The key trend over the last decade was the steady migration to public cloud infrastructure for compute and data storage. As the traditional role of the CIO evolved, they were pressured to deliver more and more economical IT solutions that met or exceeded the expectations of their customers and public cloud provided a simple means to an end. For most enterprises, cloud is a critical element to competitiveness however, there is a growing category of time-sensitive applications and workloads that – to provide the best possible experience – need to be pushed deeper into the network. These applications include Streaming Video, VR and AR, Artificial Intelligence, Industrial IoT and Critical Control System Management and they are already gaining traction in the new “Machine Economy”. However, these emerging workloads require compute models that are largely incompatible with today’s centralized public cloud paradigm.
New Demands on Global Network Architecture
A key constraint is latency. Real-Time Ad Inventory Auction and High Frequency Trading require latencies of less than two milliseconds. Autonomous Vehicle and Drone Control also require network latencies of around 2 milliseconds with maximum processing and compute times of less than eight milliseconds. Haptic VR and AR applications require combined network and compute latencies of less than 25 milliseconds. Put in the context of the human response, even world champion athletes have reaction times greater that 120 milliseconds; this is the first time in history where the network is faster than human cognitive response
While low latency is imperative, there are other fundamental requirements that need to be addressed. By the end of the next decade, ubiquitous tera-scale connectivity for 125 billion humans and machines will be required, with each connection providing cognitive operations powered by Artificial Intelligence and Machine Learning. The outcome will be highly bespoke, personalized interactions that need to be secured through contextual trust, authentication and compliant privacy management. To support these new models, a radical shift in network architecture is required.
5G Enables Accelerated Adoption Edge Computing
While High Speed Broadband is the obvious choice for these workloads today, 5G Wireless Access will provide the foundation to build and support this new functionality on Mobile Networks. 5G is a purpose-built technology designed and engineered to facilitate connected devices and automation systems. It provides “seemingly” infinite network capacity and massive throughput with imperceptible latencies. It provides tera scale device support, managing up to 1mm connections per square kilometre, and can pinpoint device with <1m accuracy to drive cognitive location-based applications. It also delivers a single network for multiple industries and use cases. Individual network ‘slices’ can be configured for different needs, thus supplying the guaranteed QoS required for critical control systems and haptics.
Edge Cloud is integral to these deployment models. This edge natively supports distributed real-time processing and optimized execution environments that enable multi-access and multi-cloud capabilities. It unlocks compute resources for application innovation and monetization. It also facilitates cognitive operations, powering new contextual services driven by Artificial Intelligence and Machine Learning that orchestrate not only downstream IoT and critical control systems, but the actual edge fabric its-self.
For businesses, it significantly increases redundancy and availability. By performing computation at the edge, any possible disruption is limited to a single point in the network instead of the entire ecosystem as is the case with a centralized cloud infrastructure. It provides much lower compute and connectivity costs; applications only send significant information instead of raw streams of sensor data and send it only short distances. This provides enormous network, compute resource and actionable data storage optimizations.
This radical paradigm swing combined with the growing requirement to support time-sensitive, contextual AI driven applications and adaptable, automated digital trust is driving an inevitable shift to the edge.
To support this new market, “Global-Local” alliances are already forming. A massive re-distribution of cloud infrastructure to the edge is underway, providing “single-hop” compute connectivity with massive scale access for both humans and machines. Our Edge Cloud platform built on the EDGE GRAVITY Unified Delivery Network (UDN) exemplifies this approach – over 100 Global Operators and numerous application partners are already scaling and monetizing the new Machine Economy for a myriad of Global Brands.
Public Cloud and Edge Computing are Complimentary
While some believe that edge compute can replace the public cloud altogether, the two approaches are relatively complementary and provide several important strategic synergies. Centralized cloud computing remains one of the cheapest and most effective ways to store or archive large data sets or process non time-sensitive information. For operations that require low latency or near real-time data processing such as device control and orchestration, edge compute in a distributed cloud will always provide the preeminent solution. Nevertheless, when used in conjunction centralized cloud and edge computing architectures can provide more effective workflows, allowing businesses to store and process actionable data in the network tier that provides the greatest advantage.
Edge Compute also greatly complements on-board device compute capabilities. While a small amount of integral compute functionality will always be required, primarily for regulatory compliance, Edge Compute provides the ability to seamlessly offload processing and storage. This has several key advantages. It eliminates low footprint memory and CPU constraints, massively increases data storage and provides flexible orchestrated workloads based on individual use cases. By providing prolonged battery life it also enables manufacturers to deliver thermally efficient, sleek designs.
On February 24th, 2019, EDGE GRAVITY will hold its fourth annual Global Edge Forum. Comprised of key cloud technology and business executives, these global leaders are collectively implementing the “edge” vision and defining the next generation of edge technology trends and experiences. Tera-scale connectivity, low-latency and human cognitive operation at the edge will be the key to creating tomorrow’s services with a single common goal: solving the digital need for new business and human value creation.