KNIME Cluster Execution

Compute clusters often run idle because of a lack of applications that can be run in the cluster environment and the enormous effort required to operate, maintain, and support applications on the grid. KNIME Cluster Execution tackles this problem by providing a thin connection layer between KNIME and the cluster, which allows every node running in KNIME and every application integrated in KNIME to be executed on the cluster. Submission of data to the cluster and collection of the results is made very simple. Long-running analysis workflows can be executed on the compute cluster, thus releasing local resources for other productive work.

Splitting compute-intensive tasks into subsets and executing them on different resources in the cluster.

Executing individual nodes on different resources in the cluster.

Advantages

Performance

An important advantage is the gain in performance for calculation intensive workflows.

Disconnect

KNIME Cluster Execution allows you to disconnect from running jobs, continue to work on other urgent tasks and later reconnect to those jobs to check for status changes and retrieve the results.

Transparent

Operation of the cluster environment is simple and fully transparent to the user.

Third Party Node Support

Third party nodes can also be routed to dedicated servers, making it possible to distribute software that does not usually offer cluster support.

Required Software

KNIME Cluster Execution is realized as a KNIME plug-in and allows the integration of the Sun Grid Engine.

Submit Clients

Linux SUSE 10/11, Fedora 10 Red Hat Enterprise Linux 5 (32 and 64 bit)

Cluster Engine

Sun Grid Engine (SGE 6.2)

Cluster Slaves

Linux SUSE 10/11, Fedora 10 Red Hat Enterprise Linux 5 (32 and 64 bit)

Please contact us for more information about KNIME Cluster Execution.