With the finalization of initial 5G standards in 3GPP during Q1 2018, 5G deployments are starting to gain momentum globally, with a few initial commercial launches during 2018 For Cable MSO’s, 5G adds new revenue opportunities, in terms of extending Mobile Broadband service to own users, and Fixed Wireless Access where there are challenges with DOCSIS/Fiber deployment, as well as IoT use cases.
One of the key attributes of 5G is the significant reduction in latency. Unlike traditional 4G LTE systems, latency requirements in 5G vary with use cases. As an example, for a traditional smartphone web browsing service, 15-20 ms round trip times may be acceptable. However, for a use case such as autonomous driving, round trip latency requirements need to be under 10 ms. Other use cases that require sub-10 ms round trip latency include industrial robotics control, drone control, web gaming, and connected, collaborative multi-site live concerts (band members playing a song across multiple locations).
Typically, low latency demands tend to be localized with communication over a short distance. In order to fulfill the low latency requirements for such use cases, new architectures need to be implemented in the network. These include implementing control functionality and local switching at the edge, which in a DOCSIS network can even be a hub site. Micro servers that can support virtual applications will need to be deployed at hub sites or cell sites and in close proximity to users. In addition, while network slicing can support multiple use cases from each radio site and can fulfill use case specific routing, bandwidth and latency requirements will need to be deployed across the networks. Present DOCSIS networks are typically designed for a median latency requirement of ~10-15 ms, which can continue to work well for traditional Mobile Broadband use cases. Also, where network slicing with Edge Servers are deployed, the current cable infrastructure may be able address the 5G requirements.
5G also introduces a Virtual RAN architecture, where Layer 3 (higher layer) RAN functionality is centralized in the cloud. The one-way latency objective between the 5G radio site and the VRAN node is typically 5 ms. To fulfill such an objective, it becomes important to maximize fiber and optical switching in the access transport network. Layer 3 ethernet switching, which can add significant delays, can be deployed between the VRAN and the Core.
The 5G scheduler is hungry, which means that it will try to get the data it receives as soon as possible to the target user. For mmWave, the scheduler has a transmit time interval (window) of ~250 micro seconds which is extremely time sensitive. The faster the data can be transferred from Core to the radio, the faster it can be forwarded to the users.
Eliminating latency bottlenecks in the transport network will be key towards maximizing the overall throughput experience of 5G networks.