A Glimpse at Edge Computing Architecture

Modern IoT systems have large number of IoT devices which generate enormous amounts of data. Since IoT device resources are usually limited, data processing is typically performed in the cloud. However, using cloud resources can raise additional issues, like latency and additional cost due to bandwidth consumption. Due to those factors a new technology has evolved during the last few years – Edge Computing. Edge computing extends cloud-computing capabilities to the edge of the network, allowing local data processing and analysis in order to reduce latency and bandwidth consumption.

Several Edge Computing reference architectures exists. Existing architectures can be classified into categories such as data-based, security-based, machine-learning-based and more, however none of them provides full solution.

When choosing the right architecture, one should keep in mind issues like security, scalability, device management and interoperability.

Note that Fog Computing (sometimes mentioned with or instead Edge Computing) is a slightly different concept. Fog Computing is a decentralized infrastructure of computing nodes in which the services provided to end-users are located between end-users and the cloud.