The difference between edge computing and distributed computing

Edge computing focuses on placing computing and storage resources outside of the enterprise data center, in or near the branch office. But edge computing is different from traditional decentralized computing. Edge computing is driven by the Internet of Things and will grow as 5G services become more popular. CIOs need to understand what is driving edge computing strategies to avoid recreating the branch IT problems that data center consolidation is designed to solve.

Edge computing is different from traditional distributed computing in terms of motivation, technology, architecture, and management concepts.

Traditional distributed computing stems primarily from ownership and control issues, where lines of business in branch offices want IT resources that they control and, because of this need, minimize the use of WAN network links.

In contrast, the edge computing strategy assumes that centralized delivery is the norm and that sufficient bandwidth will be available to provide services; it places compute and storage resources in colocation facilities at or near enterprise branch offices, solely to meet the functional requirements of a specific use case.

[[271359]]

Latency and Bandwidth

In fact, edge computing is becoming more and more popular now because we are entering an era of tasks that require very short response times or require digesting very large data sets, and it is more practical to do the analysis locally - or both. The Internet of Things is driving most of these use cases. IoT applications include real-time control of high-speed production lines or automatic control of driverless vehicles (for example, in a warehouse), which may require sub-millisecond response times and large-scale data analysis needs.

But the edge computing strategy goes beyond just fixing servers and storage at or near branch offices; it focuses on creating a centrally managed but physically distributed infrastructure that supports automated/autonomous operations.

Edge computing is based on cloud-based infrastructures that enable both fully centralized management and automated operations. In a converged or hyper-converged infrastructure, computing and storage resources require minimal human intervention to be put into production. In a private cloud, most configurations and all utilization of these resources are centrally managed. Extending the cloud paradigm to an edge computing model makes it as simple as possible for IT to configure each site and then add more resources if necessary; such plug-and-play components require minimal technician time and effort because they are automatically discovered and put into use. At the same time, professional support for enterprise edge computing is driving the development of new products in this area (for example, products that provide smaller form factors, consume less power or generate less heat so that they can be placed in small cabinets or wiring closets).

Rethinking branch networks and the edge

IT leaders are already working on use cases that functionally justify edge computing, and they can also use edge computing strategies as a means to change the economics of their branch networks to justify edge computing strategies financially.

When a strong platform is built for localized computing, IT departments can use it to reduce WAN work (no longer transmitting large data flows to other places) and reduce the performance requirements of the WAN because critical response times are achieved through edge resources. This may allow IT teams to reduce their site connectivity expenses, especially on multi-protocol label switching services. Edge infrastructure can also support Internet and security center functions, reducing the work of other branch offices, in addition to the central data center and the security core and Internet links.

Edge computing is only now beginning to gain traction with a broad range of enterprises, and we are beginning to see use cases converge with technology availability to make it a practical, sustainable option. Most enterprises don’t have use cases for edge computing yet, but more are every year driven by IoT and analytics projects.