In a contributed article for Internet of Business, Marcio Saito, chief technology officer at infrastructure management specialist Opengear, advocates a mix of centralized and edge computing approaches to keeping IoT-enabled retail systems up and running.
Too often, organizations miss the mark with IoT, by buying into trends before their ‘shape’ is fully revealed. That’s particularly true when it comes to network infrastructure. Providing the responsiveness and resilience that IoT solutions demand calls for not only greater network capacity, but also careful consideration of where those resources should reside. There may be logical cases to be made for centralizing additional network data handling and processing power, for example, or for investing instead in robust infrastructure at the network edge.
To get some perspective on this, let’s look at the retail industry. Retail store locations have undergone a tremendous expansion in their use of technology over a relatively short period of time, and are poised to repeat this achievement as IoT technology arrives.
In the recent past, a handful of point-of-sale checkout systems, some computers used for inventory management, and a security camera system may have represented the entirety of a retail location’s technology. These systems would feed information to a local data closet, containing servers to run the applications used to process transactions at checkout, keep track of inventory, and manage security video. With these systems of the past, retail locations remained their own small technology islands, with little in the way of external network connectivity – data stored locally may have been batched and sent to the corporate network as a daily or even weekly process.
Read more: Feeling the chill: Bringing IoT to cold chain logistics in retail
Jumping forward to examine the current state of retail, it’s clear that connectivity is now ubiquitous. Cash registers are often just tablets or other devices running internet-connected point-of-sale apps. Inventory data is available over the network and supports real-time insights. Workers can check stock using handheld devices or scanners, for example, while managers can oversee the supply chain like never before. Security systems now rely on web-connected cameras that offer higher quality surveillance and greater efficiency. Any local servers once found in a retail location have been replaced by cloud servers in constant communication with the store’s many connected devices. In IT infrastructure terms, network resources have shifted decidedly away from the network edge and to centralized data centers represented by the cloud.
As a result, retail locations now require constant internet connectivity to conduct business and network resilience has become an essential asset for retailers. The introduction of IoT technologies at retail locations will only deepen this dependence, and greatly so. New technologies won’t change the fact that network failures inevitably occur, but more resilient systems and automatic failover capabilities can change the impact on businesses and ensure that the retail experience and purchasing continues unhindered.
Read more: Shoppermotion uses “previously unavailable” IoT data to transform retail
In the centre, or out on the edge?
While the prevailing trend has been toward greater centralization of network resources, it’s by no means certain that this will continue into the IoT era. Because IoT solutions stand to drive exponential increases in the quantity of data that must be stored, transmitted and processed, placing considerable network resources at the network edge to facilitate this demand may be the most effective infrastructural option.
Take, for an example, a fully IoT-enabled retail store, where sensors and beacons make it possible to track the location of each item for sale, change digital price tag displays remotely, provide security, and observe and optimize the customer retail experience as a whole. Providing the tremendous data handling resources required to deliver these capabilities with real-time responsiveness will be a challenge for which nearby IT components may offer the best solution. By placing mesh wireless gateways, data aggregators, local data storage, and analysis engines at the network edge, tasks most in need of low-latency responses can be prioritized for handling locally, and related data can be processed and put into action without involving centralized network resources.
For high-demand IoT systems, edge infrastructure on the scale of micro data centers may be necessary. Effectively (and cost-effectively) implementing network infrastructure at this scale will call for remote provisioning and out-of-band management technologies. Given the size and complexity of the network infrastructure that will be required to meet the loftier goals of IoT technology, enterprises will need to rely heavily upon intelligent automation, able to stand up and manage the vast and disperse networks we’re likely to see.
Read more: Retail IoT deployment to take off in 2018, says JDA/PwC report
IT infrastructure designers should be attentive to the fact that we are at an inflection point. Where in the past, the focus was on consolidating complexity at the core, in the next few, years there will be an increased need to be able to provision, maintain and manage remote infrastructure at the edge.
To keep pace with the changes, retailers – and organizations in other sectors, too – might be wise to keep their options open, building up cloud, mobile, and network resilience capabilities that will serve them well, no matter which IoT technologies they adopt down the road. When the right opportunity to implement IoT solutions arrives, the well-prepared among them can then build the necessary network infrastructure upon a sound and reliable foundation.