The CMS experiment at the CERN LHC collider is producing large amounts of simulated data in order to provide an adequate statistic for the Trigger System design.These productions are performed in a distributed environ...The CMS experiment at the CERN LHC collider is producing large amounts of simulated data in order to provide an adequate statistic for the Trigger System design.These productions are performed in a distributed environment,prototyping the hierarchical model of LHC computing centers developed by MONARC.A GRID approach is being used for interconnecting the Regional Centers.The main issues which are currently addressed are:automatic submission of data production requests to available productioin sites,data transfer among production sites,“best-replica” location and submission of enduser analysis job to the appropriate Regional Center,In each production site different hardware configurations are being tested and exploited.Furthermore robust job submission systems.which are also able to provide the needed bookkeeping of the produced data are being developed.BOSS(Batch Object Submission System)is an interface to the local computing center scheduling system that has been developed in order to allow recording in a relational database of information produced by the jobe running on the batch facilities A summary of the current activites and a plan for the use of DataGrid PM9 tools are presented.展开更多
The data acquisition system for the CMS experiment at the Large Hadron Collider (LHC) will require a large and high performance event building network.Several architectures and swithch technologies are currently being...The data acquisition system for the CMS experiment at the Large Hadron Collider (LHC) will require a large and high performance event building network.Several architectures and swithch technologies are currently being evaluated.This paper describes demonstrators which have been set up to study a small-scale event builder based on PCs emulating high performance sources and sinks connected via Ethernet or Myrinet switches.Results from ongoing studies,including measurements on throughput and scaling,are presented.展开更多
文摘The CMS experiment at the CERN LHC collider is producing large amounts of simulated data in order to provide an adequate statistic for the Trigger System design.These productions are performed in a distributed environment,prototyping the hierarchical model of LHC computing centers developed by MONARC.A GRID approach is being used for interconnecting the Regional Centers.The main issues which are currently addressed are:automatic submission of data production requests to available productioin sites,data transfer among production sites,“best-replica” location and submission of enduser analysis job to the appropriate Regional Center,In each production site different hardware configurations are being tested and exploited.Furthermore robust job submission systems.which are also able to provide the needed bookkeeping of the produced data are being developed.BOSS(Batch Object Submission System)is an interface to the local computing center scheduling system that has been developed in order to allow recording in a relational database of information produced by the jobe running on the batch facilities A summary of the current activites and a plan for the use of DataGrid PM9 tools are presented.
文摘The data acquisition system for the CMS experiment at the Large Hadron Collider (LHC) will require a large and high performance event building network.Several architectures and swithch technologies are currently being evaluated.This paper describes demonstrators which have been set up to study a small-scale event builder based on PCs emulating high performance sources and sinks connected via Ethernet or Myrinet switches.Results from ongoing studies,including measurements on throughput and scaling,are presented.
