May 31, 2022 By Prem D'Cruz 4 min read

We are excited to announce the release of Spectrum Scale for High Performance Computing (HPC) on the IBM Cloud, powered by Intel Xeon Servers.

With this, we now have the ability to offer an end-to-end HPC system ready for consumption out of the box. The offering is pre-integrated with four workload schedulers of your choice, a large selection of balanced, compute and memory VPC profiles and high-performance file storage with parallel scale-out architecture.

Why is a good workload scheduler so critical? Without scheduling, workloads are dispatched in a haphazard fashion, leading to suboptimal resource use and longer execution times — as depicted by the white spaces in the figure below:

With sophisticated scheduling, we use resources more efficiently and can optimize for high utilization, better performance or both:

We offer you the choice of our market-leading schedulers: IBM Spectrum LSF, IBM Spectrum Symphony, open-source Slurm Workload Manager and Red Hat OpenShift as a managed service for containerized workloads.

Workload management services

IBM Spectrum LSF (“LSF” — short for load-sharing facility) is industry-leading, enterprise-class software. LSF distributes work across existing heterogeneous IT resources to create a shared, scalable and fault-tolerant infrastructure. LSF provides a resource management framework that takes your job requirements, finds the best resources to run the job and monitors its progress. Jobs always run according to host load and site policies. It is suited for Life Sciences, Weather, Electronic Design, Oil and Gas and Industrial workloads.

IBM Spectrum Symphony® software delivers powerful, enterprise-class management for running compute-intensive and data-intensive distributed applications. It provides support for a hybrid HPC cloud, enabling workloads to be forwarded to multiple clouds and data to be automatically staged to or from the cloud. Further, resources that are consumed on the cloud can be autoscaled based on workload demands and scheduling policies. It is used extensively in the Financial Services sector for risk modeling and predictive analytics.

Slurm Workload Manager software delivers powerful, enterprise-class management for running compute-intensive and data-intensive distributed applications. The software is a highly scalable cluster-management and job-scheduling offering that is open source and fault tolerant. It is used extensively in government and universities for research and development.

Red Hat OpenShift on IBM Cloud is a leading platform to quickly and securely containerize and deploy enterprise workloads in Kubernetes clusters. IBM manages the OpenShift Container Platform, and you set up the HPC system by using the automated scripts from the public git repository. Red Hat OpenShift is IBM’s strategic direction for cloud-native, and clients have the ability to integrate the IBM Cloud Paks, IBM Cloud Satellite, data, AI and multicloud into their containerized HPC applications. A Kubernetes scheduler is available for managing and submitting jobs.

Virtual private cloud (VPC) HPC infrastructure services

IBM Cloud Virtual Server for VPC offers fast-provisioning compute capacity with the highest networking speeds and most secure, software-defined networking resources. You can choose dedicated hosts or multi-tenant virtual server profiles with the core and RAM configuration that’s best for your HPC workload. Instance storage is allocated and attached at the time of provisioning. You also get fine-grained control over your virtual network and security resources for greater scaling and traffic regulation. IBM Cloud offers the highest level encryption in the industry and the ability to run regulated financial services workloads.

High-performance file system (HPFS) storage services

IBM Spectrum® Scale provides a global platform that solves modern HPC challenges with a single source of truth and a global data fabric. It is a cluster file system that provides simultaneous access from multiple compute nodes. The global data fabric breaks down silos and current data limitations to bring faster HPC insights. It is globally accessible, with active file management (AFM) secured with audit logging and immutable data, protected with comprehensive resiliency options and optimized with lifecycle management for cost and data caching for performance.

Value proposition from IBM

With this latest announcement, IBM now offers Spectrum Scale, Spectrum LSF and Spectrum Symphony as tiles in the IBM Cloud catalog. Using Terraform scripts and configuration parameters, we offer one-click provisioning of the HPC clusters and installation of the middleware software:

The architecture above clearly illustrates how the various components work together and why workload management (Spectrum LSF) and high-performance storage (Spectrum Scale) are so critical to achieve the best business value for HPC in the cloud. Optionally, for higher speed data transfers, we can use IBM Aspera with its unique transport technology.

The following are three key criteria to consider when making a Cloud HPC:

  • The ability to run workloads on the most cost-effective platform based on defined policies.
  • The ability to dynamically resize your HPC cloud hosts based on workload requirements and automate the pay for use by shutting down clusters when not required
  • The ability to manage your cloud and on-premises data with effective pre-staging and caching to control ingress and egress charges.

IBM is a leader and well-positioned to assist you with hybrid HPC that requires full interoperability between on-premises and the cloud. We have helped multiple clients safely navigate their journey to the cloud and have over 25 years of experience with on-premises HPC. We encourage you to bring your applications to us and we will help guide you on the best approach.

More from Announcements

IBM Hybrid Cloud Mesh and Red Hat Service Interconnect: A new era of app-centric connectivity 

2 min read - To meet customer demands, applications are expected to be performing at their best at all times. Simultaneously, applications need to be flexible and cost effective, and therefore supported by an underlying infrastructure that is equally reliant, performant and secure as the applications themselves.   Easier said than done. According to EMA's 2024 Network Management Megatrends report only 42% of responding IT professionals would rate their network operations as successful.   In this era of hyper-distributed infrastructure where our users, apps, and data…

IBM named a Leader in Gartner Magic Quadrant for SIEM, for the 14th consecutive time

3 min read - Security operations is getting more complex and inefficient with too many tools, too much data and simply too much to do. According to a study done by IBM, SOC team members are only able to handle half of the alerts that they should be reviewing in a typical workday. This potentially leads to missing the important alerts that are critical to an organization's security. Thus, choosing the right SIEM solution can be transformative for security teams, helping them manage alerts…

IBM and MuleSoft expand global relationship to accelerate modernization on IBM Power 

2 min read - As companies undergo digital transformation, they rely on APIs as the backbone for providing new services and customer experiences. While APIs can simplify application development and deliver integrated solutions, IT shops must have a robust solution to effectively manage and govern them to ensure that response times and costs are kept low for all applications. Many customers use Salesforce’s MuleSoft, named a leader by Gartner® in full lifecycle API management for seven consecutive times, to manage and secure APIs across…

IBM Newsletters

Get our newsletters and topic updates that deliver the latest thought leadership and insights on emerging trends.
Subscribe now More newsletters