Enterprises are using data to optimize existing markets and explore new ones. Data is the “heartbeat of modern user experiences and services,” according to IDC, which are being built using next-generation technologies, such as artificial intelligence (AI) and machine learning (ML). At the same time, the need for real-time data is growing, propelled by IoT. IDC predicts that nearly 30 percent of the world’s data will be in real time by 2025 – doubling the level today. This is because the data is increasingly intersecting with and driving the physical world of factories, automobiles and smart cities.
Processing such high volumes of data eats up time. It would be extremely challenging to backhaul data to and from an end point, such as a car, via a 5G and fiber network to cloud-based data centers for AI-based data analytics processing in real time. Data processing needs to take place as close to the user as possible, enabling a real-time response. This will make edge computing vital for a wide range of emerging applications, such as connected car, streaming video, augmented reality, location-based services and IoT.
What are edge, fog/core and mist computing?
Definitions of edge, fog/core and mist computing can be confusing. Edge computing is used as the generic term to refer to the movement of compute, storage and networking outside of the traditional data center. Fog or core computing encompasses both edge and mobile computing, covering mobile and wireline connectivity, storage and deep packet networking.
Meanwhile, we have mist computing, defined by the U.S. Commerce Department’s National Institute of Standards and Technology (NIST) as a “lightweight” form of fog computing that sits at the edge of the network fabric, bringing the fog computing layer much closer to “smart-end devices.” It uses microcomputers and microcontrollers to feed into fog computing nodes.
Moving towards the edge
Gartner predicts that by 2020, over 20 billion “things” will be connected to the Internet. Edge computing allows data from these things to be processed far nearer to where it is created, instead of having to send it on long-haul, round trips to data centers or clouds. This is key as it will enable organizations to harvest and analyze data in near real time, opening up markets for a host of new applications in financial services and health, for example.
Edge computing solutions can be located in various places, including on the end point device itself – for example in a car or sensors on an oil rig. An edge device has none of the characteristics of cloud. It’s not elastic or scalable, and the end user tends to have to own it and manage it, but it can process data and make simple decisions exceptionally fast.
Additions to the edge
Micro data center capabilities will also be required for some edge computing applications. Micro data centers aggregate data from gateways and embedded edge systems and bring greater decision making, security, storage and resiliency capabilities. This is critical in areas such as the safety of autonomous cars with the need to detect and react to unpredictable events in the real world, such as a child running out in the road.
An autonomous car needs to take in, correlate and analyze data from a large number of sources, some of which are sensors inside the vehicle (for example, the brakes), some come from satellites (GPS, weather data) and others from cameras that see other vehicles and pedestrians. All of this data needs to be brought together and processed so that decisions can be made instantly.
For other applications, universal customer premises equipment (uCPE) on the enterprise premise could host this edge computing functionality for an IoT platform, alongside software-defined routing, security and traffic acceleration capabilities.
So called multi-access edge computing (MEC), previously referred to as mobile edge computing, is another option. Here, the processing takes place on either a base station or a telco data center or PoPs. Indeed, IDC believes PoPs will, over time, be renamed Points of Service and be used for the delivery of AI-based edge computing applications in addition to network services. And we’re also likely to see the growth of in-city cloud computing instances in areas with high population densities to complement regional cloud data centers.
A mixture of compute, store and connect architectures will be required for most data analytics use cases. Data processing will often take place at the edge for real-time decision making, but some of that streaming data will also need to be captured in a central cloud for comparative analysis.
MEC will be rolled out as early as next year. MEC will provide ultra-low latency, higher bandwidth and real-time radio network data. This is achieved by enabling cellular operators to open their Radio Access Network (RAN) to authorized third parties, such as developers and service providers. This will enable them to offer up the speed and flexibility required to introduce innovative applications and services to enterprises. Use cases include augmented reality and video analytics, for example.
At the Roland Garros French tennis open last year, Orange provided a glimpse into MEC and video analytics. Using wireless cameras cited around the venue, together with AI applications, it demonstrated how it can keep track of individuals around various courts and entertainment areas. Providing intelligence on exactly how populated public spaces are can be invaluable for improving safety and security by avoiding over-crowding and enabling proactive guidance of visitors to a venue.
Standards are in the pipeline
Organizations such as the OpenFog Consortium and the LF Edge Initiative are developing the specifications that edge services will need to follow to deliver interoperability with service offerings to enterprises. These standards are key for micro data centers to be created at the edge of the world’s digital communications infrastructure.
5G and the edge
While we are seeing the first wave of 5G appear, it will be next year before we see enterprises and individuals hook into its power. But it does not mean that enterprises can’t start deploying edge right away. Retail, industrial, transport and warehousing are already exploring its potential, for example.
The rapid momentum of IoT means things are being rapidly connected. One of the big challenges with the 4G Low Power Wide Area (LWPA) standard is that it supports under 70,000 devices in one square kilometer. The IMT-2020 standard, which defines the technical specifications of 5G, outlines a minimum connection of one million devices for the same area, making it crucial for delivering on IoT’s promises.
5G will also come with speeds clocking as high as 10-20 Gbps, essential for near-real-time, power-hungry applications. Alongside this comes the mission to reduce latency to less than a millisecond – paramount for mission-critical applications such as autonomous vehicles.
With the enormous scope that these technologies offer, Orange and Dell Technologies have recently announced they are partnering on 5G innovation. They will collaborate to deliver real-time edge use cases and new service opportunities that 5G is offering up. This new distributed architecture for 5G will be designed to deliver the best of cloud and mobility.
IDC estimates that 15 percent of European IoT infrastructure will be using edge computing by 2020, as enterprises begin to embrace the flexible services that edge computing starts to offer. Growth will then increase, with 45 percent of global data forecast to be situated close to the network edge by 2025, according to the analyst firm.
Edge computing has the potential to transform many industries. Developing sectors, such as autonomous vehicles, will be the testbeds for the new 5G-based edge networks, which are being rolled out by telcos and their partners.
5G is still in its infancy. Edge computing providers, however, have got off to a head start, and advanced security and data protection features are being built into edge computing devices. So the stage will be set for a 5G/edge computing combo that will help to transform our smart cities, smart traffic, drone-enabled supply chains, virtual reality and other innovations far faster than we ever anticipated.
Read this article for more information on the work of Orange on 5G and connected vehicles.
Jan has been writing about technology for over 22 years for magazines and web sites, including ComputerActive, IQ magazine and Signum. She has been a business correspondent on ComputerWorld in Sydney and covered the channel for Ziff-Davis in New York.