An analytical performance model of robotic storage libraries

Large scale scientific projects generate and use huge amounts of data. For example, the NASA EOSDIS project is expected to archive one petabyte per year of raw satellite data. This data is made automatically available for processing into higher level data products and for dissemination to the scientific community. Such large volumes of data can only be stored in robotic storage libraries (RSLs) for near-line access. A characteristic of RSLs is the use of a robot arm that transfers media between a storage rack and the read/write drives, thus multiplying the capacity of the system. The performance of the RSLs can be a critical limiting factor of the performance of the archive system. However, the many interacting components of a RSL make a performance analysis difficult. In addition, different RSL components can have widely varying performance characteristics. This paper describes our work to develop analytical performance models of a RSL. The advantage of an analytical model over simulation is the fast solution times, making feasible a variety of performance studies. We present some sample studies, including a study of clustering, of device selection, of device sizing, and of distributing a workload between two RSLs. Finally, we show how the RSL model can be incorporated into a queuing network model. The models described in this paper, developed for the NASA EOSIDS project, are implemented in C with a well-defined interface. The source code and accompanying documentation are available through WWW at: http://www.cis.ufl.edu/(similar to)ted/.