There was a time mainframes were the size of small data centers and minicomputers were the size of a closet. With no network to connect them, LANs and WANs started to tie everything together from the late ’80s, and by the mid-90s we had client server systems that drove the emergence of large datacenters.
Meanwhile, the internet created flexible ways of communicating with these datacenters. However, they were still one-application-to-one-server configurations. The hardware configuration was in control and data centers’ pattern and application capability were set once the hardware was built.
Then the first instance of software-defined came along – virtualization. The business case was simple: if the relationship between application and server is implemented in software, it is more flexible, which helps you squeeze more from your server investment.
The cloud took virtualization to another level by scaling up and further centralizing the data center on a continental scale. It added a crucial component: sophisticated management. Virtualization and cloud have evolved to make all aspects of centralized computing software-defined, not just reliant on the hardware alone.
We now have software-defined storage and software-defined networking in which developers can create arbitrary combinations of compute, storage and networking and scale them quickly. The key issue here is configuration is controlled by the software.
Breaking Away from a Device-Centric Architecture
Device-centric architecture served the retail industry well through the ’80s and ’90s, as electronic POS was married to LANs for the first time. In early generations, it was convenient to deploy one application per device; one computer, one operating system and one POS.
However, this arrangement comes with high overhead. Complex logistics are required to get devices, networking, power, operating systems and security software into the right configuration, in the right location, to support an application.
In today’s fast-paced environment where customer expectations of speed and new experiences are elevated, retailers need a more flexible, dynamic model that is not tied to this “one device, one application,” device-orientated model. They can no longer wait six months to get a new computer into every store to launch a new service, or send a field engineer to every site just to make a change. Retailers must deploy at the speed of the cloud, in just a few clicks, or risk being held back.
Software-Defined Edge Computing is Leading the Future
Edge computing is just one of countless emerging technologies transforming retail stores. However, because it’s infrastructure, it is a critical underpinning of the technology adoption cycle for retailers.
Edge computing is about transforming the way in which workloads are hosted and managed in distributed locations. There are plenty of reasons why these workloads will never end up being centralized and require a software defined edge strategy:
Assisting latency for time-sensitive applications
Reducing the reliance on network and bandwidth
Improving data privacy and security
In retail, edge computing provides a new paradigm for running existing applications like POS and at the same time is “oven ready” for the new world of distributed microservices. Edge computing and its associated software-defined concept gives retailers the power to make things happen instantly across the business from a single point without having to rely on armies of engineers roaming the country.
Brian Buggy is CTO and co-founder of Zynstra, an NCR Company.
Today’s retailers are eager to cruise right past the recovery stage and instead situate themselves just a hop, skip, and a jump away from the future. RIS fielded some inside perspective of today’s retail analytic investment priorities through a chat with automotive retailer Servo Pacific.
Retailers and CGs are deploying AI/ML solutions to aid in supply chain efforts at increasing rates. Explore this Targeted Research report to discover where the competition is placing big tech bets and to benchmark your supply chain of the future roadmap.