Jim Holland, Regional Director, Africa, Lenovo Infrastructure Solutions Group.

Solutions Marketing Manager for Lenovo Infrastructure Group, Bob Nevins, said in a recent blog that Artificial Intelligence (AI) has the potential to create US$9- to US$15 trillion of economic value in the coming years, according to a McKinsey report.

At the same time, data creation and computing at the “Edge” is flourishing. By 2023, over 50% of new enterprise IT infrastructure will be deployed at the edge rather than in traditional datacentres, up from less than 10% in January 20212. And according to Gartner, by 2025, 75% of enterprise-critical data will be created and processed outside the datacentre or cloud.

Companies are embracing the growth of AI and Analytics solutions at the edge to gain real-time insights, transform customer experience, and gain competitive advantage. Even in South Africa, research shows AI pilots and experimentation are now prolific across local companies, with businesses showing a willingness to embrace AI and experiment using new technology.

Data, whether at the edge or in the core datacentre, is the raw material required to derive insights and take action. Data management at the edge has requirements similar to the datacentre, including protection and performance. However, edge locations also have unique challenges for simplicity, security, and data mobility. 

Edge data management solutions must integrate into standard edge networks while protecting against the increased potential for security attacks present at the edge. And many edge data locations are not managed, so edge data management must be simple, efficient, and optimised to move data from the edge to the core or cloud for end-to-end control.

Deploying effective Edge AI requires complete infrastructure solutions. 

In this article I’ll be talking about important new features of that will help businesses secure, accelerate, and manage data-rich edge environments.

Cybersecurity enhancements to protect against Ransomware

Security and performance are top considerations when deploying Edge Storage solutions. By 2025, at least 75% of IT organisations will face one or more attacks.And edge locations can be particularly vulnerable, since they lack many of the defences found in traditional datacentres.

Businesses should be seeking capabilities to detect and remediate ransomware, such as immutable read-only snapshot copies which can be locked via WORM technology - a capability by which information, once written, cannot be modified, so that attackers cannot change or delete the snapshots.

Also a good note to choose storage arrays with built in and automatic ransomware detection. These features typically use Machine Learning to understand workload patterns and behaviour on the NAS volumes.If unusual activity is detected, the software should take an immediate snapshot to provide a restore point and generate an alert to the administrators so that they can examine the activity and determine whether it’s an attack or a false positive.

All this must be done without disrupting normal operation and significantly improves cybersecurity capabilities.Even better, it’s a non-disruptive software-only upgrade.

Accelerated data access to speed application performance

Companies are looking to edge AI solutions for real-time insights, requiring high-speed performance. We’re talking about solutions that support NVMe over TCP, which operates over ubiquitous Ethernet networks.

NVMe/TCP provides application acceleration over any Ethernet network, using existing Ethernet switches, cabling and adapters, so that no infrastructure upgrades are required - making it an ideal upgrade for existing edge ethernet networks to enable high performance AI applications.

Again, this new application acceleration capability comes as a non-disruptive software-only upgrade.

Data Protection for Edge Object Workloads

Invest in true unified storage, which supports block, file, and object workloads. Integrated S3 object storage capabilities are particularly well-suited for edge AI applications, many of which rely on unstructured object data such as video, voice, etc. Support for replication of object storage provides customers with a familiar interface for fast, efficient backups of object storage to a supported target.

This can be used in several ways, including backup for applications and workloads, business continuity applications with site failover or edge-to-core object replication.