What to know about planning mobile edge computing (MEC) systems

Mobile edge computing (MEC) is a network architecture that supports compute, analytics and storage capacity at the edge of the network, and proponents say it provides substantial performance benefits for applications requiring low latency, especially IoT applications.

But MEC deployment is complicated by the lack of mature standards and the sheer number of standards and architectural options. Each IoT deployment will have unique requirements for latency, performance, the amount of date to be sent, the frequency at which it is transmitted, and cost.

IT leaders should design MEC-connectivity architectures with flexibility and adaptability in mind.

The focus of this article is on business mobile-edge computing and its relationship to IoT and does not address MEC technologies that enhance 5G RAN architectures implemented by mobile operators. (Note: Due to the popular use cases such as IoT, MEC is used as a generic term to encompass almost all edge-computing architectures.)

The impact of IoT on the enterprise

The need for MEC stems in large part to the rise of the internet of things, with millions of devices and sensors now connected to the Internet. IoT is an architecture for all  types of devices and sensors to connect to edge, centralized or cloud-based data centers.

Insights derived from IoT data help organizations improve their operational efficiency and provide improved services to their customers. IT organizations report challenges providing low-latency connectivity, and managing and securing a large number of IoT devices.

The case for mobile edge computing

IT systems have become increasing centralized in the era of public and private cloud.  The vast majority of processing, analytics and storage capacity in enterprise organizations resides in a few centralized data centers or in the public cloud. For this data to be analyzed and acted upon, traffic must flow to and from end devices and the data center. Proponents of MEC cite significant benefits for building out compute and storage capacity much closer to where the data is created – at the edge of the network.

MEC applications are typically driven by the need for very low latency. Round-trip  times between devices and data centers can be up to a second or more. MEC architectures can deliver predictable millisecond latency, which may be critical for manufacturing, health or public safety applications.

The sheer number of IoT devices and/or the volume of data they generate can create significant challenges for designing the network. MEC provides for real-time data analysis and vastly reduces the amount and frequency of the data required to be sent to a distant centralized location.

MEC architectures can provide additional benefits depending on application requirements:

• High availability – improve the redundancy and reliability of the application
• Security – by keeping sensitive data in local locations and not exposing it to the Internet
• Lower bandwidth costs – by reducing the amount of data sent over the wide area network
• Location awareness – for applications like logistics and shipping

Building a MEC system

Architecting an application-specific system for edge computing calls for a complex combination of hardware (servers), software and networking. Work load or vertically specific software is customized to meet the specific needs of the MEC application, for instancee tracking shipping containers. The server, based in central and edge locations as necessary, runs the custom software that translates large amounts of IoT data into actionable information.

A specific network, such as Wi-Fi or 4G, is designed to connect the devices and deliver the actionable analytics with appropriate latency, reliability and cost. Open APIs are generally available on both software and networking systems to customize traffic flows.  As a result, systems-integration skills may be required to deliver MEC systems with the required specifications.

The cost and complexity of deploying enterprise IoT systems have slowed early development of MEC systems.  Every new project requires highly specialized software and integrated custom networking to deliver on project goals across such diverse applications as industrial control, smart cities, health services, public safety and energy management. So projecting the expected return on investment remains challenging for most enterprise MEC projects.

Too many standards and architectural choices

More than a dozen standards organizations are involved with setting the architecture for MEC including ETSI, OpenFog, EdgeX, and OpenStack StarlingX. A similar diversity has complicated and stalled SDN and NFV architectural standards for more than five years. Because of the diversity of IoT applications and the range of their specific and custom needs, it may be difficult to apply the emerging MEC standards horizontally across industries. IT organizations currently evaluating MEC deployments must select from a wide array of hardware, software, and networking options and figure out how to tightly integrate them into hardened production systems.

Recommendations for IT leaders

MEC offers potential benefits for IoT and other edge applications including predictable low latency, lower WAN costs, improved reliability and security. Since each MEC application has significant differences and with no clear blueprint for deployment, MEC architectures should be designed with flexibility and adaptability to meet changing business requirements. IT organizations should move cautiously with initial edge computing deployments involving pilot projects with a clear business case.