With 5G-based applications, communication is no longer one-to-one. New communication patterns are driven by app downloads, software updates, sending advertising to digital signage, and connected cars – and those are just a few examples. More frequent use of video on mobile devices will increase the need for broadcast and multicast technologies to alleviate network bandwidth demands. IoT and other mission-critical apps will use many-to-one data collection in the upstream direction to report data or video from multiple monitoring points to one central location.
End users, whether that be people or things, may need to access multiple applications or content domains located in different places, simultaneously. Similarly, individual users may upload information to a cloud application that is accessed by many other users. In these instances, it’s clear that applications require rapidly changing any-to-any connectivity – in other words, dynamic interconnectivity. But can our transport networks deliver?
New application requirements
LTE-Advanced (LTE-A) is being rolled out and 5G is entering the early deployment phase. We can categorize lead applications based on these mobile technologies in three groups:
- Evolved mobile broadband (eMBB)
- Ultra-reliable, low-latency communication (URLLC)
- Massive machine-type communication (mMTC)
Each of these application types has specific network requirements.
For eMBB, we can expect traffic volume to exponentially increase due to increased demand for high-definition video, AR and VR, and real-time gaming. We know users expect faster, higher-quality experiences while simultaneously improving visual, audio and potentially even tactile interfaces. Over-the-top (OTT) application providers like Netflix already use distributed content architectures to place content caches closer to users, and use varying methods to reduce cost by taking advantage of off-peak capacity to shift large volumes of data around the network.
However, it isn’t just content that needs to be stored closer to users – applications like autonomous vehicles and smart factories require low latency, ultra-reliability and security, which means network operators must place cloud-based computing resources for application processing closer to end users to support these requirements. Multi-access Edge Computing (MEC) is increasingly being used by operators to deliver unparalleled experiences and allow software applications to benefit from low latency.
To accommodate the growth of mMTC from the Internet of Things (IoT), mobile operators will deploy small cells in much higher numbers to increase coverage and bandwidth. However, operators still face the challenge of connecting these small cells into the transport network and providing the necessary capacity, synchronization and security.
Mobile architecture changes
In 4G, we saw the advent of centralized RAN with optical fronthaul used between remote radio heads (RRHs) and baseband units (BBUs). 5G takes centralization further. The 3GPP has defined a number of possible functional splits between distribution units (DUs) and centralized units (CUs), with the goal of providing the ideal combination of latency, throughput and cost effectiveness for a variety of potential applications.
Virtualizing the RAN functions and moving from dedicated radio access platforms to cloud-based systems is the next step in the process. With these systems, operators gain the flexibility to run radio functions in large centralized data centers, smaller distributed sites, or a combination of both, which allows them to efficiently manage all components as a single cloud.
This RAN evolution is occurring at the same time as mobile core functions are being distributed. For example, packet core functions are moving from centralized locations over to local aggregation sites, which are closer to end users to address the need for high bandwidth with low latency. The 3GPP has defined control and user plane separation (CUPS) to provide the flexibility to support centralized or distributed control and user plane functions with independent scaling, and to facilitate network slicing. These cloud-based RAN and core functions are dynamic in nature and can increase or decrease in scale as needed. As a result, traffic volumes will also dynamically scale up and down.
In summary, as we move toward 5G, RAN and mobile core functionalities are being split and located in different parts of the network, resulting in data plane and control plane traffic being dynamically routed to these various processing points – and this routing must be flexible in order to make appropriate real-time adjustments.
The transport network of the future
Local content and computing, any-to-any connectivity, and distributed virtual network functions require dynamic interconnectivity as traffic patterns and volumes change. To accommodate these requirements at scale, operators will need IP routing that encompasses IP, MPLS and segment routing protocols across the transport network – right out to cell sites and business demarcation points. Upgrades for multigigabit cell site connectivity are required. The access and aggregation layer requires high GE and 10GE port density and 100GE throughput to accommodate increasing numbers of connectivity points and traffic volume. It requires fast re-convergence to keep up with changing network implementations and provide resiliency. A uniform, end-to-end IP routing environment provides all this plus operational simplicity, resiliency, faster problem resolution, and easier service implementation, reporting and management.
As 5G-era applications change traffic patterns and volume on transport networks, high-capacity IP routing is a key underpinning for the dynamic interconnectivity required between centralized and distributed content and computing that makes our mobile communications and applications run flawlessly.
Ellen Warren has worked in telecommunications for over 30 years in product management and marketing roles. She has experience with a wide range of technologies including ISDN, satellite, fiber optics, and is now working in the IP Routing group at Nokia Ottawa. After the telecom meltdown in 2002, Ellen co-founded a startup that developed a lane departure warning system for long-haul trucking.