How to include edge computing in your 2020 IT budget

Internet of Things (IoT) devices are gaining a great deal of traction among enterprises, as valuable information lives on the edge — embedded in products, on factory floors, at remote work sites, and within vehicles. To fully leverage the benefits of edge computing, careful planning is needed, as hidden or unexpected costs can arise, making your edge initiatives more expensive than anticipated.

In a very broad definition, edge computing encompasses the generation, collection, and analysis of data where the data is generated. This is a contrast to centralized locations such as data centers. Internet of Things devices are typically considered to include some edge computing, since some of the data can be processed on-device, while in other circumstances, data is transmitted in real-time or data aggregates are transmitted on programmable intervals to a central system.

Some edge computing (or IoT) devices may be single-use, or have a shorter shelf life than standard computing devices, due to limitations such as non-replaceable batteries. 

Realistically, the costs of edge computing can vary wildly, depending on the size and scale of your deployment, the amount of data being collected and processed, and the geographic location of your edge computing deployment. 

First, consider the number of sensors that are needed. In certain deployments, this can work as a hub-and-spoke model, with smart tags that are recorded when they pass through exits (or other physical thresholds in a building). For manufacturing settings — among others — this can also work with barcodes and scanners, minimizing costs of single-use scanners. Expenses associated with RFID-enabled smart tags can add up quickly, as standard tags run for 15 cents at wholesale prices.

The amount of data being collected and processed can also play a role — for IoT devices that trigger serverless events in Lambda, the frequency and length of events may be a consideration. When properly optimized, serverless computing should be more cost-effective than fixed-rate virtual machines, though comes with an inherent variability as charges are based on the number of events and run time.

Likewise, location makes a significant difference for your deployment. For on-premises devices hardwired to a network or using wi-fi, the connection cost is essentially free — there’s not really an additional cost for transit, just power and internet connectivity already available at your location. For remote devices, using mobile (cellular) networks typically incurs a cost, as charges are typically metered on a per-gigabyte basis.