ETSI’s MEC plans were signified with a name change and expanded focus to include hetnets using LTE, 5G, fixed broadband and Wi-Fi technologies.
Highlighting the rapidly expanding scale and scope of mobile edge computing on the greater telecommunications space, the European Telecommunications Standards Institute renamed its Mobile Edge Computer Industry Specifications Group to Multi-access Edge Computing. The newly named organization will continue to use the MEC abbreviation.
In addition to the name change, ETSI said the group’s focus will expand to “address multiple MEC hosts being deployed in many different networks, owned by various operators and running edge applications in a collaborative manner.” This expansion is set to include work on heterogeneous networks using LTE, “5G” technologies, fixed deployments and Wi-Fi technologies, with current work to simplify application programming interfaces, standards-based interfaces for multi-access hosts and an alignment with network functions virtualization architectures.
ETSI explained the move to “embrace the challenges in the second phase of work and better reflect non-cellular operators’ requirements.”
“In phase two, we are expanding our horizons and addressing challenges associated with the multiplicity of hosts and stakeholders,” explained Alex Reznik, who was recently elected chairman of the MEC ISG. “The goal is to enable a complete multi-access edge computing system able to address the wide range of use cases which require edge computing, including [internet of things]. We will continue to work closely with [Third Generation Partnership Project], ETSI NFV ISG and other SDOs as well as key industry organizations to ensure that edge computing applications can be developed to a standardized, broadly adopted platform.”
The MEC system is said to provide a standardized and open system to support diverse virtualization development and deployment models, with the ability “for an application to discover applications and services available on other hosts, and to direct requests and data to one or more hosts.”
ETSI last year released a trio of “foundation-level” group specifications through its MEC ISG, which it claimed defined MEC terminology, study technical requirements and use cases, and specified the framework and reference architecture of MEC.
ETSI earlier this year touted significant success from a recent NFV Plugtests event in Spain, which included NFV interoperability tests for network service on-boarding, instantiation and termination. The organization claimed 98% success for interoperability tests of network service on-boarding, instantiation and termination, but more cautious “encouraging initial results” for more complex operations like scaling and network service updates. The organization said it plans to drive the findings back into its NFV ISG.
As telecommunication networks evolve in support of higher data speeds and lower latency, operators are expected to rely more heavily on MEC architecture to support more demanding use cases.
At the recent Open Compute Project event in Santa Clara, California, Said Berrahil, VP of CIMS and telco cloud infrastructure at Nokia, explained that today’s telco cloud deployments are designed to support network services with 100 milliseconds of latency, a maximum of 10 megabits per second of throughput, and no more than 10 billion devices that cost up to $1,000 each and with a battery life of one day. For tomorrow’s network, Berrahil said the telco cloud will need to provide latency of no more than one millisecond, support network speeds in excess of 10 gigabits per second, and host up to 1 trillion devices that cost $1 each.
In order to meet the latency requirements for 5G and IoT, Berrahil said the telecom space would need to significantly bolster today’s centralized data center approach with edge data centers and mobile edge computing.
As an example, Berrahil said the 100-millisecond latency expected from today’s networks can be served by a cloud data center up to 10,000 kilometers away from the end user. However, when looking to push latency down to four milliseconds, the cloud data center needed to be no more than 30 kilometers away, while services in need of latency less than one millisecond would require MEC deployments almost within eyesight.
“Latency matters and latency changes the network,” Berrahil said.
For the telecom space to meet those requirements, Berrahil said it would likely need architecture with a centralized data center housing hundreds of OCP-based racks, which were in turn connected to edge data centers housing tens of OCP-based racks each, which were in turn connected to MEC deployments housing individual OCP-based racks.
Bored? Why not follow me on Twitter.
