IBM HPC is NOW

By Andy Morris, IBM Cognitive Infrastructure

November 5, 2019

 

[Attend the IBM Spectrum Scale User Group at SC19!]

 

HPC is NOW” as the theme of this year’s International Conference for High Performance Computing, Networking, Storage, and Analysis (SC19). The Technical Seminars offered by the IBM HPC team fall in line very well with the conference theme because they introduce and explore data storage and management technologies available today to help organizations of all types and sizes effectively leverage the advantages of the AI and Big Data applications rapidly sweeping across modern business and science landscapes. You don’t have to wait for innovation to catch up to where you want to go. IBM HPC is NOW.

The very first SC19 IBM Technical Seminar is titled – AI needs IA: The Critical Role of Information Architecture for AI Success. Early adopters of AI-enhanced applications and technologies have quickly learned that these new solutions need unprecedented amounts of training data, processing power, and system performance to achieve acceptable levels of insight and accuracy.1 Like never before, data collection, preparation, movement, optimization, and storage – or what is termed the data pipeline – has been pushed center stage. An effective data pipeline needs a high performance, highly resilient, cost-efficient, multicloud IT infrastructure, often referred to in this context as an information architecture.

The “AI needs IA” session will explore the data management requirements of different elements of the AI data pipeline and how key storage management technologies work together to provide the infrastructure needed to initiate and grow AI projects. Solutions for unstructured file data will be a key focus of the session, and when any conversation turns to high-performance file system management, IBM Spectrum Scale will surely be mentioned.

[Also read: Building a Solid IA for Your AI]

 

This year, innovation within the IBM Spectrum Scale platform has continued at such a pace that the topic is worthy of an entire SC19 Tech Seminar of its own, titled The Secret [Weapon] Behind the World’s Smartest Supercomputers. The title is a reference to the fact that IBM Spectrum Scale provides the data management for two of the fastest supercomputers on the planet right now – Summit and Sierra – as Pangea III the world’s most powerful commercial supercomputer for Total, a supermajor global energy company operating in more than 130 countries. The IBM POWER9-based supercomputer will help Total more accurately locate new resources and better assess the potential associated revenue opportunities. In addition, according to Total, Pangea III requires 1.5 Megawatts, compared to 4.5 MW for its predecessor system. Combined with the increased performance of Pangea III, Total has reported that they have observed that the new system uses less than 10% the energy consumption per petaflop as its predecessor. 2

IBM Spectrum Scale simplifies data management with integrated tools designed to help organizations manage petabytes of data and billions of files. Today, it is installed in clusters supporting enterprise and HPC applications ranging from climate modeling and tornado simulation to big data MapReduce analytics, gene sequencing, digital media, and scalable file serving.

IBM Spectrum Scale is deployed globally across environments as diverse as the financial, energy, automobile manufacturers, telecommunications, retail, digital media, and healthcare and biotechnology industries, as well as in many science and government use cases.

As organizations move to multicloud environments and begin to adopt AI applications, the scalability, flexibility and data sharing capabilities of IBM Spectrum Scale have increased its popularity. For example, innovative transparent cloud tiering enables non-disruptive, intelligent data migration between flash, disk, tape, and even cloud storage tiers, helping enterprises more easily bridge storage silos on-premises plus add the benefits of cloud storage to their overall data solutions.

[Read also: Breaking Boundaries with Supercomputer Storage]

The IBM Tech Seminar at SC19 will highlight a number of enhancements and new features to IBM Spectrum Scale. A new Erasure Code Edition enables deployment on storage-rich commodity servers for greater flexibility and cost-efficiency. The platform has moved forward to version 5.0.4, with new support for the industry-standard Container Storage Interface (CSI) across all IBM Spectrum Scale configurations and deployment models. The latest version provides improved fidelity for the Apple SMB protocol, automated monitoring and reporting of an even wider range of system data, and seamless scaling of the monitoring capability to hundreds of nodes. There is also a new pre-tested blueprint that simplifies deployment of storage solutions optimized for AI and Big Data analytics workloads.

You can’t talk for very long about IBM Spectrum Scale without mentioning one of its most powerful and popular deployment models – Elastic Storage Server (ESS). Recently, IBM has released a new grid storage solution based on the basic ESS architecture called Elastic Storage System (ESS) 3000. The IBM Tech Seminar titled – Faster Time to Value and Flexibility for Your AI/ML/Analytics Workloads introduces ESS 3000 and explains how it offers faster time-to-value and easier maintainability for resource-limited IT organizations.

By utilizing containerized software delivery, the new solution is designed to be one of the simplest ways to deploy IBM Spectrum Scale’s world-class file system technology. A key innovation is the Non-Volatile Memory Express (NVMe)-enabled flash storage leveraged to achieve ultra-low latency and massive throughput. ESS 3000 brings ground-breaking capabilities for AI-based applications and workloads, providing more than enough performance and throughput to enable full GPU resource utilization in servers such as IBM Power AC922, as well as the fastest benchmarks with NVIDIA DGX systems.

As always, IBM will have a substantial presence at SC19. The IBM HPC team will be available all conference long at booth #1525. And of course, SC19 attendees can schedule single client briefings at the nearby Hyatt Regency, or attend a variety of User Group meetings. One of the most popular and beneficial element of the SC19 technical program is the Technical Seminars we host. IBM subject matter experts will cover a variety of hot topics in the Tech Seminars – from artificial intelligence (AI) and supercomputers to containers and Kubernetes for HPC environments.

 


References

[1] TechTarget Whatis definition: Deep Learning (https://searchenterpriseai.techtarget.com/definition/deep-learning-deep-neural-network)

 

2 IBM Infrastructure website: The world’s most powerful supercomputers https://www.ibm.com/it-infrastructure/power/supercomputing

Return to Solution Channel Homepage

IBM Resources

Follow @IBMSystems

IBM Systems on Facebook

Do NOT follow this link or you will be banned from the site!
Share This