IBM Platform Data Manager for LSF

IBM Platform Data Manager for LSF takes control of data transfers to help organizations improve data throughput and lower costs by minimizing wasted compute cycles and conserving disk space. Platform Data Manager automates the transfer of data used by application workloads running on IBM Platform LSF clusters and the cloud, bringing frequently used data closer to compute resources by storing it in a smart, managed cache that can be shared among users and workloads.

By conducting data transfers out of band, Platform Data Manager eliminates wasted processing cycles to maximize processing resource utilization. Data is staged in and out independent of workloads, freeing Please Recycle compute resources to work on other jobs while data is transferred back and forth among compute resources and storage. Platform Data Manager also leverages the underlying file transfer infrastructure already in place, allowing it to be easily integrated into existing cluster infrastructure.

Platform Data Manager uses a smart, managed cache to reuse data common to multiple workloads. This preserves disk space and bandwidth by eliminating the need to repeatedly transfer the same data between storage and compute resources as different workloads are processed. With smart caching, Platform Data Manager reduplicates data transfers, optimizing the use of storage resources and bandwidth. Transferred files are automatically cached on the execution cluster with an optional time to live. Cached copies can be leveraged for all workloads that need access to the data, and can be shared between multiple users where appropriate. Workloads running on execution clusters can also write intermediate data to the local cache (such as restart data) for use by other jobs.

Key Points

Product Features