Compute infrastructure is undergoing a massive transformation. With the explosion in cloud-based computation in the past two decades, a few critical limitations of cloud computing are coming to the fore. Data latency, the need for persistent connectivity, and the cost of high-volume data transmission make cloud computing infeasible for edge applications.
Applications at the edge are proliferating rapidly. Monitoring and tracking of assets, people, and processes span industries and verticals. Similarly, imaging infrastructure’s safety and security is also a significant use case as businesses seek to automate expensive and error-prone human security processes. Marketing stakeholders utilize the edge to engage customers in new ways and understand their buying behavior more intimately. Industrial businesses apply edge computing to drive factory automation and improve personnel productivity. As companies look to the future, AI-based optimization is a crucial application, with edge infrastructure playing a pivotal role in algorithm inferencing. All these applications need enough compute power at the edge to succeed.
The limitations of cloud-based computing are increasingly relevant as business and user value shift more towards the edge of the network. Edge computing overcomes these challenges by bringing computing power right to the edge and facilitating data collection, analysis, and real-time insights. IoT devices, camera infrastructure, industrial process control devices, and retail store applications are just a few examples of a broad landscape of edge functionality expected to unlock new business opportunities and revenue models.
Edge Environments Influence Infrastructure Choices
In considering infrastructure for the edge, it is important to recognize that the edge is not a homogenous environment. For example, it would be sub-optimal to use the same edge infrastructure in retail banking and in-process plants as the two environments are very different. Retail banks have temperature-controlled dedicated spaces to house their IT infrastructure, while industrial settings are often corrosive with little temperature control. Some edge nodes operate off a battery in other settings, as wired power is cost-prohibitive, while others do not have persistent connectivity. Edge infrastructure can be indoors in a controlled environment or outdoors exposed to the elements depending on the application. Stakeholders need to account for their environment, the specific application, and the compute required to make cost-effective choices.
Many infrastructure options are available to decision makers. Ruggedized servers, GPU accelerators for AI, storage devices, and varying combinations of computing options are all possibilities. Deployment teams can now protect the edge infrastructure from environmental factors and ensure that the necessary compute power is available for their application in a form factor that suits the space constraints typical of network edge locations.
Arrow and Dell Technologies OEM Solutions have compiled a comprehensive approach to edge infrastructure selection, including a full accounting of the various edge environments, typical edge use cases, and infrastructure considerations. Download the ‘Resilience at the Edge’ eBook to learn about edge solutions for your application.
