databases, Master thesis UCL in collaboration with EURA NOVA. February 2011.
[3] Thibault Dory, Study and Comparison of Elastic Cloud Databases, Master thesis UCL in collaboration with EURA NOVA. April 2011.
[4] Nicolas Degroodt, L’élasticité des bases de données sur le cloud, Master thesis ULB in collaboration with EURA NOVA. June 2011.
[5] Thibault Dory, Boris Mejías, Peter Van Roy, Nam-Luc Tran, Measuring Elasticity for Cloud Databases, in the proceedings of the The Second International Conference on Cloud Computing, GRIDs, and Virtualization. September 2011.
[6] D. Agrawal and al., Database Scalability, Elasticity, and Autonomy in the Cloud, in the Proceedings of the 16th international conference on Database systems for advanced applications – Volume Part I. 2011.
[7] S. Das and al., ElasTraS: an elastic transactional data store in the cloud, in the proceedings of The conference on Hot topics in cloud computing, HotCloud 2009.
[8] C. Curino and al., Relational Cloud: A Database Service for the Cloud, 5th Biennal Conference on Innovative Data Systems Research. January 2011.