Online Performance Modeling for NoSQL Databases using Extreme Learning Machines

Victor A. E. FariasPedro R. A. PinheiroFlávio R. C. SousaJoao P. P. GomesJavam C. Machado

NoSQL databases rise as a solution to manage large amounts of datain the cloud. Mechanisms to guarantee Quality of Service in can significantlybenefit from performance predictability. Building an accurate predictive modelto estimate a DBMS performance in a cloud environment is challenging sincei) workload and resources allocation change dynamically; ii) concurrency anddistribution introduce nonlinearity on performance metrics and iii) predictivemodels should be trained and updated online to capture unseen workloads. Thispaper presents an online performance modeling approach for NoSQL databasesusing extreme learning machines. Experimental results confirm that our performance modeling can accurately predict throughput under several scenarios.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: