This chapter describes the IBM technical computing software portfolio. First, for a better understanding of the portfolio, we explain the IBM technical computing software concept. Then, we introduce each product line. Finally, we cover the latest current user roadmap.

The following sections are covered in this chapter:

The IBM Platform Computing products can simplify the setup, integration, and management of the heterogeneous technical computing infrastructure while it drives up server utilization, increases application throughput, and helps to greatly improve the time to results. And they also help integrate servers, storage, parallel execution environments, and applications. This integration enables the delivery of complete solutions that greatly simplify and accelerate deployment and management of high-performance clusters, grids, and High Performance Computing (HPC) clouds. IBM Platform Computing products are divided into two main categories: workload management and cluster management.

HPC applications need huge computing power. The purpose of workload management in technical computing is to allocate an HPC application, such as a service-oriented architecture (SOA) workload or batch workload with huge computing process, to the large-scale distributed computing environments. Workload management uses computing resources efficiently to complete workloads as fast as possible. To enable an efficient workload allocation, an intelligent scheduling policy is required. An intelligent scheduling policy is based on understanding shared computing resources, the priority of the application, and user policies. At the same time, a mission-critical application, which performs a large-scale workload with complicated calculations and big data, requires reliability, scalability, and high processing power. And the larger the environment, the more heterogeneous it is, so it also requires the seamless integration of the heterogeneous environment. Therefore, the best workload management software must offer comprehensive features for optimization of HPC application.

An HPC cluster generally consists of a large-scale distributed environment. It includes a server, storage, and network hardware, as well as operating systems, a hypervisor, and middleware. And it is a heterogeneous environment that consists of multiple vendors or versions. For a large-scale complex environment, we have to install, configure, manage, and monitor it throughout its lifecycle. So, integrated management software, which can be operated from a single point in an efficient way, is required. It is easier to provision a cluster, including the operating system/hypervisor, middleware, and configurations. Depending on the usage situation, it needs to be able to increase, decrease, and change computing resources in a simple way. All operational processes must be as automated as possible to avoid human error.

IBM Platform Computing products are comprehensive products, which provide workload management and cluster management. In addition, they complement the IBM systems and technology portfolio by providing simplified management software to help eliminate the complexity of optimizing cluster, grid, and HPC Cloud environments. Figure 2-1 shows the IBM Platform Computing portfolio.

In this section, we summarize each product. We provide more detailed explanations of each product in later chapters.

The IBM Platform Load Sharing Facility (LSF) product family provides powerful workload management for demanding, distributed, and mission-critical technical computing environments. It includes a complete set of resource-aware scheduling, monitoring, workflow, analysis, and license management capabilities, all designed to work together to address HPC needs.

IBM Platform LSF V8.3 includes a comprehensive set of intelligent, policy-driven scheduling features. These features enable the full utilization of compute infrastructure resources and position them for a high return on investment. The highly scalable and available architecture of the IBM Platform LSF allows users to schedule complex workloads and administrators to manage small clusters up to peta FLOP-scale resources while it increases application throughput. With one of the best support structures in the HPC industry, the IBM Platform LSF product family provides one of the most complete HPC data center solutions for workload management.

The IBM Platform LSF product family can help you in the following ways:

IBM Platform LSF runs on various x86 hardware and operating environments, including the latest generation of System x servers. It is also certified on IBM Power Systems™ servers that run the AIX® and Linux operating systems. By pre-qualifying and certifying these platforms, IBM helps you take the risk out of mission-critical high-performance technical computing deployments.

IBM Platform LSF manages and accelerates workload processing for compute-intensive or data-intensive applications across distributed compute environments. With support for heterogeneous compute environments, IBM Platform LSF can fully use all the infrastructure resources that are needed for policy-driven, prioritized service levels for always-on access to resources. A comprehensive set of intelligent scheduling policies ensures that the right resources are automatically allocated to the right jobs, for maximum application performance and efficiency. Through a powerful command-line interface, users can monitor that the existing IT infrastructures are optimally utilized. More work is done in a shorter amount of time with fewer resources and hardware and administration costs are reduced.

IBM Platform LSF offers these add-ons:

IBM Platform Analytics is an advanced tool for visualizing and analyzing massive amounts of workload data. It enables managers, planners, and administrators to easily correlate job, resource, and license data from one or multiple clusters for data-driven decision making. With better insight into high performance computing data center environments, organizations can identify and quickly remove bottlenecks, spot emerging trends, and plan capacity more effectively. Traditional business intelligence (BI) solutions require significant time and multiple steps to translate raw data into usable information. However, IBM Platform Analytics incorporates innovative visualization tools that are built on top of a powerful analytics engine for quick and easy results. You can utilize the preconfigured dashboards or construct your own. You can quickly answer questions about your technical computing infrastructure and applications, and use that information to optimize technical computing resource utilization.

By allocating a virtualized pool of licenses that is based on the distribution policies of an organization, IBM Platform License Scheduler enables license sharing of FLEXlm licenses between global project teams. It prioritizes license availability by workload, user, and project so that licenses are optimally used. An intuitive web-based console enables license usage to be monitored in real time. Whether the application software environment is simple or complex, IBM Platform License Scheduler helps organizations optimize their use of enterprise software licenses, improving productivity and containing cost.

By reducing or removing the need for operator intervention to trigger computational workflows, IBM Platform Process Manager compresses end-to-end cycle time. Using an intuitive web-based interface, designers can describe workflow steps and dependencies so that lengthy, repetitive tasks that are prone to human error are automated. User-defined and system-defined calendars can be combined so that workflows or individual jobs can be run automatically at predefined times. Technical users employ the XML-based file formats and the rich command set that allows time-dependent jobs or flows to be triggered and managed through scripts as an alternative to the graphical interface. The result is a more cost-effective, logical, self-documenting solution for workflow automation and scheduling.

IBM Platform RTM is an operational dashboard for IBM Platform LSF environments that provides comprehensive workload monitoring, reporting, and management. It makes cluster administrators more efficient in their day-to-day activities and provides the information and tools that are needed to improve cluster efficiency, enable better user productivity, and contain or reduce costs. Dashboards provide comprehensive reports to support the day-to-day administrative tasks that are associated with managing single and multiple cluster environments. Timely information on the status of the HPC environment helps improve decision-making, reduce costs, and increase service levels.

IBM Platform Application Center provides a flexible application-centric portal for users and administrators to interact with their HPC cluster or grid in a natural and powerful way. The web-based interface simplifies workload management with remote job monitoring, easy access to job-related data, and the capability to manage jobs, such as stopping, suspending, resuming, or requeuing jobs. Intuitive, self-documenting scripting guidelines provide standardized access to applications. This standardized access enables administrators to better enforce site policies and simplify the creation of job submission templates, which results in reduced setup time and minimizes user errors during job submissions. To further simplify application integration, predefined templates for many applications are available with the product:

By configuring these templates based on the application settings, users can start running jobs without writing custom wrapper scripts. For users that want to integrate their custom applications directly with their cluster, IBM Platform Application Center includes an extensive web services application programming interface (API). This API is for custom application integrations, extended visualization support, and integration with other IBM Platform Computing products, such as IBM Platform Process Manager.

IBM Platform Message Passing Interface (MPI) V8.3 is a high-performance, production-quality implementation of the MPI. It is widely used in the high performance computing (HPC) industry and is considered one of the standards for developing scalable, parallel applications. IBM Platform MPI maintains full backward compatibility with HP-MPI and applications that are supported by it. IBM Platform MPI incorporates advanced CPU affinity features, dynamic selection of interface libraries, superior workload manager integrations, and improved performance and scalability.

IBM Platform MPI supports the broadest range of industry-standard platforms, interconnects, and operating systems to help ensure that parallel applications can run almost anywhere. It runs on various hardware and operating environments, including the latest generation of System x servers. By pre-qualifying and certifying these platforms, IBM helps clients take the risk out of mission-critical high performance technical computing deployments. IBM Platform MPI can help clients:

IBM Platform Symphony V5.2 is an enterprise-class grid manager for running distributed application services on a scalable, shared, heterogeneous grid. It accelerates various compute and data-intensive applications, quickly computing results while using the optimal available infrastructure. The IBM Platform Symphony efficient low-latency middleware and scheduling architecture are designed to provide the performance and agility that are required to predictably meet and exceed throughput goals for the most demanding analytic workloads. Designed for reliability and having advanced management features, IBM Platform Symphony helps organizations realize improved application performance at a reduced total cost of ownership. IBM Platform Symphony can help you achieve these goals:

IBM Platform Symphony runs on various hardware and operating environments, including the latest generation of System x servers. By pre-qualifying and certifying these platforms in large-scale environments, you take the risk out of deploying mission-critical grid computing applications.

IBM Platform Symphony offers these features:

IBM Platform HPC is easy-to-use, yet comprehensive technical computing management software. The robust cluster and workload management capabilities are accessible by using the latest design in web-based interfaces - making it powerful, yet easy to use. IBM Platform HPC simplifies the application integration process so that users can focus on their work, instead of managing a cluster. For applications that require MPI, the robust commercial MPI library accelerates and scales HPC applications for shorter time to solution. Other HPC cluster solutions combine multiple tools and interfaces, which are not integrated, certified, or tested together. IBM Platform HPC is a single product with a unified set of management capabilities that make it easy to harness the power and scalability of a technical computing cluster, resulting in shorter time to system readiness and user productivity as well as optimal throughput. Backed by the best client support in the industry, IBM Platform HPC incorporates nearly two decades of product and technology leadership.

IBM Platform HPC delivers the following key benefits:

IBM Platform Cluster Manager V3.2 Advanced Edition automates the self-service assembly of multiple heterogeneous HPC and technical computing environments on a shared compute infrastructure. The cluster manager creates an HPC Cloud for users to run technical computing and analytics workloads. This cloud offers the following capabilities:

IBM Platform Cluster Manager Advanced Edition can deliver these capabilities:

IBM Platform Cluster Manager Advanced Edition is designed to provide more function than the traditional cluster management solutions:

IBM Platform Cluster Manager Advanced Edition runs on the latest generation of IBM System x iDataPlex®, Intelligent Cluster™, and other rack-based servers and is also supported on non-IBM industry standard x64 hardware. By pre-qualifying and certifying these platforms at scale, IBM can help you take the risk out of deploying mission-critical grid computing deployments.

Table 2-1 shows the current user roadmap.