Data acquisition architecture studies for the KM3NeT deep sea neutrino telescope

KM3NeT is a European consortium whose goal is a future underwater neutrino telescope of cubic kilometer size in the Mediterranean Sea. The science case includes the study of high energy phenomena in the Universe involving the emission of neutrinos. The detection principle is based on an extended array of photomultipliers detecting single Cherenkov photons emitted by the charged products of neutrino interactions. This paper describes the conceptual design of a data acquisition and trigger architecture for the KM3NeT telescope. Its main features are based on the experience of the NEMO, NESTOR and ANTARES neutrino telescope pilot projects. The main issues addressed by this design include the integration of hundreds of acquisition nodes interconnected through a high bandwidth network and the seamless management of high rate data flows resulting from challenging levels of background noise. The networking technologies used -e.g. dense or coarse wavelength division multiplexing- address optimization issues such as minimizing the number of deep-sea fiber connections. The network topology is optimized for “all data to shore” transmission in which a real-time distributed data acquisition application manages a fluctuating data flow. The data are organized as time-slices and routed accordingly to a workstation farm running trigger algorithms which are expected to reduce the data flow by a factor of 10⁴. The control and configuration schemes that allow the proper operation of the neutrino telescope are specified together with their associated database organization principle. These principles address the issues of hardware description management, configurations and run conditions and their association with the acquired data. We will illustrate how the KM3NeT data acquisition system is intended to make the most of the available and affordable software and hardware technologies in a challenging data flow context involving embedded, real-time processing.