Gaming Isn’t Homework: Predicting Demand at the Neighborhood Level (2022)

By Christopher Stump, May Tan, Adam Gruchevsky; Comcast

Efficient network planning requires accurate forecasting of customer behaviors in the near and long term. The goal is to anticipate the location and magnitude of capacity augments months, and even years, into the future using forecasted monthly bits per second demand (Mbps). This prevents outages, increases reliability and prepares the network to handle faster speeds. However, no longer does a blanket growth rate apply to every neighborhood. Network architecture is transitioning to run fiber deeper into the network and make Service Groups (SGs) smaller, decreasing households passed while increasing speeds. This adds further challenges as smaller SGs are more sensitive to individual household patterns and faster speeds create more usage spikes. As an example, single releases of popular games downloading to a console are enough to create network congestion as are increasingly frequent live sports streaming events.

In the existing analog node world, an office park could hang off the same SG as a housing complex. The difference between day versus night peak, and weekday versus weekend, are washed away due to the size of the SG and augment decisions could be made by the macro trend to consume more traffic at faster speeds. Now, in the digital node world, SGs are smaller. An office park may have a dedicated SG and a residential community may have its own dedicated SG. If we sum their traffic at the headend level, we will certainly see the macro trend, but now the individual SGs have their own driving forces. The office park is driven by weekday, daytime peaks with seasonal dips for vacations and holidays, not to mention pandemics. The residential component is driven by nighttime and weekend peaks that spike heavily in work-from-home situations or when popular games have new releases (and homework lay forgotten). In fact, popular gaming releases can more than double the number of SGs that peak on a single day. With the popularity of gaming and the increase of sporting events dedicated to streaming, these spikes are anticipated to become more frequent. As such, the trends of these SGs cannot be accurately predicted using a headend or national level CAGR (Compound Annual Growth Rate

By clicking the "Download Paper" button, you are agreeing to our terms and conditions.

Similar Papers

Winning the Gaming War: Play for Cable Operator
By Alon Bernstein, Rajiv Asati & Sangeeta Ramakrishnan, Cisco Systems
Optimizing Wi-Fi Channel Selection in a Dense Neighborhood
By Yonatan Vaizman, Hongcheng Wang; Comcast
Why Gaming Needs An Edge
By Alan Evans, EDGE GRAVITY by Ericsson
Access Capacity Planning: Staying Well Ahead of Customer Demand Helped Ensure Stability During COVID-19
By Bruce E. Barker Jr., Claude Bou Abboud & Erik Neeld, Comcast Cable
Photon Avatars in the Comcast Cosmos: An End-to-End View of Comcast Core, Metro and Access Networks
By Venk Mutalik, Steve Ruppa, Fred Bartholf, Bob Gaydos, Steve Surdam, Amarildo Vieira, Dan Rice; Comcast
Right Technician at the Right Time: Using Machine Learning to Predict Network Maintenance Issues
By Anastasia Vishnyakova, Rama Mahajanam, Mike O’Dell, May Merkle-Tan, Catherine Hay & Lisa Pham, Comcast Cable
Planned Maintenance Tool (PMT): A Data-Driven Approach to Recommending the Best Time for Planned-Maintenance
By Pete Quesada, Julianne Heinzmann, Nishesh Shukla, Resmi Vijayan, Mike O’Dell, May Merkle-Tan; Comcast Cable
Training Machines to Learn From Signal Meter Readings
By Gary Ventriglia, Jack Birnbaum, Robert Gonsalves, Anastasia Vishnyakova, Michael Kreisel & Larry Wolcott, Comcast Corporation
The Dennis Botman Story: A Tale of Next-Level Chatops
By Michael Winslow, Ryan Emerle & Mia Kuang, Comcast Corporation
Comcast Underground: Innovative Fiber Deployments Over Existing Underground Critical Infrastructure
By Venk Mutalik, Pat Wike, Doug Combs, Alan Gardiner, Dan Rice; Comcast
More Results >>