Glenn Katz, Chief Commercial Officer, Telesat. Senior executive with deep expertise in complex networks and innovative service delivery.
Modern enterprises run on data. Consequently, the volume of real-time data they are generating is exceeding the latency and uninterrupted connectivity requirements of traditional centralized data center and cloud models. To counteract bottlenecks, many businesses are now exploring edge computing as a means to process large amounts of time-sensitive data more efficiently.
Large-scale edge computing places compute and storage resources close to where data is generated, such as by various devices, including the rapidly expanding Internet of Things (IoT). This is especially useful in enterprise applications.
Consider a large hospital campus where all workers and most equipment are connected to edge devices. Data generated on such a large scale needs to be coordinated with its applications. Instead of that data going to a cloud environment and back to the applications that need it, the hospital can use a large-scale edge architecture to process it locally for on-campus requirements and sharing with other organizations like emergency responders and insurance companies. While such architectures are more complex, they can ensure near real-time access to potentially life-saving information.
Shifting To The Collaborative Edge
Still, everything cannot be done through a single-edge server. Applications will often need to talk to other edge computing devices, and data will go back and forth to the master cloud (or wherever the master application is held). To do that, organizations need a massive amount of low-latency, high-speed connectivity to edge computing devices, referred to as the collaborative edge. This could be done as a backup to fiber, but such solutions can be cost-prohibitive as already vast amounts of data continue to grow.
Low Earth orbit (LEO) satellite constellations help enable a mesh network that backs up operations to guarantee resiliency or expand operations, delivering carrier ethernet-level performance at the lowest latency and highest speed.
For decades, satellites have been used to transmit data over long distances. The growth of large and powerful LEO constellations helps position the satellite industry to enable the low-latency transmission and resiliency that large-scale edge computing environments require. Such connectivity can support collaborative edge operations anywhere and will be especially useful for industries with remote operations that may struggle with current coverage limitations.
For instance, the maritime sector can’t rely on best-effort internet connectivity for connecting on- and offshore edge data to and from ships at sea for future autonomous shipping applications. Many autonomous commercial ships sailing the oceans will generate huge and costly data requirements while underway. Although some of that data can be managed locally, each ship will require a land-based digital twin replica for the network operations center (NOC). This digital twin is needed for technicians and others to keep tabs on operations and remotely investigate any error indicators that arise. These connections must be low latency to minimize the delay of transferring data like real-time video, which may be crucial to safety and operational integrity.
Agriculture also relies on the broad application of IoT, with millions of devices across enormous tracts of land generating volumes of data on water usage, nutrition density and other environmental factors affecting crop growth. However, getting adequate internet coverage in rural locations is challenging and expensive. As more farms implement autonomous IoT, applications like on-the-ground video surveillance and tracking the status of autonomous farm vehicles in motion can expand high data rate implementations.
What’s Needed For Collaborative Edge Networking
These use cases require low-latency connectivity with dedicated bandwidth and strict SLAs based on the compute edge application requirements. When consumer or best-effort broadband isn’t sufficient, what options exist for remote compute edge architectures?
If all locations can be reached via terrestrial networks, building a collaborative compute edge network is feasible. The best approach is to share detailed network requirements with multiple telecom providers to get estimates on cost and feasibility. Key requirements should include minimum guaranteed data rate (CIR), maximum latency, maximum jitter, maximum packet loss and guaranteed availability. All of them should come with service level agreements (SLAs).
Evaluate whether the provider’s terrestrial network can support the necessary topologies:
• Star Network (where remote locations connect to a central edge hub).
• Mesh Network (where remote locations communicate with each other).
• Hybrid Network (a combination of both star and mesh networks).
These capabilities are typically available through Carrier Ethernet services, which usually adhere to MEF Forum specifications. Avoid providers that don’t comply, as their networks may lack interoperability with others.
If Carrier Ethernet isn’t viable, LEO satellite networks might be an option. While LEO networks are still developing, some currently offer reliable layer 3 broadband, though they are not yet suited for collaborative compute edge implementations. However, in the coming years, more LEO networks will emerge with greater capability.
When planning, focus on companies offering MEF-compliant layer 2 Carrier Ethernet services or similar, along with the necessary user terminals and interconnect requirements. The LEO network provider should also adhere to industry standards to ensure compatibility with terrestrial telecom networks and support star, mesh or hybrid topologies.
Two Technology Shifts Transforming The Industry’s Options
Gartner predicts that by 2025, 75% of enterprise-generated data will be created outside of centralized data centers. That means far more low-latency bandwidth is required fast. With today’s terrestrial Carrier Ethernet capabilities and tomorrow’s LEO networks, building collaborative compute edge networks in any location can help meet that need.
Moore’s Law dictates that the amount of potential onboard satellite processing power will continue to grow exponentially as technologies become ever more miniaturized. That creates a new avenue for quickly and economically bringing needed bandwidth anywhere in the world.
Within the next decade, it’s not unrealistic that compute edge-type capabilities could actually run applications on an LEO satellite and provide direct-to-satellite backhaul. As enterprises chart their strategic direction for the next decade and more, the collaborative edge combined with powerful new LEO constellations can help imagine entirely new connected environments, wherever they need to operate.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
