Pushing to the Edge with Hybrid Cloud
Over the last few years, technology has evolved through an acceleration of innovation across industries, bringing forth new combinations of technologies, new use cases and new business models. Technologies such as the Internet of Things (IoT), cloud computing, machine learning and big data have combined to solve business challenges that plagued industries for decades.
What is a hybrid cloud?
According to IBM, a hybrid cloud is infrastructure that connects at least one public cloud and at least one private cloud, but the definition can vary.[i] A hybrid cloud provides orchestration, management, and application portability between public and private clouds to create a single, flexible, optimal cloud infrastructure for running a company’s computing workloads.
Despite the growth in public cloud computing, enterprises often need to use a combination of public and private (on-prem) clouds. Often overlooked in the hype around public cloud computing, private clouds offer greater flexibility, security and compliance.
A private cloud environment is generally accessible only through private and secure network links, rather than the public internet. Industries such as healthcare and finance have specific regulations about storing and processing data and thus favor using private clouds. A company can run a private cloud on-premises in its data center, local server room or access it as a securely hosted offering by a cloud service provider (CSP).
Crucially, hybrid cloud computing enables companies to accelerate their digital transformation efforts, primarily if they work with legacy hardware and infrastructure. They can extend their existing infrastructure by adding one or more public cloud deployments — modernising applications and processes in stages rather than a complete digital transformation upheaval.
What is computing on the edge?
IoT technology is ubiquitous, with connected devices collecting more and more information through sensors, cameras, accelerometers, LiDAR and depth sensors. All this information requires collection, storage, processing and analysis to create data-driven insights. Some of this data comes from mission-critical applications where a split-second delay can have significant consequences. For example, factories, smart traffic consoles, an insulin pump, and smoke and noxious gas monitoring.
As a consequence, edge computing use cases have grown. Edge computing places processing (and some storage) capabilities close to the data source, enabling fast data analysis in real-time. It’s particularly useful in poorly connected environments such as oil refineries, mines and wells. Companies are moving more of their compute and financial investments toward edge computing. Grand View Research predicts that companies will spend $43.4 billion on edge computing by 2027, a compound annual growth rate of 37.4%.[ii]
Despite the predictions of some analysts, this does not mean the death of cloud computing. Cloud computing and edge computing have a beneficial functional relationship. And this relationship extends the hybrid cloud concept.
According to Gartner, “Edge computing augments and expands the possibilities of today’s primarily centralised, hyperscale cloud model and supports the systemic evolution and deployment of the IoT and entirely new application types, enabling next-generation digital business applications.”[iii]
Combining hybrid cloud and edge computing
A hybrid environment with workloads at the edge and various cloud locations offers advantages to companies seeking greater efficiency and cost savings. Running business and time-critical workloads at the edge ensures low-latency and self-sufficiency. This means transactions can occur even in rugged environments where internet connections are poor.
Take the example of industrial IoT and a factory that uses sensors to monitor machines for temperature, sound, pressure and vibration. The factory can use a locally hosted compute device from a nearby cloud provider, or even something like a Raspberry Pi, to process and filter and aggregate data from the machines in near real-time. If this edge compute instance detects an urgent anomaly, then it can generate an alert for investigation. It can send the filtered and aggregated data to a public cloud instance during regular operation to perform further analysis, machine learning processing, decision making, and storage with a service that provides better efficiency and value for such tasks.
Connected cars are another example, which are effectively data centers on wheels with hundreds of in-car sensors creating a deluge of data. Autonomous driving systems, such as those tested by Equinix customer Continental, must aggregate, analyse and distribute that data, as well as data from other sources such as traffic and weather information, in real-time — with all the necessary security and privacy controls in place. And as the degree of autonomy advances (from level 1 for some driver assistance to level 5 for fully autonomous), the amount of data to aggregate and analyse will continue to soar. Current test drives for L2 autonomy are generating up to 20 terabytes (TB) of data a day, while more advanced sensor sets for higher levels of autonomy (L4 and above) may generate up to 100 TB/day.
A car needs some of this data in real-time to make split-second decisions, like whether to move lanes or whether the road is clear of pedestrians. The processing of this data could happen on the onboard computer or on any available local edge compute instances the vehicle happens to be near at the time. When the car returns to a WiFi connection, it can then upload any other less important data to a public cloud instance, receive software and machine learning model updates, a driver can review their data, or the manufacturer can download for analytical purposes.
The communication between edge computing and the rest of the hybrid cloud needn’t be in one direction. Once compute services have processed, analysed and reached decisions on the data they have, they can then push relevant updates to edge compute instances.
Are you looking to introduce a hybrid cloud solution?
Like many other aspects of modern infrastructure, containers and orchestrating them with Kubernetes can help standardise edge and cloud deployments. Kubernetes’ standard runtime layer enables you to develop, run and operate workloads consistently across computing environments and move workloads between edge and cloud.
Equinix Metal provides the foundational building blocks that give businesses the ability to create and consume interconnected infrastructure with the choice and control of physical hardware and the low overhead and developer experience of the cloud. Digital leaders use Equinix Metal to create digital advantage by activating infrastructure globally, connecting it to thousands of technology ecosystem partners, and leveraging DevOps tools to deploy, maintain and scale their applications. This means that on-demand bare metal servers with dedicated GPUs optimised for edge-type workloads such as machine learning are within your reach.
Metal integrates with a range of common hybrid cloud tooling such as Anthos, VMWare Tanzu, and RedHat OpenShift, allowing public cloud vendors and users alike to leverage any existing infrastructure and tooling.
Equinix Fabric supplements Equinix Metal by offering software-defined interconnection to connect Equinix Metal and your other infrastructure together, including all leading cloud providers. Equinix Fabric helps companies who want to take advantage of hybrid multicloud but need to reinforce privacy and security for data as it travels between edge and public cloud locations. On top of providing these security guardrails, Equinix Fabric is affordable and performant, not adding any other overheads to applications.
To learn more about how enable the hybrid cloud for your organisation today, download the Equinix Whitepaper on Enabling The Hybrid Cloud.
[i] IBM, “Hybrid Cloud,” October 19, 2019.
[ii] Grand View Research, “Edge Computing Market Worth $43.4 Billion By 2027 | CAGR: 37.4%,” March 2020.
[iii] Gartner, “2021 Strategic Roadmap for Edge Computing,” Bob Gill, 3 November 2021 – ID G00723410.
Originally published here.
The Acer ConceptD 300 desktop features up to 64 GB of DDR4 2666 MHz memory and real-time...
Digital leaders leverage "cloud precipitation" to optimise on-premises and cloud-native...
The events that transpired in year 2020 have changed the way the retail industry interacts with...