By Hugh Taylor, CEO of Edge Site Partners
Edge computing seems poised for significant growth, given the right market conditions. Major players from multiple industries have a strategic interest in making edge computing a success and have been busy elaborating on plans for investment. Commercial use cases are developing, but have not yet matured. When they do, it is likely that success will require seamless interoperation between multiple, independent edge networks owned by separate corporate entities. For this to work, there will have to be standards for managing edge data center locations across multiple provider networks — what we’re calling Edge Location Management and Operations Standards (ELMOS).
Hypothesis and drivers of need
The excellent work being done by the LF Edge Akraino project and others is establishing the basis for pervasive edge computing at the operating system, device and software layers. Existing open standards projects and commercial platforms for edge computing do not offer much for the management of multiple edge data center locations, however. Few, if any, standards or commercial platforms are emerging to handle what could be the hundreds of thousands of physical edge data centers needed to support pervasive edge computing use cases.
Additionally, few if any, emerging standards or commercial platforms appear able to handle edge computing deployments that span multiple edge network providers. It is a virtual certainty that edge applications will need to deploy across multiple edge networks, however. For example, no single telecom provider can provision an application that supports assisted driving or autonomous vehicles along America’s four million road miles. To work, that application would have to be deployed on multiple telecom service provider edge networks. The infrastructure to support the application could easily encompass half a million micro-edge data centers.
Consider the infrastructure management requirements that come with such a deployment:
- Provisioning applications and data on a location-centric basis. There is a need to identify available edge data center rack space, each within one-quarter mile of 100,000 specific locations. (If an existing edge data center is not present, we need to contact property owners and explore the idea of building one on their land and/or acquiring the property for that purpose.)
- Facilitating transactions to deploy across multiple edge networks. There is a need to sub-lease rack space (or KWs of data center capacity) in 100,000 micro edge data centers owned and operated by five different telecom and tower providers — including requesting the sub-lease, procuring the capacity, negotiating a service contract(s), arranging for physical support providers on a local basis, and handling payments.
- Enabling inter-operation between edge applications hosted at micro data centers owned by different corporate entities.
- Monitoring hardware and responding to outages or performance problems on a location-by-location basis; for example, a server running the EVE operating system at the micro data center at 41.495722 latitude and -81.519856 Longitude is down — and repeat this use case across 100,000 sites.
- Managing infrastructure on a location-by-basis, including updating the EVE operating system at a handful of locations among 100,000 sites.
- Responding to physical issues on a location-by-location basis; for example, rolling a truck to investigate vandalism at that location or summoning a utility to reconnect a power line.
- Provisioning location-sensitive failover instances, e.g., we need an instance within one-quarter mile of a location in order to maintain 1-millisecond latency.
At this time, it does not appear that existing standards projects or commercial platforms either have or are developing such location-centric infrastructure management capabilities. Using manual methods to manage such massively distributed and multi-supplier infrastructure is prohibitively expensive and slow.
A proposed standard
Key edge computing stakeholders would benefit from coming together to define and develop standards for the inter-operation and management of edge data center locations. The goal should be to develop standards that can handle the exchange of data required for location-centric edge infrastructure management and multi-edge computing, including:
- The location of an edge data center.
- The available infrastructure elements (compute, storage) at an edge data center.
- An edge data center’s capacity (e.g., kilowatts, rack space, etc.).
- Commercial terms of sub-leasing of edge data center capacity.
- An edge data center’s owner and operator.
- Local utility and physical support contact information for an edge data center.
- Billing and commercial transaction data for procuring the use of an edge data center.
- Requirements needed to deploy software to an edge data center site, based on location.
- Data to support monitoring and management of edge infrastructure, based on location.
It’s time for diverse players to come together
Edge Site Partners wants to bring key industry stakeholders together to develop ELMOS. Participants in the ELMOS process might include telcos, tower companies, colo providers, data center REITs, and many others. Any business that wants to inter-operate and transact business related to edge computing resources should find the proposed ELMOS to be of interest.
Edge Site Partners has published a comment draft of an ELMOS white paper. Feedback is welcome.
About the author
Hugh Taylor is CEO of Edge Site Partners, a startup focused on solving the real estate challenges inherent in edge computing.
DISCLAIMER: Guest posts are submitted content. The views expressed in this post are that of the author, and don’t necessarily reflect the views of Edge Industry Review (EdgeIR.com).
edge data center | Edge Site Partners | facilities management | micro data center | observability | orchestration | standards