The concept of edge computing on a fundamental level refers to the information technology architecture that allows for client data to be processed as close to the original source as possible. Essentially, IoT sensors and connected devices transmit data to a nearby edge computing device, for example a switch or router that processes or analyses the data, rather than sending it back to the cloud or a remote data centre. In other words, edge computing enables analytics and data gathering to occur at the source of the data. While not an entirely new concept, the necessity and popularity of edge computing has notably increased in more recent times. So what has triggered this movement, and what are the benefits?
Why do we need edge computing?
Until now the role of edge computing has largely been to ingest, store, filter, and send data to cloud systems. The advancement of technology, however, means that these computing systems can now pack more compute, storage, and analytic power to utilise and act on the data at the machine location.
Additionally, a report by Goldman Sachs shows that the cost of IoT sensors has halved in the last ten years, the cost of bandwidth has reduced by 40X, and the cost of processing has reduced by 60X. This means that the ability to scale the numbers of sensors being used, as well as data collected, has increased dramatically. Of course, this all requires more support within global and local data centres to cope with the sheer amount of information and processor power needed to sift through this data and act on it.
All this said, it’s not just the reduced cost and ease of deploying IoT sensors and connected devices that is pushing this move to the edge – it’s also us.
As our homes, cars, watches, phones, TVs, and a variety of other household items become more connected, the number of devices needing connectivity and links back to those central data centres increases. More devices, means more activity is recorded, meaning that the potential for running data analysis across the information collected increases. This pushes more focus back to what else can be connected to transmit data, and in turn, everything else increases – the data centre capacity, the number of devices and the connectivity.
What does this mean for managed service providers?
As these devices start to better engage with each other, the need to quickly transmit information between sensors will explode. For example, temperatures recorded in a factory need to be transmitted, analysed and quickly adjusted in a self-regulating IoT enabled environment. This is fine for a couple of sensors, but what about hundreds of thousands of sensors, all of which are transmitting information that moves through complex algorithms to deliver recommendations or automated changes based on machine learning.
This type of data collection requires computing to be close, not in a data centre 3000 miles away as the move to IoT enabled devices pushes the requirements for better connectivity and speedier services. The result will be a true hybrid approach to cloud computing, with the large, central data centres storing vast amounts of information, and local data centres delivering super-fast connectivity to local customers.
Consequently, organisations that would have traditionally consumed vast amounts of cloud services globally in central data centres, are now supplementing their global delivery with local deployments in colocation facilities in order to support geographically dispersed customers. For example, if hundreds of thousands of customers reside in the Birmingham area and need more resource, then the organisation needs to move IT infrastructure locally to provide the level of service required.
What are the benefits?
The main benefits of edge computing address the concerns of many business decision makers, including: better safety, reduced cost, and an increase in speed.
The nature of edge computing means that more data is stored in local centres, therefore if an issue was to occur at the central data centre, a business can continue running without disruption to the service. Equally, data transmission costs are decreased as the amount of data being transferred back to the core location is condensed. It is the decision of the business as to how much data is kept locally, and how much is kept at the core data centre, therefore increasing flexibility and personalisation of service.
The future of edge computing
It has been predicted that by 2020 there will be 25 billion IoT- based installed devices. As the quantity of traffic between devices and machines continues to increase, so too will the pressure on central data centres. Edge computing is proving to be an effective solution to this problem, offering both businesses and consumers increased reliability and efficiency of data.
Dave Ricketts is head of marketing at Six Degrees