Edge data centers include hardware, software, applications, data management, connectivity, gateways, security, and advanced analytics. They come in all shapes and sizes depending on the use case: Small ones can fit into a wiring closet. Larger ones resemble traditional data centers. There are also modular, prefab edge data centers that can be set up outside a manufacturing plant or retail location to avoid expensive and time-consuming renovations to the facility creating the data.
Edge data centers serve as an extension to the organization’s existing on-premises data-center/hybrid/multicloud IT infrastructure strategy. They are an intermediary that collects, filters, and processes some types of data on site, and that sends other data that requires additional analysis back to a central data center, the cloud, or both.
[Download our editors’ PDF edge data center buyer’s guide today!]
In this buyer’s guide
- Edge data centers explained
- Why enterprises need an edge data center
- What to look for in an edge data center
- Major trends in edge data centers
- Leading edge data center vendors
- What to ask before buying edge data centers
- Essential reading
Why enterprises need an edge data center
The main driver for edge data centers, but certainly not the only one, is the internet of things (IoT). Remote sensors, security cameras, vehicle fleets, industrial and warehouse robots, smart cities’ assets, and other connected endpoints generate massive amounts of data.
Processing at the edge addresses the inefficiency and cost of transmitting raw data back and forth to a central site or the cloud. It also provides the benefits of reduced latency for time-sensitive applications and enables data processing on-site for actionable information based on real-time analytics.
In most cases, edge implementations are driven by business units with a specific issue that needs to be addressed. IDC has identified more than 400 use cases for edge computing across various industries and domains, with the largest investments in 2023 targeting production asset management, autonomic operations, omnichannel operations, freight monitoring, and augmented customer service agents.
Here are some leading-edge data center use cases in major industries:
Manufacturing: By deploying sensors, plant managers can quickly detect problems in real time (such as “this engine is overheating”) and step in to avoid disruptions to manufacturing processes. Edge data centers can analyze data over time to provide recommendations for predictive maintenance. At higher levels, edge data centers enable real-time quality assurance, optimized supply chains and automated processes. Edge data centers are an integral part of an organization’s industrial IoT (IIoT) strategy.
Healthcare: Remote patient monitoring via cameras, telemedicine, the deployment of wearable devices that gather medical information, medical imaging systems, and disease detection and prevention analytics are data-intensive use cases that benefit from the deployment of edge data centers in healthcare.
Retail: A retail edge data center might initially be deployed to support video cameras designed to detect theft. The organization might quickly discover that in-store cameras can also collect data about customer behavior. This can lead to better product placement, even new customer experience applications, such as sending an instant discount coupon to customers’ phones based on what product they’re looking at.
Energy and utilities: Edge data centers let energy and utility companies conduct remote monitoring of equipment to detect problems, avoid outages, and increase system efficiency and reliability.
Transportation: Organizations with fleets of vehicles can deploy edge data centers to provide real-time alerts to drivers and maintenance crews, and to enable predictive maintenance.
Smart cities: In smart-city implementations, edge data centers collecting camera data can provide real-time traffic management, as well as other features such as parking systems that point drivers to the nearest open parking space.
What to look for in an edge data center
An edge data center is fundamentally different than a traditional enterprise data center in these ways:
Hardware: Edge computing hardware won’t necessarily be deployed in a controlled environment. In industries such as manufacturing, oil and gas, mining, energy, and agriculture, organizations might need ruggedized servers that can handle temperatures extremes, dust, vibrations, etc.
Applications: There is no cookie-cutter way to build an edge data center; it’s totally application-dependent. For example, an edge data center designed to process high-definition security camera video has bandwidth, storage and analytics requirements far different from an agricultural deployment of IoT sensors that measure soil conditions.
Protocols: IoT devices have their own alphabet soup of data messaging protocols (AMQP, MQTT, and CoAP) and network connectivity protocols (ZigBee and LoRaWan).
Gateways: The edge is an intermediate piece of the organization’s larger IT infrastructure. Specialized edge gateways are required to keep data moving smoothly, irrespective of protocols.
Data management: Data management systems determine which data can be safely ignored, which data can be processed on site, and which should be passed along to more robust systems, either cloud or on-premises.
