Toward Semantic Enhancement of Monitoring Data Repository

Most existing monitoring systems do not provide semantic information for their monitoring data. Thus, monitoring data in the repository can only be understood by the system that has generated the data. Round Robin Database (RRD) is an important effort toward building monitoring data semantics. However, the RRD model has missing semantic information, such as what QoS metric or which system entity a monitoring data stream is for. Our goal is to enhance the semantics in RRD to facilitate semantic based monitoring data retrieval and to improve the interoperability of monitoring systems and analysis software. We leverage the RRD data semantics and develop the new SE-RRD (Semantic Enhanced RRD) model. We also develop the SE-RRDtool to provide semantic-based accesses to monitoring data. With SE-RRDtool, it is also possible to reason and automatically perform data alignment between the data requests and the monitoring data streams available in the repository. SE-RRDtool greatly enhances the capability of RRDtool and cloud monitoring systems can use it for semantic-rich monitoring data repository.