As we will explore later in the chapter regarding cloud and fog computing, edge routers are capable of providing computing resources close to where the data is being generated. This is an important characteristic in edge routers, especially for IoT solutions. As more IoT solutions are built in remote and varied locations, having the ability to compute locally is essential for many use cases. Computing at the edge can involve affecting user data through techniques such as:

• Filtering and aggregation
• Denaturing data
• Security and intrusion detection analysis
• Key management
• Rules engines/event processors
• Caching and storage

Edge routers will vary in size and computational capacity. They are not data center servers or rackmount hardware. Typically, the resources on an edge router will be comparable to the following table:

The available resources need to be shared with the primary functions of the router as a network agent as well as with whatever the user designs as edge or fog software.

An edge device needs to manage and secure customer-deployed solutions on the edge, as in many situations the device will be very far from the programmer or administrator. The edge needs to be thought of in the same context as cloud-developed software and services. A router that offers edge computing services will provide an API or an SDK interface to use resources on the device and deploy programs.