Edge Computing in Action: Call Center
Ready Networks developed an edge computing solution for an insurance company based in Italy that had acquired a number of smaller firms in the U.S. The solution involved moving its call center application to Amazon Web Services and leveraging natural language processing (NLP) to automatically route calls that come into an 800 number to the appropriate call center.
“So, we took NLP, a series of Lambda scripts and technologies, and built an AWS framework that was distributed over AWS architecture,” explains Sam Barhoumeh of Ready Networks. From a network latency perspective, Ready Networks designed the data warehouse and framework so that all workloads were as close to the consumer as possible.
“Whenever an inbound call came in from the East Coast versus West that call traffic was routed through an intelligent application load balancer and directed to the appropriate region. When you [the caller] got handed off to the live agent, that agent knew a lot of information that you provided the automation behind it in real time. The outcome for the customer is better customer service, so less wait time and less compute because it's distributed.”
In addition, he says, “The application itself was driving data around data metrics, so we can actually use Lambda and the databases behind it to make other determinations based on analytics.”
Edge Computing in Action: Food Storage
A food storage company turned to its MSP, Buchanan Technologies, to develop a more efficient way to ensure its freezers maintained a safe temperature range. “They had some sensors that were built into their refrigeration devices. …It just provided raw data that needed to be transformed into information,” explains Buchanan’s Chuck Ma. “So, if you're giving me a 5 or 7 or 9, what does that mean? And then from there we need to process that and turn that into an alert or notification,” enabling the plant manager to take action.
Buchanan developed an edge solution that monitored data from the on-site SQL database server via a cloud API. When the temperature range became unacceptable the application sent a PSTN call to the plant manager. “Email was going to be too slow, and the technology they had at the time was a phone,” Ma says.
The solution did require programming. “That's typically what we see,” he notes. “There's got to be some kind of development efforts because you're taking something that was done in a more traditional fashion, and now you're trying to change it to really take advantage of the edge.”