Bright Cluster Manager for HPC

The importance of middleware

Bright Computing logo

No matter how advanced the hardware in an HPC cluster, if the nodes are not operating efficiently, if jobs cannot be managed easily, or if software updates require frequent downtime, then that cluster will perform very poorly. The solution that fuses individual computing nodes into a unified cluster, allows a cluster administrator to manage operating systems and applications on that cluster, and permits users to submit computing jobs through an intuitive user interface is called middleware. And the most comprehensive middleware solution, one that is used in the most demanding HPC environments in the world, is the HPC Cluster Manager by Bright Computing.

Bright Cluster Manager is a comprehensive, all-in-one solution for HPC. You can also check out these two smaller scale applications as stand-alone products. (Both are integrated into Bright.)

HPC suite editions

A powerful and comprehensive HPC management tool deployed in leading enterprise organizations, inluding Dell and Boeing

- DICE Evaluation of Cluster ManagerIcon representing PDF file

- 451 Group Market Insight on Bright ComputingIcon representing PDF file

Bright Cluster Manager is a comprehensive cluster management solution for managing all types of HPC clusters and server farms, including CPU and GPU clusters, storage and database clusters, and big data Hadoop clusters. Bright also manages your clusters in the cloud. Bright Cluster Manager was designed by a group of highly experienced cluster specialists who perceived the need for a fundamental approach to the technical challenges posed by cluster management. Best of all, it is designed with those with little technical experience in mind. Focus on running and growing your organization, and let Bright Cluster Manager manage your supercomputing.

Bright Cluster Manager is available in two editions: Standard and Advanced. You can easily upgrade from the Standard to the Advanced edition as your cluster grows in size and complexity.

  1. Bright Cluster Manager Standard Edition includes all features and functionalities required for small and medium size clusters, for example up to one to two hundred nodes. The Standard Edition does not include support for redundant head nodes, cluster health management and off-loadable provisioning.
  2. Bright Cluster Manager Advanced Edition includes all features and functionalities developed for the complete application. The Advanced Edition is suitable for complex clusters of any size. Examples of advanced features are support for redundant head nodes, cluster health management and off-loadable provisioning

Bright Cluster Manager Key Features

Both editions of Bright Cluster Manager come with many of the major components which work together to offer complete control of the HPC cluster. Below are some shared key features of the HPC management tool, as well as a couple features exclusive to the advanced edition.

Cluster Management GUI

Easy and streamlined management; single and consistent interface

Bright Cluster Manager includes a first-party developed Graphical User Interface (GUI) that was designed to make cluster management simple, appealing to those with minimal Linux or HPC knowledge, while not sacrificing efficiency or functionality for experienced users.

The cluster management GUI is multi-user, allowing for the creation of user roles with different permission levels. Cluster Manager can be run on the head node of the cluster or as a stand-alone desktop application that can be installed on a local desktop PC or laptop which connects to one or more clusters. It is available for Linux, Windows and soon on Mac OS X.

Cluster Optimization

Linux optimization; node provisioning; node identification

Bright Cluster Manager is based on Linux and offers x86/64 distributions. Bright improves Linux integration by optimizing the distribution on a HPC cluster, removing unnecessary components and bringing specs up to Intel Cluster Ready standards.

Bright Cluster Manager builds upon its streamlined and efficient base with sophisticated node provisioning and image management systems, allowing you to install individual nodes or complete clusters, create/manage/use as many different node images as needed, and much more. Bright also utilizes the capabilities of managed Ethernet switches to detect which regular node is connected to which switch port, simplifying the process of identification and saving large amounts of time.

MPI Libraries

OpenMPI, MPICH, MVAPICH, and more

Bright Cluster Manager offers a choice of message passing libraries, including MPI libraries optimized for high-speed interconnects such as InfiniBand and Myrinet. All MPI libraries installed on Bright can be used with - and are already cross-compiled with - the compilers installed on Bright Cluster Manager.

  1. OpenMPI - OpenMPI is an open source implementation of version 2.0 of the MPI standard, which includes support for most common cluster interconnects, like TCP/IP networks, InfiniBand and Myrinet.
  2. MPICH2 - MPICH2 is an open source, portable implementation of version 2.0 of the MPI standard.
  3. MVAPICH - MVAPICH is an open source implementation of version 1.2 of the MPI standard. It is a port of MPICH on top of InfiniBand developed by Ohio State University for the InfiniBand interconnects.

See the complete list of libraries

GPU Management & Monitoring

Easily see GPU metrics alongside CPU

Bright Cluster Manager includes powerful GPU management and monitoring capabilities that leverage functionality in NVIDIA's Tesla GPUs to take maximum control of the GPUs and gain insight in their status and activity over time. Bright also includes the necessary CUDA and OpenCL libraries.

Bright Cluster Manager allows for alerts and actions to be triggered automatically when GPU metric thresholds are exceeded. Such rules are completely configurable to suit your requirements, and any built-in cluster management command, Linux command, or shell script can be used as an action. In addition, Cluster Health Management can also include health checks for GPU cards and GPU Computing Systems in GPU clusters. Any of the supported GPU metrics can be used in regular and prejob health checks.

ScaleMP Management & Monitoring (Advanced Only)

Aggregate multiple cluster nodes into a single virtual SMP node

Cluster Manager Advanced Edition can aggregate multiple cluster nodes into a single virtual SMP node using ScaleMP's Versatile SMP (vSMP) architecture. Capabilities include on-the-fly deployment, provisioning, configuration, monitoring and management of virtual SMP nodes.

Bright Computing's integration with vSMP Foundation means that creating and dismantling a virtual SMP node can be achieved with just a few clicks within the cluster management GUI. Virtual SMP nodes can even be built and launched automatically on-the-fly using the scripting capabilities of the cluster management shell.

Multi-Cluster Management (Advanced Only)

Run clusters in New York, Tokyo, and Oslo from your laptop!

In the advanced version of Bright Cluster Manager, you have the ability to run multiple clusters from a single installation and GUI. Each cluster is shown separately in the resources tree, allowing easy and simultaneous management of geographically dispersed clusters.