Edge computing has been on our radar for the past couple of years, thanks to the rise of the Internet of Things, augmented/virtual reality and 5G. But during the pandemic we’re experiencing an unprecedented demand on networks — from video conferencing to content streaming — that requires high bandwidth and zero latency.
Indeed, global consultancy Bain & Company suggests that COVID-19 lockdowns may accelerate the shift to edge computing, since “dramatic shifts in traffic patterns have exposed weaknesses in network infrastructure, strengthening the case for investments in technology that reduces bottlenecks.”
But moving computing closer to the edge isn’t exactly easy — and perhaps not always necessary. Some of these benefits and challenges were addressed in a recent IDG TECHtalk Twitter chat, Edge Computing: A New Business Edge Emerges.
The online chat was moderated by Isaac Sacolick, president and CIO of StarCIO and author of Driving Digital, who posed questions to participants on how they define edge computing, its use cases and the challenges to implementation. Here we sum up the chat and some of its key findings.
IDG: How do you define edge computing?
Edge computing is defined by @willkelly as a “computing model that brings computation and data storage closer to the location where you need it, saving bandwidth and response time.” Or, as @joannefriedman defines it, “the ability to process and act on data wherever it’s created.”
In an Internet of Things context, it means computing happens closer to the ‘things.’ But there’s two aspects to this,” says @webjedi, “as there’s an edge of the provider/enterprise and the edge of the consumer/user. Those touchpoints are the edge. Users have the edge via their devices & network (LTE/WiFi) and enterprises are cloud or on-prem configs.”
IDG: How is your organization using (or working toward using) edge computing in a COVID world?
When IDG polled participants on whether their organization was moving toward edge computing, half of respondents said yes while the other half said no. Despite this, “there are actually 100s of global enterprises (not just tech companies) employing or beginning projects for #edge,” said @mthiele10.
But for @nyike, edge computing “is a byproduct of cloud hype” that resulted in quick decisions made without considering workload types, costs, safety and long-term scale.
“A lot of organizations spent a lot of time and political capital convincing senior leaders and #CFO[s] to spend on centralization and #cloud,” said @DigitalSecArch, and edge “requires a high use case to reverse some of that project work and redirect CapEx.”
Several participants pointed out that it’s not easy to implement edge computing and many organizations don’t have the resources or expertise. So, naturally that leads to …
IDG: What is the most promising edge computing use case, and why?
According to @sarbjeetjohal, edge use cases can be divided into two categories: consumer and industrial. Consumer use cases suggested by participants include gaming, connected homes and driverless cars, while industrial use cases include robotics, connected offices and connected factories.
@ArsalanAKhan suggests algorithmic trading as a potential use case (because latency can be expensive), @EvanKirstel sees possibilities for ag-tech and @willkelly is curious to see how edge could be applied to DevOps and DevSecOps.
But the lines between industrial and consumer aren’t always clear, such as with finance and IoT applications. And new models might emerge as suggested by @joannefriedman, who said “the best use case for #edge in manufacturing is creating direct-to-consumer business models that don’t exist to engage consumers as early as possible in product ideation & design.”
Several participants see use cases in healthcare. @itlinchpin says that “for #healthcare SaaS providers, edge computing is needed to reduce latency and improve performance especially to those clients who support such a large influx of new patients due to COVID.” But several participants pointed out that hospitals may need a more effective cybersecurity strategy before moving to edge computing.
IDG: What are the biggest benefits that edge computing offers?
In another poll, of the six participants who responded to the question, three voted for speed and latency and one each for cost savings, reliability and scalability. Several felt that reduced latency would produce cost savings and improve customer relationships.
If we dig deeper, said @sarbjeetjohal, the need for edge computing is driven by the need for better networks, and that in turn is driven by a need for speed and reliability. “I believe in ‘speed is the new scale’ thinking. #Edge gives you [a] speed advantage. Just like cloud, if you focus on cost alone, you will damage #edge.”
IDG: What is your biggest concern around edge computing?
Some participants were concerned about management buy-in and the cost effectiveness of deployment. But most agreed with @joannefriedman that security is the biggest hurdle to overcome.
“Especially when trying to retrofit security after deployment,” said @benrothke. “When it was never engineered into the infrastructure in the first place. Kiss all those #edge cost savings goodbye.”
That’s why @joannefriedman is “a huge supporter of #Securitybydesign” and says it’s time we engineer and build secure hardware — SRAM, PCA/PCB, FPGA and NIC — to #buildforedge.
IDG: What advice do you have for those looking to adopt edge computing? What are some ‘edge’ best practices?
“Don’t believe all the hype around #edge. Have your teams spend lots of time defining ‘your’ specific use cases. Know ‘your’ exact requirements. Only then can you know if you can benefit. If you just deploy it w/o your requirements, expect wasted time & money,” said @benrothke.
@sarbjeetjohal suggests taking the approach of ‘cloud when you can, edge when you have to,’ since “#edge operations management will be more expensive than cloud for [a] considerable number of years.” Several participants agree it’s worth getting expert help and working with partners instead of trying to do it all yourself.