Edge Computing is one of the top areas of IT investments, especially in the light of companywide digital transformation initiatives such as the internet of things (IoT). Computing technologies that enable the delivery and analysis of data and resources to people and things in a timely fashion are collectively referred to as Edge Computing. Edge Computing is the moving of compute to where the data is generated, and therefore aids in reducing the "time to value" i.e., the insight gleaned from it.
At first glance, Edge Computing appears very homogeneous, and includes all activities that are performed outside of the "Core", which could be the location of the primary corporatewide IT (Information Technologies) infrastructure. In fact, Edge Computing is a multitiered mix of assets arranged in a use-case and workload centric fashion. An "intelligent" Edge Tier is a crucial link between the Core and Endpoints, that provides a distributed compute, data persistence and network aggregation layer, and serves as the intermediary analytics of collected data. For OT (Operational Technologies) use cases the Edge tier hosts actuation and control equipment while in CT (Communications Technologies) use cases it can even interface with communications equipment such as booster and relay equipment. The lack of industrywide standards and architectural approaches for Edge Computing means that today most of the Edge deployments are highly custom in nature (though the introduction of architectural overlay concepts such as "Fog Computing" may eventually lead to the ratification of proposed standards).
CIOs and other IT leaders must prepare their wider organization and key stakeholders for the necessary steps to successfully deploy and manage an Edge Computing infrastructure. Firms that acknowledge Edge Computing as a long-term investment generally fall into these two categories:
- Firms that have embraced Edge Computing early on. They have invested in custom or semi-custom approaches and are well on their way to realizing the business benefits of moving compute closer to data. Many of the early adopters of the Internet of Things fall in this category or
- Firms that are still on the fence with Edge Computing. They are evaluating the way in which they can deploy Edge Computing in production, the changes do they have to make their IT processes, and how they can reap long-term business benefits. These firms will generally take an approach that involves an industry-standard infrastructure.
(Note: this classification does not include "traditional" Edge deployments like remote and branch offices. Rather it examines Edge deployments for newer use cases like the Internet of Things, and by Telecom Service Providers, Oil and Energy providers, and Retail merchants.)
For firms that fall in the second camp, I recommend the following areas of diligence – areas that will significantly improve value that organizations obtain from their Edge Computing infrastructure. Avoiding any of these practices could result in wasted spend on an incomplete and suboptimal solution, or lost revenue and additional costs to address crises that arise from an incomplete or improperly implemented infrastructure.
- Infrastructure Architecture involves the selection of an appropriate computing platform, along with connectivity and data persistence. The platform must support bare-metal and virtualized workloads. Organizations must treat the Edge as an extension of their Core from an architecture perspective.
- Asset and Application Management requires control of asset sprawl, and a catalogued deployment and management IT, OT and CT apps. Deploy a software-defined infrastructure solution that treats the Edge like a cloud
- Data Management and Governance involves the lifecycle management of data generated or collected at the Edge. Ephemeral data has to be analyzed and then disposed of, and what is persistent has to be protected. Organizations must define a Core-Edge-Endpoint data management paradigm that ensures that data is managed according to its value.
- Infrastructure Security requires managing device, user, application and data security. It requires that Edge devices be managed according to corporate governance, risk and compliance policies. Here, organizations can take a multi-pronged “always-on” approach to security
Organizations must evaluate infrastructure platforms that are designed from the ground up for Edge deployments, and:
- Support running analytics (e.g. predictive asset maintenance) SCADA, historian and HMI applications that are used in OT deployments
- Have high availability and virtualization built-in, making it cost-effective to deploy bare-metal, virtualized and containerized applications.
- Are self-managing, self-protecting and can be deployed with minimal or no IT support. These capabilities help reduce unplanned downtime and ensure availability of business-critical industrial applications.
- Are ruggedized for true edge deployments where it can be placed without rack mounts, in embedded or remote locations with less forgiving power and cooling options.
- Come with remote management – alleviating customers from performing ongoing maintenance – which can be performed by the hardware vendor or a managed service provider.
Edge Computing forms an important tier in an organization's next-gen infrastructure. It can offer enormous benefits to a firm when deployed in the right context, for the right use case, with the appropriate architecture and with the right investments in technologies. Firms that could benefit from Edge Computing or have already invested in it should refer to the following IDC Perspective as a guide in making the most of their future-ready infrastructure.