- Network (DNS and DHCP)
- Central user management :
- Directory Service (Kerberos + LDAP)
- Integration with TUM/LRZ Active Directory Service
- Home directories (NFS)
- Network Attached Storrage (NAS)
- E18 Linux
- Web server
- Mail server
- Database server
- Login server
- High performance computing: batch farm (cluster)
Two racks, located in the central server room of the Physics department and dedicated to E18, provide the space for the central computing infrastructure of the chair. These racks are provided with a total cooling power of 30kW. Emergency power combined with UPS and redundant hardware grand high availability capacity.
Servers and several dozen workstations at the chair run a customized variant of linux, based on CERN CentOS 7. It includes all central services like user management, including single sign on, storage and cluster access. Besides that it provides a huge set of preinstalled packages of standard and self maintained packages, which add up to a powerful, homogeneous development environment for users.
To satisfy the high computing demands that are typical for modern particle physics experiments, substantial computing infrastructure has been built up at the chair. This includes a small, but never the less powerful Linux cluster, which is run as a batch farm.
GridEngine servers as cluster scheduler for the batch farm, which runs E18 linux as operation system.
Hardware wise the compute cluster consists of 37 compute nodes, providing a total computing power of ~20TFlops/s on 944 cores. In total, the compute cluster is equiped with ~4.3 TB of RAM.
Network attached storage is provided using the NFSv4 and SMB protocols with Kerberos authentication. The total amount of ~600 TiB (netto) redundant storage is provided by four servers.