Security and data privacy: Edge data centers constitute another attack surface that needs to be protected. IoT sensors themselves provide an entry point for hackers, and edge data centers might be situated in locations that don’t have dedicated IT/security professionals on site. Data privacy is another concern, particularly as these connected endpoints gather and transmit, for example, patient data in healthcare or customer data in retail.
Enterprises should take a platform approach to edge data centers. Trying to build an edge data center infrastructure in a piecemeal fashion could result in compatibility issues and could lead the organization into a technological dead end.
Major trends in edge data centers
“Data gravity” refers to the dramatic shift in how and where enterprise data is generated, driven primarily by the explosion of IoT devices across virtually every industry. The basic concept is that IT infrastructure should move to where the data is, rather than the other way around. That’s the fundamental driver of edge data centers regardless of application or industry.
IDC predicts there will be 41.6 billion IoT devices by 2025, capable of generating nearly 80 zettabytes of data — nearly half of the 175 zettabytes of data generated globally. (A zettabyte is a trillion gigabytes.) Cisco Systems predicts the number of IoT devices will hit 500 billion by 2030, and the trendline indicates that the volume of edge-generated data will exceed all other sources in the near future.
Enterprise and service provider spending on hardware, software, and services for edge solutions will reach nearly $317 billion by 2026; and 75% of organizations say they intend to increase their edge spending by an average of 37% over the next two years, according to IDC.
Leading edge data center vendors
Vendors are coming at the edge data center market from all angles. The cloud hyperscalers are extending their cloud infrastructure to the enterprise edge as a managed service. The traditional IT infrastructure powerhouses have do-it-yourself and fully managed edge data center solutions. Telecom service providers are leveraging their distributed physical infrastructure to deliver edge services. And content delivery network (CDN) providers are also getting in the edge game.
Here are some leading examples (in alphabetical order):
Akamai: Akamai’s Gecko (Generalized Edge Compute) platform leverages the company’s more than 4,000 global points of presence to bring compute capacity closer to users, reducing latency for traditional use cases and enabling new ones like immersive retail, spatial computing, and industrial IoT.
Amazon Web Services: AWS Outposts delivers AWS infrastructure and services to on-premises or edge locations. With AWS Outposts, organizations can run some AWS services locally and connect to other services in the cloud.
Dell: The Dell Native Edge platform includes hardware, storage, software, gateways, and consulting services to help businesses centrally manage and securely scale their edge across multiple locations.
Google: Google’s Distributed Cloud Edge is an on-premises service whereby Google provides, deploys, and maintains the edge data center server hardware, storage, and software, with remote management by Google.
HPE: HPE’s GreenLake managed service provides an integrated platform that covers an organization’s cloud, data center, colocation, and edge deployments. The GreenLake service offers a pay-per-usage, self-service model that features governance, control, visibility, cost management, data analytics, and insight.
IBM: IBM offers hardware, software, and consulting services for the edge, including IBM Edge Application Manager and IBM CloudPak for Data, which extends applications and models to the edge for real-time insights.
Microsoft: Microsoft’s edge portfolio, under the Azure Edge Stack and Azure IoT Edge Services banner, provides ruggedized servers and a full suite of hardware, software, and management services.
Oracle: Oracle’s Roving Edge Infrastructure, a mix of computing, storage, and platform software, running in a ruggedized enclosure, enables low-latency processing closer to the point of data generation for Oracle Cloud workloads.
Verizon: Verizon has two multi-access edge (MEC) offerings. Private MEC is an onsite deployment of compute, storage with 5G RAN (radio access network) connectivity to support specific low-latency applications. Public MEC places compute and storage at the edge of the Verizon cellular network.
What to ask before buying edge data centers
Because every enterprise is different and because edge data centers are both complex and varied in their capabilities, you need to get a clear grasp on your specific needs, capabilities, and resources before engaging prospective vendors and then choosing specific solutions.
10 key questions to ask yourself before buying edge data centers
10 critical questions to ask vendors about edge data centers
Essential reading
- What is edge computing and why does it matter?
- Edge computing: When to outsource, when to DIY
- AI, 5G, and IoT spur edge data centers
- What is IoT? The internet of things explained
- Get ready for the convergence of IT and OT networking and security
Source:: Network World