A retail store with IoT infrastructure transmits as much as 10 GB of data every hour. That amounts to more data than an average internet user uses every month. While retail store’s consumption is bound to be more than an average user, it still becomes a significant challenge when it comes to processing this data, and adds heavily to the operational costs.
Currently, cloud computing sends this data over to data centres first, before it can be processed. While that is all well and good, edge computing can do it faster and more cost efficiently. With edge computing, this data is processed near, or at the ‘edge’,of the IoT network, performing analytics and knowledge generation at or near the source of data.
Let’s look at the challenges with cloud computing
Although cloud computing provides several benefits, there are also increasing challenges when it comes to handling massive volumes of data:
Cost
While cloud computing itself is affordable, the costs can skyrocket when data volumes increase. Tuning the platform to receive and transmit increased amount of data means expenditure on bandwidth, improved infrastructure, and also higher maintenance costs.
Latency
How fast data is transferred from point A to B is a major focus for enterprises as business literally depend on the speed of data movement. With the increasing adoption of IoT devices for performance monitoring, the whole game is based on responsiveness: real-time, timely response to the data gathered. And cloud latency issues cause delays that result in serious losses.
Edge computing rectifies all these problems, easily
When data is processed at the edge of the network:
- It reduces the load on the network and saves bandwidth costs
- Data is processed and analyzed faster and the results can be swiftly relayed back to the systems for appropriate action. The IoT network is hence more responsive and issues are identified and rectified faster.
Here’s how Edge computing works
Here’s a look at the technologies used to enable edge computing:
Fog computing
It is a decentralized computing infrastructure which not only pushes the data to the the edge of the network, but does so efficiently by putting it in the most relevant places between the cloud and the origin of data. This process is called placing it ‘out in the fog’.
Micro data centres
These data centers provide all the essential components of a traditional data center. But what makes them better than the latter is their size. They are much smaller, and are especially beneficial for SMEs that cannot afford to have their own data centres.
Mobile edge computing
It is a network architecture that improves the network efficiency by placing the computational and storage resources within the Radio Access Network(RAN). By doing so, the delivery of content to the end user becomes more efficient.
Cloudlets
It is an architectural element, formed because of the convergence of cloud computing and mobile computing, and represents the second tier of the three tier hierarchy – mobile device > cloudlet > cloud. Its primary purpose is to improve resource-intensive mobile applications by providing computing resources with lower latency to mobile devices.
With effective use of Fog computing and micro data centres, and optimal placement of storage resources within the RAN, they’re able reduce the data processing time significantly. Many industries are adopting the edge approach, and rightly so.
How are different industries leveraging Edge Computing?
Retail
Brick and mortar retail owners are always looking for ways to get an edge over their competitors online. With the help of instant edge analytics, they’re able to collect valuable customer insights, like which coupons were used by them and which products the customers picked up from the display. Edge devices like Beacon help collect this information and allows retail businesses to make the user experience more personalized and effective, by targeting promotions and sales items as the customers enter the store. When a customer is near the store, (or, when he enters the ‘proximity boundary’) Beacon notifies him about things like active deals, products, events and more.
Oil and gas
Predicting a disaster before it happens is what makes the real difference in the oil and gas industry, since it is so prone to on-site explosions. The traditional centralized data infrastructures can assess what caused the explosion, but they can’t help predict such a disaster. Edge computing, on the other hand, can quickly analyze data, spot anomalies, and raise alerts to helps prevent recurrence of such disasters.
Transportation industry
Traffic management is an ideal application for edge computing. City authorities can deploy intelligent transportation management system to compute applications locally, on physical traffic infrastructure like signals, which would reduce the noisy data at the edge. This would, then, significantly reduce the amount of data to be transmitted via networks which would lead to online storage cost reduction. Moreover, since the data is handled at the edge, it would also reduce the latency that would’ve occured if it was transmitted via cloud.
Manufacturing industry
There hasn’t been a significant rise in the annual productivity gains in the manufacturing industry over the years. But that might change, soon.
With the rise of Industrial Internet of Things (IIoT), manufacturing industry stands a chance to improve that. Monitoring the factory floor and the multitude of precision operations can be done better with the help of connected devices. And all of this data can be instantaneously processed via edge computing, to spot inefficiencies across the manufacturing and supply chains.
Nearly every industry one could think of will be impacted by the Internet of Things (IoT). As volumes of data increase, edge computing will become a necessary addition to the enterprise IT infrastructure. Not to say edge will completely replace cloud computing, but it will definitely emerge to be a more viable investment in a lot of cases. And as with all technology evolutions, early adopters will have a significant advantage over competitors.
Author Bio: Sriram Sitaraman: Practice Head for Analytics and Data Science at Srijan Technologies.
With over 20 years of experience in designing and delivering innovative business solutions, Sriram leverages his expertise in machine learning, statistical modelling, and business intelligence to enable digital transformation in industries as diverse as healthcare, manufacturing, retail, banking and more.
