The Economic Benefits of Distributed Broadband Services

There are several key market trends that are driving the need for new distributed, disaggregated, and cloud-based broadband network architectures. Traffic is growing rapidly because of 4K/8K video, cloud gaming, work at home, and other new applications. High-bandwidth, low-latency applications and services are driving the requirement for service intelligence in the metro access network. New CUPS architectures are disaggregating the user plane and the control plane in both mobile and fixed broadband networks, and we expect to see increasing convergence of mobile and fixed IP networks.

‍
Although legacy centralized BNG architectures have served many service providers well for decades, there are a few challenges with this legacy approach in existing networks:

• BNG ports provide rich features combining aggregation and service functions and are relatively more expensive

• BNG expenses rise with increasing port density requirements

• Complex feature sets combining BNG features and aggregation functions in legacy BNGs can require considerably more coordination and planning for upgrades

‍
Traditionally, BNGs use 1:1 redundancy for high-availability implementations. A new approach to fixed broadband BNG service uses a spine-leaf design similar to today’s cloud data center approaches. The spine-leaf design consists of access leafs and service leafs that disaggregate access node functional requirements from BNG service functions. This architecture supports distributing edge nodes or service leafs to the access network providing BNG services for high-bandwidth and/or low-latency intensive applications such as video caching, IoT gateways, AR/VR servers. Likewise, this spine-leaf design can be implemented in centralized or regional data centers with similar advantages.

‍
This paper compares the legacy centralized BNG architecture with the spine-leaf design including:

• Identification of some challenges with legacy centralized BNG architectures

• Details of the new spine-leaf design

• Assumptions and results of the total cost of ownership (TCO) model

In our analysis we compare the legacy centralized BNG architecture with a distributed spine-leaf BNG architecture and show a five-year TCO savings of 35%. We also compare scenarios where both the legacy BNG and the spine-leaf BNG designs are deployed in central office data centers to compare their TCO. These approaches produced the most significant five-year cumulative TCO savings of 64%.