Centralised Computing Facilities
The Center CERIT-SC operates 11 PB of storage of two types: 1) Shared networked storage for user data – standard disk arrays to keep data used for computations on the clusters 2) High-capacity data storage (Hierarchical Storage Management, HSM with capacity above 9 PB) for storing semi-permanent and permanent data from scientific equipment and computations
The hardware resources are complemented by a wide portfolio of development tools (compilers, code analysers, debuggers), e.g., Intel and Portland C, C++ and Fortran compilers, Intel vTune, TotalView, Allinea DDT, and more than 200 modules of application software related to all scientific disciplines of the Centre’s users. In general, the software is purchased in tight cooperation with CESNET to maximise its impact.
The Center CERIT-SC operates more than 5,500 CPU cores in two main usage modes: 1) Cloud compute also known as "Infrastructure as a Service" paradigm – run virtual machines on demand with complete control, using standardised interfaces (industry standards OpenNebula and OpenStack as well as standardised OCCI interface). 2) High throughput compute use traditional computing environment – common operating system, hundreds of preinstalled software packages, payload distributed with batch system. Resources are allocated to individual users based on fair-share policies reflecting scientific results the users achieve using CERIT-SC.