File and Object Storage

Containerized IBM Spectrum Scale brings native supercomputer performance data access to Red Hat OpenShift

By Matthew Geiser posted Tue October 27, 2020 08:42 AM

  

Organizations are moving beyond cloud to a hybrid cloud architecture and strategy.   It is estimated in the next few years that most enterprises will build their IT strategies around a mix of on-premises private cloud, public clouds and their existing platforms.  Hybrid cloud is not easy and platforms such as Red Hat OpenShift are clearly focused on making that process easier.  Enterprise Strategy Group estimates that 74% of organizations are using containers in production now or planning to do so within the next year.[1]  As hybrid cloud continues to advance in deployments, challenges with data access and data complexity with the hybrid cloud remain.  Organizations need a hybrid cloud strategy for both current applications and workloads along with new modern applications on platforms such as OpenShift.   

Now IBM Spectrum Scale Version 5.1 introduces container-native storage access.  This new enhancement scheduled to release in December 2020 combines a container native storage application with the high performance global parallel file system of IBM Spectrum Scale.  This changes the landscape of container native storage and creates a simple way to create  container native storage for kubernetes and combine it with the high performance of IBM Spectrum Scale.  Performance is optimized with parallel access and enhanced throughput to storage resources.  Our current test results show that performance for containerized applications is the same as for non-containerized applications.  This means that containerized applications can run at the same level of performance as the fastest supercomputers in the world.   This is the best of both worlds and eliminates silos of data and duplicate data within a single YB scale cluster of data.  Access to high performance data can now be combined with other applications and workloads along with containerized applications from Red Hat OpenShift.


Unlike some enterprise storage options that provide Kubernetes storage using just a container storage interface (CSI) and sometimes an operator interface (CSI operator), IBM Spectrum Scale includes a software defined containerized software application.  The container-native storage access can scale performance and data availability as required within seconds.  IBM Spectrum Scale containerized access is also not limited to capacity restrictions on the compute nodes like other containerized storage options.  Capacity is clustered  from one or more IBM Spectrum Scale node(s) and is maintained as part of the federated file system, which can be integrated from edge to the core data center to public cloud. 

IBM Spectrum Scale with container-native storage access is perfect for customers looking to leverage high performance data access with enterprise data services for their Red Hat OpenShift platforms.  Customers can now bring their AI, HPC and big data applications to Red Hat OpenShift with the confidence that they have an intelligent storage infrastructure that will meet the high performance and growing capacity needs they require.   

"I can see how IBM Spectrum Scale with container-native storage access is ideal for users who want to leverage high performance data access for Red Hat OpenShift," said Joel Zysman, Director of Advanced Computing at University of Miami. "Users can now bring AI, Big Data, and analytics applications to Red Hat OpenShift, knowing they have scalable storage that will meet their high performance, availability, reliability, and growing capacity needs."

Customers can also leverage the parallel architecture and optimized, easy to scale capacity of IBM Spectrum Scale for other hybrid cloud applications and to consolidate and create an intelligent AI information infrastructure for the entire organization.  Using the continuous real-time data catalog of IBM Spectrum Discover customers can also add data governance, data index or classification and AI data optimization to list of use cases for containerized data.


IBM Spectrum Scale container native storage access is a powerful addition to IBM’s storage portfolio for the hybrid cloud.  To learn more about IBM Spectrum Scale you can visit: https://www.ibm.com/products/spectrum-scale

[1] Source: ESG Research Report, Data Storage Trends in an Increasingly Hybrid Cloud World, March 2020, https://www.esg-global.com/research/esg-research-report-data-storage-trends-in-an-increasingly-hybrid-cloud-world

0 comments
289 views

Permalink