Edge computing is a distributed computing paradigm that encompasses infrastructure and applications outside of centralized, dedicated and cloud data centers, located as close as necessary to the places where data is generated and consumed.

The ability to deliver IT services at the edge is increasingly becoming a requirement in our digital world. It can help companies respond faster, reduce the cost of moving and storing data, and improve compliance with data sovereignty regulations. Worldwide Edge Spending Guide (February 2024) concludes that global spending on edge computing will reach $232 billion this year, an increase of more than 15% over 2023. Further growth is expected in the coming years.

On the surface, the cost argument for deploying edge infrastructure is fairly simple: by processing data closer to where it is created, companies can reduce their network and connectivity expenses while improving performance. But the scale and scope of edge projects can quickly drive up costs.

Invests in edge infrastructure and operations always the right choice? To make the right decision for a particular implementation, an assessment of hard and soft costs is necessary.

IDC EdgeView 2024, n = 800

The hard costs of edge computing

In many ways, this is the easy part. IDCs Assessing the cost and ROI of edge computing (June 2024) sets out the fixed costs.

Fixed costs include:

• Setting up the infrastructure (servers, connectivity, storage, gateways, sensors/input devices and hardware) and integrating the edge deployment into it. The hardware required alone ranges from very basic to enterprise-class rack-based systems consisting of standalone, converged or hyper-converged infrastructure. If these systems are not already installed at the required location, setting them up can be very expensive.
• Software to support computing at the edge. This includes an operating system or operating environment, remote management and security tools, analytics and data management solutions, and application-specific software.
• Manage and orchestrate the flow of resources between multiple devices, infrastructure, and the network at the edge.
• Redesign or rebuild applications to work better in edge environments and to be automatically deployed, patched, updated, and monitored
• Data transport and storage
• Power supply and cooling at the edge location
• Remote monitoring platforms
• Deployment and ongoing management services

Indirect or hidden costs of edge computing

Indirect or hidden costs are equally important but much more difficult to quantify. These include:

• The additional costs for remote system maintenance, troubleshooting and upgrades
• Labor costs at the edge location
• Remote management
• Costs of data transfer

Ways to reduce edge computing costs

One of the best ways to reduce the cost of edge computing is to decide which application scenarios are best suited to edge, with cost being the most important factor. Take stock of the applications you plan to build or use in the next five years and their performance requirements. Those with the highest, most critical performance requirements, large data sets, and latency-sensitive workloads are good candidates.

Rather than implementing and running an edge deployment in-house, some organizations find that it is more cost-effective to use an as-a-service model. However, there are reasons other than cost to consider a service-based approach to edge computing. This is often a good idea for organizations that don’t have much experience with edge deployments, or for edge locations where technical experts aren’t available to troubleshoot problems. This is especially true in industries like healthcare and retail, where edge locations can be very remote and distributed without IT staff available at all locations. This model also allows organizations to start small with a few use cases at the edge and scale over time without spending too much money.

The three options are:

• Managed edge services. This model is similar to an outsourcing agreement where a partner owns and operates the infrastructure and applications. EdgeView 2024 According to a survey, 36% of companies say new investments in the edge will include managed edge services.
• Edge as a service. In this model, a partner owns and maintains the equipment, but the company operates it. Companies pay a monthly fee.
• Cloud services for edge. One service type in this category is a dedicated offering such as Amazon Outpost, where a public cloud provider provides dedicated equipment for a customer in a location of their choosing and all workloads running on it are dedicated to that customer. Another service type is shared cloud resources, where a public cloud provider provides resources in a location close to where the customer needs them. The same IDC survey found that 49% of enterprises say their edge investments will include edge solutions from cloud service providers.

Here’s how to judge: If your company can afford to buy the infrastructure and connectivity solutions for all locations, it can probably manage and administer edge deployments itself. If that’s not possible, it’s probably worth considering colocation or some sort of managed edge services.

Edge computing is here to stay and is unlikely to go away anytime soon. Looking at costs on a case-by-case basis is the best way to benefit from the edge model without breaking the bank.