Forecasting bandwidth utilization growth
By Tabin Dharanikota, Sudheer Dharanikota, Rajesh Abbi and Dennis Edens
December 31,2018
What is the Problem?
90% of Telecom operator investment goes into access network upgrades. The key factor that determines the planning of this access network is the customer bandwidth utilization growth.
We have been using historical utilization growth to project the future access investment. Many times, this process proves to be incorrect that leads to sub optimal investment and resource forecasting.
This paper talks about a methodology on how to predict growth.
Key Takeaways
Growth prediction, albeit is a difficult problem, we created a three-step framework for analysis in this work. Namely understand the demand growth impact due to,
- Demographics of a zip code
- Application adoption of demographics, and
- Concurrency of applications
Taking these into account we provide a mechanism that improves accuracy of bandwidth utilization growth forecasting.
In the current white paper, we present only the first part of the three-part puzzle. Will address the remaining in the later white papers.
Executive Summary
Predicting bandwidth utilization growth, which is key factor in determining Telecom operators 90% of the yearly investment, has been at a minimum a projection of the historical information if not a complete coin toss. Such forecasting is too general causing inaccurate spend in different markets.
Improving the growth forecast by geographic location not only improves spend effectiveness, it also helps prevent customers from having a poor experience caused by high levels of congestion when node actions are not taken in a timely manner. This new approach involves understanding application bandwidth requirements, concurrency, as well as, the demographics that drive the first two.
If your company is open to improving their network bandwidth utilization forecasts by physical location, read on to get Duke Tech Solutions analytical approach.
