Edge computing moves your data closer to your customers

Small companies are finding success using cloud services to extend their on-premises infrastructure to the edge.

Edge computing has been gaining momentum thanks to a highly distributed workforce since the COVID-19 pandemic began.

Edge computing allows compute resources, including analytical and other data processing, to be carried out closer to the sources of the data, says David Linthicum, chief cloud strategy officer at Deloitte Consulting. "Thus, we don't have to ship all of the data back to some sort of centralized processing existing in a traditional data center or on a public cloud,'' he says. "This increases the performance of those systems and does not require constant connectivity back to a centralized processing center."

These mechanisms have a lower cost and faster ROI, which appeals to small- and medium-sized businesses, Linthicum adds.

Momentum for edge computing is clearly building. "Edge computing will transform every aspect of the computing and communication ecosystem and the value chain,'' writes tech consulting firm Chetan Sharma. Among the emerging use cases are augmented and virtual reality (where processing takes place locally), autonomous vehicles (where processing is run in the vehicle and in local stations), and the use of facial recognition for security (where analytics are run on the camera vs. on a central computer), Sharma says. 

Another common use case is content caching, where large web pages and videos may be stored closer to the users that request them. 

Forrester's February 2021 report "Transform IT To Deliver A Future Fit Tech Strategy" calls edge computing one of the emerging technologies for new opportunities to "empower customers, partners and employees to decide and act immediately."  

Deloitte agrees, noting that "growth opportunities abound for tech companies that execute on all forms of digital transformation," particularly in areas including cloud, AI, analytics, RPA, cybersecurity and edge computing. "Because edge computing can process huge chunks of data in real-time, less data needs to be processed in the cloud, taking pressure off the network," according to the firm's 2021 outlook for the US tech industry.

How to excel at the edge

Edge computing means that you rent space on servers that are physically closer to the applications, people and devices that need them. SMBs are adopting edge computing by deploying a mix of traditional, on-premises systems with public clouds, Linthicum says. Like any technology, the challenges for edge computing include security and operations.   

"If you're dealing with sensitive information, either regulated or not, you still have to protect that data no matter if it's on an edge device or in a public cloud,'' Linthicum says.  "This means that you'll have to invest in security systems that are able to support security for data in flight and at rest, on both the edge devices and centralized systems."   

Operations is a challenge because IT still needs to manage and monitor many edge devices, Linthicum says. IT also has to perform preventative maintenance and fixes on these devices, including updating firmware, replacing devices that fail, and fixing network issues, among others. "Edge operations, or EdgeOps, are often underestimated in terms of cost and level of effort required," he notes.

Canadian-based TeamBuilding, which offers online team building events, uses edge caching for faster performance for clients like Apple, Amazon, Google, Johnson & Johnson, Netflix, Chipotle and NASA, says CEO Michael Alexis.

The company has over 100 employees, and prior to the pandemic, it organized local activities like renegade museum tours and guacamole-making competitions, Alexis says. "When COVID hit, we shifted to online events, [and] part of the reason our technical performance is so important to me is that it's important to our clients,'' he says.

There is also utility in marketing and sales. "About 10 years ago, Amazon shared that 100ms of latency on their site cost them 1 percent in sales,'' Alexis notes. "A similar review by Google found that adding .5 seconds to generating a search page dropped traffic by 20 percent. I had these studies in mind when we designed our client booking app [to be] lightning-fast." Clients can choose events and dates without delay, which helps increase bookings, he says. 

Caching at the edge means that the data is stored close to clients whether they are in the U.S., Canada, Europe, Asia or elsewhere, Alexis says.

"This upgrade wasn't a spontaneous investment, it came from months and years of testing and improving technology. Specifically, we've noticed a direct correlation between the speed of our applications and other important metrics, like duration of use and lead volume,'' he adds. "In addition to edge caching, we've started experimenting with other technologies like JavaScript Service Workers, which can further improve performance."

Generally speaking, large static files such as images, videos and long scripts are best for edge caching, he notes. Dynamic content that changes frequently by visit and user, is not a great candidate, according to Alexis.

However, this "leads to a bit of a tradeoff," Alexis adds. While TeamBuilding wants its applications to run as fast as possible, some variables may get "stuck" and not update, he explains.

It isn't a major issue, but some users may notice the discrepancy and it affects their experience, Alexis says. Post-COVID, he anticipates becoming a hybrid organization. "This shift to remote has directly correlated with our interest in edge caching technologies — for us, it makes sense for a digital workforce and service to have the best digital infrastructure available."

Like Alexis, Miranda Yan, co-founder of VinPit, a Singapore-based car data company, believes computing at the edge "is the next level to taking a business to the digital landscape. It enables businesses to cache their server data to the locations near to their customers, improving their performance and reducing the bottleneck issues."

Yan started moving data to the edge at the end of 2020 and has also opted for a hybrid environment. "Due to high traffic, the site used to load slowly … We deployed [on] numbers of data centers for various regions to resolve the issue." Already, Yan says, she is seeing better results.

Getting to the edge

For organizations thinking about moving to the edge, there are some considerations to take into account. 

"It really comes down to need first and technology solutions second for small- and mid-sized businesses,'' Linthicum says. "This means understanding what the technology is capable of, and where it might be applied to the benefit of the business."

He advises companies to focus on a proof-of-concept first, before making any significant investments, since that will lead to a more pragmatic understanding of the technology as well as cost and risk.    

Alexis suggests that organizations interested in caching at the edge to treat their investments with the 80:20 principle. "You can often get 80 percent of the results for 20 percent of the input,'' meaning cash, time and effort, he says. "With technology, it can seem important to aim for 100 percent, as it's often a large investment, but I believe that even hitting 80 percent would result in major gains for many organizations."

Learn more about edge computing solutions from Dell here.