NEMO: A network monitoring framework for high-performance computing

Abstract

The volume of data generated by the Large Hadron Collider (LHC), several PetaBytes (PB) per year, requires a distributed tier-organised structure of computing resources for mass storage and analysis. The complexity and diversity of the components of this structure (hardware, software and networks) require a control mechanism to guarantee high-throughput high-reliability computing services. NEMO is a monitoring framework that has been developed in one of the computing clusters that receive data from LHC and has been designed to measure and publish the state of a cluster resources, maximize performance and efficiency and guarantee the integrity of the cluster.