Edge Computing: Why small data centers are big winners
February 24th 2018, Author / Editor: ISG / Lenz Nölkel
The rapid growth of the Industrial Internet of Things is not seeming to stop soon. Changed requirements for computer centers are one of the consequences: Massive amounts of data have to be transferred between central IT and the terminals in real time – a challenge which is more and more impossible to accomplish for geographically distant data centers. Remedy can be found through Edge Computing.
Shortened Latencies through Edge Computing
The classical network architecture knows the data center as central focus: Off-grid connected devices send data to the central computer which then processes them. This process however comes not without latencies. Edge Computing on the other side is about smaller, decentralized data centers at the outer edge of the network, located near to where data is created, significantly reducing the delays. The terminals are equipped with sensors, receiving data and transmitting it to the edge computer which immediately after processing and filtering sends it back. Latencies and data volume are thus optimized.
Purposes needing short latencies in digitalized industries are, for example, production controls critical for the production process. These controls cannot bear delays in the interaction with machines and robots.
Another example is autonomous driving. In case of an emergency, decisions have to be made in less than five milliseconds, meaning that data processing has to take place as fast as possible.
Edge computing is a convincing way to guarantee the high speed transmission necessary, to which optimal mobile transmission rates are key.
Ever faster networks and more performative services
5G, the successor to LTE, will be able to guarantee low latencies which are for instance necessary for autonomous cars and scenarios from digitalized industries. However, regarding current efforts of 5G pioneers such as Qualomm, the technology won’t be on the market until 2022. Many users don’t want to wait that long.
The so called Low Power Wide Area (LPWA)-protocol could be one solution in the meantime. The most well-known technology is the narrowband-iot (NB-IoT), being characterized by low latencies, module costs and maintenance costs. NB-IoT is specified by the 3rd generation partnership projekt (3GPP), some telecommunication companies are already offering it.
Some first edge-computing-applications are already in place. Numerous providers are offering ready to use-edge computing solutions, consisting of rack, server, storage, network and the facilities, including power supply, AC, firefighting system, access protection and monitoring.
These products are set up quickly and can be easily expanded modularely if needed. Plus, there are also already as-a-service-solutions, relieving the user from arising administrative expenses. All in all, nothing bares the way to a fast implementation of edge computing anymore.
This piece is party based on an article by ISG.
Image Source: iStock/BeeBright