Tejo: A Supervised Anomaly Detection Scheme for NewSQL Databases

Guthemberg Silvestre (1) , Carla Sauvanaud (1) , Mohamed Kaâniche (1) , Karama Kanoun (1)

1 LAAS-TSF - Équipe Tolérance aux fautes et Sûreté de Fonctionnement informatique

The increasing availability of streams of data and the need of auto-tuning applications have made big data mainstream. NewSQL databases have become increasingly important to ensure fast data processing for the emerging stream processing platforms. While many architectural improvements have been made on NewSQL databases to handle fast data processing, anomalous events on the underlying, complex cloud environments may undermine their performance. In this paper, we present Tejo, a supervised anomaly detection scheme for NewSQL databases. Unlike general-purpose anomaly detection for the cloud, Tejo characterizes anomalies in NewSQL database clusters based on Service Level Objective (SLO) metrics. Our experiments with VoltDB, a prominent NewSQL database, shed some light on the impact of anomalies on these databases and highlight the key design choices to enhance anomaly detection.

Mots clés: cloud computing; anomaly detection; dependability; machine learning; voltDB; NewSQL databases

Domaines: Performance et fiabilité [cs.PF]; Informatique [cs]

Guthemberg Silvestre, Carla Sauvanaud, Mohamed Kaâniche, Karama Kanoun. Tejo: A Supervised Anomaly Detection Scheme for NewSQL Databases. 7th International Workshop on Software Engineering for Resilient Systems (SERENE 2015), Sep 2015, Paris, France. ⟨10.1007/978-3-319-23129-7_9⟩. ⟨hal-01211772⟩