The Internet of Things (IoT) is all around us, with devices gen­er­at­ing con­tinu­ous data that needs to be stored, and for critical ap­plic­a­tions, evaluated in real time. Edge computing evaluates this data directly at the source, bringing about a paradigm shift in the age of cloud computing.

What is edge computing? A defin­i­tion

Edge computing is a design approach for IoT en­vir­on­ments that provides IT resources like storage capacity and computing power as close as possible to the devices and sensors gen­er­at­ing the data. Edge computing is an al­tern­at­ive to tra­di­tion­al cloud solutions with central servers.

The term ‘edge’ alludes to the fact that in this approach, data pro­cessing does not take place centrally in the cloud but instead in a de­cent­ral­ised manner at the edge of the network. Edge computing is intended to provide what the cloud has not been able to offer so far: servers that can evaluate mass data from smart factories, supply networks or traffic systems without delay, allowing them to take immediate action in the event of an incident.

Edge computing basics at a glance

Edge computing uses es­tab­lished tech­no­lo­gies in a compact design under a new name. Here is an overview of the most important terms of edge computing:

  • Edge: In IT jargon, the ‘edge’ is the edge of the network. However, which com­pon­ents are assigned to the network edge depends on the situation. In tele­com­mu­nic­a­tions, for example, a mobile phone can be the edge of the network; in a system of networked, autonom­ously driving cars, the in­di­vidu­al vehicle.
  • Edge device: Every data gen­er­at­ing device at the edge of the network functions as an edge device. Possible data sources are sensors, machines, vehicles or in­tel­li­gent devices in an IoT en­vir­on­ment. This could be, for example, washing machines, fire detectors, light bulbs or radiator ther­mo­stats.
  • Edge gateway: An edge gateway is a computer located at the trans­ition between two networks. In IoT en­vir­on­ments, edge gateways are used as nodes between the Internet of Things and a core network.

Edge computing vs fog computing

Adding local pro­cessing instances to the cloud is not a new approach. As early as 2014, the US tech­no­logy group Cisco es­tab­lished the marketing term ‘fog computing’. Data generated in IoT en­vir­on­ments is no longer sent directly to the cloud, but is first con­sol­id­ated in small data centres, evaluated and selected for further pro­cessing steps.

Today, edge computing is seen as part of fog computing, where IT resources like computing power and storage capacity move even closer to IoT terminals at the edge of the network. A com­bin­a­tion of both concepts is also possible. The following graphic shows an ar­chi­tec­ture with cloud, fog and edge layers.

Image: Schematic representation of a cloud architecture with cloud, fog and edge layers
Schematic rep­res­ent­a­tion of a cloud ar­chi­tec­ture with cloud, fog and edge layers.
Tip

Reference ar­chi­tec­tures for fog and edge computing en­vir­on­ments are being developed as part of the Open Fog Con­sor­ti­um, an open con­sor­ti­um of industry and academia.

Why choose edge computing?

Currently, central data centres carry the majority of the data load generated by the internet. Today, however, data sources are often mobile and too far away from the central mainframe to ensure an ac­cept­able response time (latency). This is par­tic­u­larly prob­lem­at­ic for time-critical ap­plic­a­tions like machine learning and pre­dict­ive main­ten­ance.

Note

Pre­dict­ive main­ten­ance is set to re­volu­tion­ise the main­ten­ance and man­age­ment of future factories. The new main­ten­ance concept is designed to detect risks of defects using in­tel­li­gent mon­it­or­ing systems so that issues can be iden­ti­fied before an actual defect occurs.

Edge computing is not seen as a re­place­ment, but as a sup­ple­ment to the cloud, which provides the following functions:

  • Data col­lec­tion and ag­greg­a­tion: Edge computing relies on data col­lec­tion close to the source, including pre-pro­cessing and data pool selection. Uploading to the cloud only takes place if in­form­a­tion cannot be evaluated locally, detailed analyses are required, or data is to be archived.
  • Local data storage: For large numbers of data, real-time trans­mis­sion from the core data centre in the cloud is usually im­possible. This problem can be cir­cum­ven­ted by storing cor­res­pond­ing data de­cent­rally at the edge of the network. Edge gateways act as replica servers in a content delivery network.
  • AI supported mon­it­or­ing: Edge computing enables con­tinu­ous mon­it­or­ing of the connected devices. Combined with machine learning al­gorithms, status mon­it­or­ing in real time is possible.
  • M2M com­mu­nic­a­tion: Edge computing is often used in con­junc­tion with M2M com­mu­nic­a­tion to enable direct com­mu­nic­a­tion between networked devices.

The following graphic il­lus­trates the basic principle of a de­cent­ral­ised cloud ar­chi­tec­ture, in which edge gateways act as an in­ter­me­di­ary between a central computer in the cloud and IoT devices at the edge of the network.

Image: Schematic representation of an edge computing environment
Schematic rep­res­ent­a­tion of an edge computing en­vir­on­ment: Edge gateways receive data from the Internet of Things and load it into the public cloud or a private data centre as required.

How can edge computing ar­chi­tec­tures be used?

Uses for edge computing usually originate from the IoT en­vir­on­ment. An important growth driver for edge computing tech­no­logy is the in­creas­ing demand for real-time capable com­mu­nic­a­tion systems. De­cent­ral­ised data pro­cessing is, for example, clas­si­fied as a key tech­no­logy for the following projects:

  • Car-to-car com­mu­nic­a­tion: Edge computing is important for cloud-based early warning systems or autonom­ous means of trans­port­a­tion.
  • Smart grids: Thanks to de­cent­ral­ised energy man­age­ment systems, elec­tri­city grids should be able to adapt to power fluc­tu­ations. Data that is trans­por­ted to gen­er­at­ors makes it possible to react to changes in con­sump­tion in real time.
  • Smart factories: Self-or­gan­ising pro­duc­tion plants and logistics systems can be im­ple­men­ted with edge computing.

What are the ad­vant­ages of edge computing?

Compared to tra­di­tion­al cloud ar­chi­tec­tures, edge computing offers a number of ad­vant­ages:

  • Real-time data pro­cessing: Pro­cessing takes place closer to the data sources, helping to avoid issues with latency.
  • Reduced data through­put: Due to local data analysis, sig­ni­fic­antly less data needs to be trans­ferred across the network.
  • Data security: Com­pli­ance re­quire­ments can be im­ple­men­ted more easily.

What are the dis­ad­vant­ages of edge computing?

Despite the many ad­vant­ages, there are also dis­ad­vant­ages to edge computing that should be taken into account during im­ple­ment­a­tion:

  • More complex network structure: A dis­trib­uted system is more complex than a cent­ral­ised cloud in­fra­struc­ture.
  • Ac­quis­i­tion costs: Edge computing requires a lot of local hardware and therefore comes with enormous ac­quis­i­tion costs.
  • Main­ten­ance costs: Due to the large number of com­pon­ents, both main­ten­ance and ad­min­is­tra­tion costs cannot be ignored.
Go to Main Menu