Standards must come together, with latency as the cornerstone

Standards must come together, with latency as the cornerstone

Craig Thomas, Vice President Strategic Marketing and Business Development at Broadband Forum, and Greg Mirsky, PEAT Project Stream Lead in the ATA Work Area at Broadband Forum from Ericsson

Standards Development Organizations (SDOs), including Broadband Forum, IETF, and IEEE are concentrating on specifications for measuring metrics of network performance. Broadband Forum’s Broadband Quality Experience Delivered (Broadband QED), uses Quality Attenuation methods  to look beyond conventional measurements, such as Quality of Service (QoS). This initiative helps optimize networks and deliver greater Quality of Experience (QoE) to the end-user by analyzing the performance of different network segments including latency, consistency, predictability, and reliability.

Leveraging the latency tool

Latency measurement and attenuation is a great tool and one corner of the jigsaw, but it has to be part of a solution that we can regulate and automate. The likes of automated intelligence and management and cloud technologies can allow operators to steer sessions dynamically because those measurements and tools are telling you there is a problem per application. It is a useful tool to identify network problems, so that operators can find a solution that brings a return on investment.

At our Network X demo, we showcased automation with latency measurements, and how the operator can dynamically perform Subscriber Session Steering. The solution probes the network, recognizes that there is a problem with the network performance between the subscriber premises and any location in the network, and that latency has fallen down in certain parameters. By moving an individual session, not just the connection, to a different Broadband Network Gateway, the latency tools and measurements can be utilized in an automated solution for a more responsive network.

Latency has to be applicable to the operator’s cloud network, the first mile, the access network, and just as importantly, inside the home all the way to the end-user device. Unless you are bringing in tools driven by SDN/NFV, service containerization, and in the future, a move to Metro-Edge Computing, then operators will identify problems without an automated solution. That is why standards have to come together, with latency as the cornerstone.

Differentiating existing offerings

Service providers are coming to us at the Broadband Forum and saying that they want to talk about containerization. Not just to be more competitive with other service providers, but to differentiate their services to their existing customer base. We are continually identifying niche markets to tap into. This includes gaming, and asking the customer whether they are willing to spend an extra two, three, or four dollars on gaming latency measurement? From an SME point of view, are they willing to pay for an Industrial IoT application?

The access network is very quickly becoming multi-service. It is applicable to consumer and Industrial IoT, enterprise, mobile xHaul, and residential. As we look at next-generation 10G, 25G, and 50G networks, every service provider we talk to is looking at using the access network for multiple customers, not just residential. This increases the relevancy and demand for network slicing and inevitably appropriate latency as well as security and agility per network slice.

Standardization delivers interoperability

The importance of standardization is manifold. We write the dictionary so that we can, as designers and implementers, better understand and educate the industry. We produce specifications that enable developers to bring interoperable solutions to market and help operators avoid vendor lock-in. Interoperability is one of the main benefits of standardization.

Both Broadband Forum and ITF have a concentrated effort on developing data models, primarily YANG data models, for controlling the measurement of latency and exporting measurement results. When operators’ networks collect measurements, the systems from different vendors, need to speak a unified language. Standardized data models can ensure that data results can be predicted and used to carry out network diagnostics.

An agreed global standard

A big question facing the industry when it comes to latency is net neutrality and how service providers can monetize latency without running afoul of some regions’ net neutrality requirements. This is where we will see a continued move to Metro Edge Computing as a way to offer different latency packages. We have a responsibility to work with government bodies and regulators to let them know what multi-vendor solutions there are out there that can help them monitor and measure the wider network requirements of the home environment for service providers.

It is paramount that we have an agreed global standard for latency, and a consensus is agreed upon between multiple SDOs. The liaison between SDOs working together ensures that we do not cause confusion in the marketplace by having differing standards. This will create a clearer path for market deployments and adoption by regulators. We must cooperate and work closely together. If there is one universal standard, it is easier to adopt.

Additionally, to help serve providers monetize a differentiated application appropriate latency service, that same standards work needs to align with subscriber awareness. This is a much wider task that will need concerted work with the application creators and providers, consumer device innovators and the press.

Listen to Craig Thomas and Greg Mirsky’s panel discussion during the Understanding Latency Webinar by Broadband Forum member, Domos here.