What is Edge Computing?
Edge computing is decentralised data processing that takes place at the edge of the network. The data collected by sensors is filtered, compressed and sent on directly on or in the end device. By simply analysing the data, previously defined tasks can also be carried out centrally. Accordingly, edge computing is particularly important when large amounts of data, including Big Data called, arise and are to be processed and filtered quickly on site. This enables quick decisions and reduces the amount of data to be transmitted.
Areas of application and examples
Edge Computing and IIoT
The abbreviation IIoT stands for Industrial Internet of Things. The term IIoT is also often used as a synonym. Industry 4.0 used. It refers to intelligent and digitally networked machines or plants in industry to create more efficient and self-organised production. Enormous amounts of data are generated by sensors and control devices. For example, the approximately 6000 sensors of the Airbus A350 deliver around 2.5 TB of data every day. To prevent this amount of data from being unnecessarily Cloud must be transported, they are filtered and evaluated on site and only a fraction of it sent to the cloud.
Also in the area of Predictive Maintenance edge computing plays an important role, as the collected data is also primarily important on site and enables short decision-making paths. For example, that a machine needs maintenance because of conspicuous measured values.
Edge Computing and IoT
The IoT (Internet of Things) describes networked and intelligent electronics, such as those used in the smart home. Similar to the IIoT, a lot of data is generated here, which is particularly useful on site.
Edge computing and autonomous driving
Here a Combination of Edge and Fog Computing used. Control devices, sensors and actuators also cause the autonomous driving large amounts of data. Between 5 and 20 terabytes of data per day are not uncommon. Through local data analysis (code to data) in a mobile mini-data centre, with the help of Fog Computing, the data is evaluated on site and only the results are transferred. This means that the required Data processed in real time and decisions made quickly become. Because in ongoing road traffic, delays can be life-deciding. Edge computing also works offlineThis means that an autonomously controlled vehicle can easily drive through a dead zone or a tunnel and still be fully functional.
Edge computing in healthcare
Data sets in the health sector have increased by over 800 percent between 2016 and 2018. But not all of this data needs to be stored. Through edge computing, it is possible to store the relevant Filter out data directly at the terminal. For example, inconspicuous heart rates can be detected and deleted. However, abnormalities can be detected simultaneously and forwarded without latency. This makes it possible to react to emergency situations in real time.
What are the advantages and disadvantages of edge computing?
- Real-time data processing with latency minimisation
- Enables real-time monitoring and services
- Also works without an internet connection, especially important in rural areas or where there are wireless holes
- Limitation of transmission delays and service failures
- Bypasses bandwidth restrictions
- Sensitive customer and company data remains on site and does not have to be transported to the cloud
- Irregular computing or memory requirements
- More control and greater protection of terminals needed to prevent misuse and failures
- constant availability of the equipment must be guaranteed
- Elaborate initial set-up
Edge Computing vs. Cloud Computing
Edge and Cloud computing both belong to data processing models. One of the main differences is where the data is processed. With edge computing, this happens on or even in the end device, with cloud computing in a central IT structure, the cloud. Both function independently of each other, but can also be used together. Thus, large amounts of data can be filtered and reduced in advance through edge computing before they are transported to the cloud.
Applications such as complex data analyses (big data analyses), being able to access data from anywhere and the long-term storage of data are very feasible with cloud computing. The disadvantage here, however, is the speed with which a decision is made, as the data is first sent to the cloud, processed there and then the decision is sent back to the device. This can result in higher latency times. In addition, unfiltered data can lead to excessive data volume and bandwidth overload for data transport. Edge computing can relieve these problems.
Edge Computing vs. Fog Computing
Edge and Fog Computing will often used as synonyms, although they describe different approaches. In a way, Fog Computing is a mediator of the cloud infrastructure. If the cloud hovers like a cloud centrally above all end devices, Fog Computing is like a fog closer to the end devices. Thus, instead of all data being routed to the cloud, it is already processed in mini data centres (the Fog Nodes) nearby. This reduces latency and processing times.
Fog nodes are able to communicate with each other, which is not possible with edge computing. This means that More complex analyses in Fog Computing The result is that the data is much more feasible than in edge computing, which is limited to very simple analyses and, above all, to the filtering of data.
Fog, Edge and Cloud Computing work particularly well in combination with each other. For example, edge computing can pre-filter and reduce the amount of data, and initial analyses can be carried out in the Fog Node. Complex and time-consuming tasks are then forwarded to the cloud. In this way, the strengths of the different models can be used very well.