The Growing Prospects for Edge Data Centres

Opportunities will only increase in this exciting sector

The tech industry loves to promote the next big trend, offering a window into opportunities and concerns on the horizon. One of the latest technology advances under the spotlight is the edge data centre. Although the value this provides isn’t in question, for many companies, the journey to adoption is far from fixed.

Edge is the trending prefix to computing and storage capacity for processing data closer to the source rather than in a centralized data centre. This is especially beneficial for application workloads where latency is critical, as in the case of fast-acting decision-control facilities used in manufacturing.

Beyond reducing latency, edge computing can use advanced analytical processing served by artificial intelligence constructs such as machine learning. This allows for more sophisticated control and predictive insight closer to the point of application where it offers the greatest value. Consequently, edge data centres serve as small outposts located at the edge of the network, with lower power usage, providing edge computing processing capability.

Edge Drivers Are Plenty

Operational drivers place growing focus on more-distributed processing, storage and network management. This is because of the breadth of IT and communication advances that enable the rapid digitalization of industries and extend the reach and capacity for digital operations.

For example, several new capabilities benefit smart manufacturing, healthcare and retail solutions. They draw from connected products and systems defined by the concept of the Internet of things which reside at and extend the edge of the operational network.

5G networks offer significantly lower latency, faster speeds and greater capacity for high-bandwidth solutions at the network edge. Advances in network function virtualization and software-defined networking support greater levels of software-driven operations, flexibility and connectivity. Such technologies play an important role in digitizing the management of critical infrastructure that serves industrial edge operations.

Collectively, these network-enhancing technologies form the bedrock for many new computing workloads and the generation of vast amounts of data. Building in support for cloud-based operations opens the way for edge workloads that require great levels of reliability, efficiency, autonomy and data processing. And with machine learning analytics, companies can better achieve the fine level of control needed in time-sensitive edge operations.

The Scope for Commercially Viable Uses Is Still Being Explored

There are some clear uses that highlight potential commercial opportunities. Sophisticated machine learning processing at the edge can support computer-based operations including using computer vision to validate the quality of manufactured goods or inspect the condition of physical fixtures. Other examples are in smart city support services, which see a range of connected things communicating at close quarters using ultralow latency.

Such implementations can lead to support for safety-critical functions that enhance road safety and enable real-time traffic management and alerts for collision warnings. Autonomous vehicles have real-time needs that are best served by edge operations, and video analytics solutions demand high-bandwidth edge processing support for cameras deployed in different locations.

The world of industrial operations, which employs dedicated on-premises proprietary systems and private networks, presents a wealth of further opportunities, some of which are starting to be realized.

The fourth industrial revolution centres on the cost benefits, flexibility and deep-level data processing offered by cloud-based operations, advanced analytics and network advances such as 5G. New processing capabilities from leading equipment and chip manufacturers such as Intel can aid time-sensitive network operations, allowing for more cost-efficient edge processing of time-sensitive workloads. These opportunities can be achieved by replacing current latency-focussed proprietary edge control systems and enabling them to be supported by general-purpose processors and IT-grade high-level systems such as Linux or Windows.

Transformation of on-premises proprietary process control systems that provide operational control in continuous industrial processing systems could prove cost-effective. Further benefit could see greater application of finely tuned time-sensitive incremental control for advanced optimization deployed to a broader range of edge operations.

The Edge Is Another Dimension, Navigating It Won’t Be Easy

Those involved in maintaining edge processing operations benefit from specialist knowledge and experience; they understand the complexities and implementation concerns of the technologies needed to support edge operations and allow its value to be achieved.

What’s often missing, though, is clear articulation and mapping of the different solutions, supporting network communication strategies and partner relationships. This can make it very hard for non-specialist roles to understand the operational needs, teams and technology solutions that need to be in place.

Edge operations require user organizations and suppliers to think beyond infrastructure and architectural needs. New automation and orchestration challenges will arise, often across transactional boundaries and occurring between different companies and industries rather than just different parts of the network. They must also think about ownership of the software and infrastructure stack and the likely path of service engagement — be that through a telecom operator, hyperscale public cloud provider or others.

Providers of edge operational services also need to decide how they support multiple customers according to their individual needs. This will be especially necessary for applying operational-specific artificial intelligence algorithms, and may result in multi-layered partner offerings.

Why Is It So Darn Complicated?

Getting real at the edge means considering a swathe of attributes and strategies that address analytical intelligence, resilience, security, visibility, performance, maintenance and failover.

Securing the edge. Being at the edge of the network means that security, both physical and digital, becomes vital as avenues of attack and remoteness increase susceptibility to interference.

Operational resilience. Resilience and fault tolerance with the expectation for automatic failover to another edge computer or data centre are crucial. Such operations are already addressed in several critical-services industries, but they rely on complex architectural management strategies with a high degree of automation. Consequently, the level of plumbing and maintenance to keep operations effective can be challenging. And as with all things concerning infrastructure systems, the ability to manage edge locations through a unified control plane remains paramount for efficient hybrid operations.

Collaboration rights. Distributed architectures underpin many business services and form the basis for future deployments such as autonomous vehicles. But such complex environments require the right set-up, with strong collaboration between software and systems engineers, application developers and infrastructure operators.

All this will require organizations to think more carefully about how they extend their data centre operations to enable greater levels of edge processing, work with cloud providers or hook into another provider’s edge data centre network.

If You’re Not Netflix, Is Edge Processing for You?

The biggest drivers for edge data centres are coming from industry sectors where edge operations are already well-established. Gaming, video streaming services, branch offices and more already have a good command of edge operations. They run complex and highly distributed content delivery networks with stringent latency constraints to meet the expectations of a global audience. Netflix, a well-known customer of Amazon Web Services (AWS), and its streaming peers use cloud computing and delivery services. However, many others use dedicated proprietary systems and hardware appliances.

Therefore, deciding whether any edge-based business operation needs to meet such strict latency service-level agreements will provide insight into the complexity underpinning edge data centres.

Be Prepared for Edge Data Centres

The latest foray into edge computing operations and data centres is still relatively nascent. However, the underlying mechanics, particularly network communication strategies and architectural models from telecom service providers and hyperscale cloud platforms, are starting to take form.

Telecom operators are sharpening their portfolios of mobile edge services. Some of these are being served through collaborations with leading public cloud providers, which add cloud-based functionalities to the latest connectivity advances brought by 5G networks.

At CCS Insight, we’ve reported on the growing band of telecom operators deploying mobile edge services with AWS Wavelength, which embeds AWS computing and storage at the edge of an operator’s 4G and 5G networks. Our research highlights four network operators — KDDI, SK Telekom, Verizon and Vodafone — with mobile edge services based on the solution.

The biggest challenges will be navigating the maze of suppliers and service partners that support edge computing and edge data centre operations, and understanding ownership responsibilities in service delivery engagement and management.

Many organizations are unlikely to manage their own edge data centres and will look to tap into a network of providers. Their choice of services will range from failover support to performance service-level agreements with various dashboards providing management, deployment and visibility. Industry regulators will need to consider multiple paradigms of operations and communications along with ownership responsibilities.

The edge data centre market is fizzing with potential, as the number of edge computing solutions grows, offering a range of capabilities and delivery services. But it remains early days, and the market will continue to be shaped by topics such as the sustainability of power consumption and distribution as this becomes increasingly important.

A version of this article was also published on Computer Weekly; you can access that here.