Edge computing continues to be increasingly important for organizations to make better data driven decisions. Since edge computing is a big generator of data and since data requires storage, storage is a major player. Let’s examine briefly how IBM positioned its approach to edge computing at its recent Think 2020 Digital conference, how it is expanding its role in edge computing and how IBM Storage, notably IBM Spectrum Scale and IBM Cloud Object Storage, is assisting those efforts.
IBM Expands Its Role in Edge Computing
Edge computing is generally thought of as a distributed IT architecture, where the data collected by an edge device can be processed locally to improve response time and save on bandwidth. This definition is somewhat confining as we shall see. Now, an edge device may be an endpoint on a private or public network, such as a mobile computing device like a smart phone or a pervasive computing device, say a sensor-based Internet of Things [IoT] solution.
At Think 2020, IBM and Red Hat announced new edge computing solutions for 5G network technologies, which bring with them unprecedented speed and extremely low latency. The goal is to enable enterprises to overcome the complexity of managing workloads across a massive number of devices. In particular, the IBM Telco Network Cloud Manager offering is designed to help telcos to quickly deliver 5G-based edge-enabled services to customers. In general, IBM plans to enable users to deliver value-adding insights through the application of AI and analytics close to the edge.
Now, back to the definition of edge computing. How local is local? Local should mean close enough to do the job properly in the most cost-efficient manner. All talk about computing being distributed or decentralized vs. centralized or core would seem to be irrelevant. For example, IBM uses Red Hat OpenShift as a linchpin for its edge computing work. Red Hat OpenShift even works at the mainframe level with IBM’s storage (see IBM Continues to Extend Its Mainframe Storage Solutions). Yet it would seem likely that, unless the mainframe can be used for edge computing from its current site, other IBM storage options could be used to place the necessary storage as close to the edge as necessary.
The Role of IBM Storage in Edge Computing
Although IBM’s data and AI storage solutions are often installed in central computing installations, they can also be deployed as appropriate to the edge as edge computing storage solutions.
Start off with the two members of the IBM Spectrum Storage software-defined storage (SDS) suite that provide the software horsepower for its data and AI storage solutions — namely IBM Spectrum Scale and IBM Cloud Object Storage. Now IBM Spectrum Scale would be the choice if the decision is made to manage the data at the edge as file data whereas IBM Cloud Object Storage would be the choice if the data can best be managed as objects.
Please note that IBM’s SDS software can be sold separately from its storage hardware. This means that a partner within IBM’s edge ecosystem could choose non-IBM storage hardware to go along with the chosen Spectrum Storage software. Naturally, IBM would prefer to sell software and storage bundled together and the company believes that the value proposition of its integrated storage hardware/software solutions is profound.
In the case of IBM Spectrum Scale, the bundled solution for edge storage goes under the name of Elastic Storage System (ESS) 3000 whereas for IBM Cloud Object Storage the bundled solution has the same name, IBM Cloud Object Storage, as the software standalone product. Let’s focus on what each software product brings to the edge computing table as the hardware is covered elsewhere starting with IBM Reinforces Storage Portfolio for AI and Big Data.
IBM Spectrum Scale’s global single namespace enables management of data as a single virtual pool and at any scale necessary. This capability has enabled IBM Spectrum Scale to bring many years of successful deployment in large centralized installations, but IBM Spectrum Scale also brings features, such as Active File Management (AFM), and capabilities that make it attractive for edge computing. AFM automatically shares and caches data across geographically distributed size to achieve performance levels similar to local data. This allows nondisruptive, intelligent data integration between edge locations and central data centers using a single-pane window management.
IBM Cloud Object Storage uses its concurrent parallel access capability to bring data from the edge quickly and efficiently with geo–dispersed data protection. IBM Cloud Object Storage enables enterprises to store and manage massive amounts of data efficiently and securely with extreme system reliability and accessibility from any location.
IBM Spectrum Scale and IBM Cloud Object Storage are also key parts of IBM Storage Suite for IBM Cloud Paks, enterprise-grade container software packages that are designed to offer customers faster, more reliable ways to build, move and manage applications and workloads in hybrid clouds. Enterprises use this Suite, which is a pre-tested reference architecture with data resources for file, block, and object requirements, to speed up their development processes.
Mesabi musings
At IBM Think 2020, IBM and Red Hat launched new edge computing solutions for the 5G era. But edge computing is not only about processing power for storage also takes a major role. And IBM can play this major role with the use of IBM Spectrum Scale and IBM Cloud Object Storage, which deliver the flexibility and scalability with the AI and analytics capabilities that many edge computing applications, such as telcos, demand. Edge computing in the 5G era should be quite exciting.