Edge computing is when data collected from the sensors is not being computed centrally in the cloud, but rather stored and computed on the Edge devices (nodes) of an IoT network. Thus we can look at Edge computing as some kind of cloud computing optimization method, since some part of the IoT application is shifted on the Edge which is actually in contact to the physical world.
Why would we now go back to computing on the Edge of the network when we have all moved to the cloud? There are numerous reasons, but one of them is definitely the cost associated to storage, power consumption and cooling in the data centers. Another important reason for Edge computing is that you don’t need to send all the collected data over the network, you may just need aggregated and analyzed data since this will influence the associated connectivity costs. Another reason is reducing the latency/delay which helps to increase the overall Qulity of Service (QoS). At the end of the day the most important thing is the fact that it will enable you to make faster decisions and generally take actions on time. Edge computing is a good fit for many applications, until the day when we’ll get super high bandwith and super low latency pipes between the cloud and our gateways.
Who needs this kind of computing solution? Well, the list is almost endless, anybody who can’t afford any delay in processing and needs analysis of important data in near real time – I’ll just name some: healthcare, autonomous vehicles, finance, asset management, critical power issues, process optimization, predictive analytics and all the other latency-sensitive micro services. IDC predicts that by 2020. Edge computing will take up to 18% of overall IoT infrastructure spending.
Edge computing can bring a lot of benefits. From security perspective Edge represents a smaller attack surface since devices are distributed and less data is sent over the Internet. Companies, like Microsoft with its Azure Sphere, are taking care of very important security aspect – management of IoT devices. Azure devices are certified microcontrollers with managed Linux OS and a cloud service, so they are automatically updated and also securely managed centrally.
Another usage of Edge computing could be a hyper-personalized experiences on the web sites – dynamic, content-based user profile with preferences, location, time of day, previous interactions.
Who is into Edge Computing market development? Well, we have already mentioned Microsoft with its Azure Sphere, but we also have other big players: Amazon with its AWS Greengrass and Google with Edge TPU and Cloud IoT Edge. Besides them, we have a bunch of other big names in Edge Computing’s arena: HPE, Nokia, Scale Computing, Vertiv, Huawei Technologies, Fujitsu, NVIDIA, General Electric, Intel, Dell, IBM, Cisco, SAP SE.
As you can see there are numerous reasons (it’s faster, it’s making cost savings, it’s more intelligent) why we need Edge computing as an add-on to today’s traditional computing approach which involves: data collection, getting the data to the cloud or data center, process, data analysis and so on. If you have some more interesting information or Edge computing use case you’d like to share with us, please write to firstname.lastname@example.org.