How to Keep Edge Computing Sites Truly Autonomous

Hyperconverged infrastructure (HCI) solves many problems that historically have plagued data center managers–clutter, management complexity, and high upgrade and maintenance costs created by multiple Ethernet boxes, servers, and storage arrays. Hyperconvergence removes those elements by combining data center functions into a single node.

As such, HCI is ideal for unstaffed edge computing sites, which can number in the thousands as part of a single network. As companies deploy edge environments, IT managers are coming to managed service providers (MSPs) and IT solution providers with questions along these lines: “My CIO wants me to deploy a standalone edge micro data center at each of our branch locations by the end of this month. How can I possibly install and maintain these sites located across such a wide territory?”

These “from the trenches” questions present an opportunity for IT solution providers to add significant value and cement their trusted advisor role by delivering and remotely managing hyperconverged edge sites. It’s also a great opportunity to adapt the business to a fast-growth market. In fact, a recent IDC survey revealed that 40% of end users plan to outsource their edge initiatives, making this a rapidly growing market opportunity for IT solution providers.

Keeping Edge Computing Sites Resilient

So, what is really on the minds of customers when it comes to edge computing sites? It’s all about reliability. The overriding priority at the edge is to keep everything running uninterrupted. After all, edge computing networks support mission-critical applications. They are essential to digital transformation, as organizations leverage the Internet of Things (IoT) and advanced technologies to gain agility and improve the customer experience. Even when there is a break in a site’s connection to the wide area network (WAN) or offsite infrastructure, it’s important to keep the site operating.

Currently, HCI solutions are two-node systems that can run a complete micro data center or edge site by providing processing, storage, data protection and power backup. Yet, these two-node systems require a third node or “witness” to achieve data consistency across the clustered file system in the HCI. The witness, operating remotely at a third site, maintains consistency in the cluster even if one of the nodes goes offline. But this approach is costly, especially when multiplied by all the sites on an edge network.

Affordable Innovation

To address this challenge, Schneider Electric and Cisco have teamed up to offer a cost-effective innovation that boosts reliability at the edge. Now, two-node Cisco HyperFlex solutions can have a Local/Connected Witness, without the cost of another compute node. The witness functionality is embedded in the latest generation of APC’s Network Management Card (NMC). This card resides slotted into the rear of the APC uninterruptible power supply (UPS) that supports the cluster at an edge site. Combining a witness-embedded NMC with HyperFlex helps prevent downtime during a power disturbance, as the UPS helps the edge clusters to continue functioning.

This Cisco-validated Local/Connected Witness keeps costs down and removes the need for