Archer is based on Cray XC30 high performance compute nodes. Each node has dual Intel Ivy Bridge processors and uses the Lustre open source file system that allows processing jobs to be run across multiple XC30 instances in parallel. User data first comes into a pre-processing area where it is checked and prepared for running on the on the main Archer systems. The XC30 compute nodes have direct-attached disk – to capacity of around 5 Petabytes in total – to allow rapid I/O (input-output) during modelling operations with 100GBsec peak throughput. The second tier of storage to Archer comprises 214Tb in NetApp filers where users can store their data between simulations.