Skip to Main Content
The ATLAS event monitoring service (Emon) is a CORBA-based peer-to-peer data distribution framework, which has been implemented for ATLAS - one of CERN's forthcoming particle physics experiments. Its purpose is to sample collision event data from various stages of the data flow chain and efficiently distribute them to the users. This paper discusses various technical details of the current peer-to-peer-based implementation. A special focus lies on dependability and the provision of scalability and predictable performance. A comparative study of message complexity and scalability as well as practical test results obtained during the HLT/DAQ large scale and performance test periods in July 2005 and November 2006 will be presented. Apart from that, a recent addition that facilitates the parallel analysis of event data on a set of machines will be presented.