Service Group Capacity Determination via Utilization Curve Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Service groups of cable modems often face bandwidth overload when multiple modems actively download data, exceeding the maximum bandwidth and potentially leading to customer dissatisfaction.
Innovation Solution
A method is implemented to objectively determine if a service group has sufficient capacity by fitting a smooth curve through historical service group utilization values and calculating mean residual values, ultimately determining an aggregate probability value that indicates the likelihood of cable modems obtaining their maximum allocated bandwidth.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple cable modems actively download data simultaneously in a service group, then the data transfer capacity increases, but the service group bandwidth is overwhelmed and exceeds maximum bandwidth
Solution Approach 1:
The system performs preliminary actions by continuously monitoring and analyzing historical service group utilization data before overload occurs. It fits smooth curves to historical data and calculates mean residual values to establish baseline utilization patterns, enabling early detection of capacity issues before the service group becomes overwhelmed
Solution Approach 2:
The system implements feedback mechanisms by calculating aggregate probability values that compare current utilization patterns against historical baselines. When the aggregate probability value exceeds thresholds, the system sends alerts to network operations centers, providing continuous feedback about service group capacity status and enabling proactive capacity management
2Reliability
If service group capacity is increased to handle more cable modems, then bandwidth availability improves, but hardware changes become more costly and time-consuming
Solution Approach 1:
The system performs preliminary analysis by fitting smooth curves to historical utilization data and calculating mean residual values to establish baseline patterns. This preliminary action enables the system to predict capacity issues before they occur, allowing planners to schedule hardware upgrades during off-peak times and minimize service disruption
Solution Approach 2:
The system implements dynamic capacity management by continuously monitoring service group utilization and calculating aggregate probability values that reflect current capacity status. This dynamic approach allows the system to optimize existing hardware utilization before hardware changes are needed, delaying costly upgrades by maximizing the use of available capacity
3Measurement precision
If service group utilization is monitored continuously, then overutilization is detected early, but system complexity and computational requirements increase
Solution Approach 1:
The system performs preliminary data processing by fitting smooth curves to historical utilization data and calculating mean residual values to establish baseline patterns. This preliminary action simplifies subsequent monitoring by providing reference baselines against which current utilization can be compared, reducing the complexity of real-time analysis
Solution Approach 2:
The system transforms raw utilization data into meaningful parameters by calculating aggregate probability values that represent the likelihood of capacity issues. This parameter transformation simplifies monitoring complexity by converting complex multi-dimensional utilization data into a single interpretable metric that can be compared against thresholds
Data Source
AI summary
A computing system fits a long-term utilization line through aggregate utilization values that correspond to ones of a plurality of intervals over a time period. The computing system determines, for each respective interval of/ordinal intervals in each second time period within the time period, a mean residual value based on a difference between a utilization value identified on the long-term utilization line at the respective interval and an actual aggregate utilization value at a same ordinal interval in each second time period. An aggregate probability value, based in part on a probability that a cable modem of a plurality of cable modems can obtain at one or more intervals a maximum bandwidth allocated to the cable modem, is determined, and in response to determining that the aggregate probability value is a non-preferred value, an alert, indicating that the service group is overutilized, is sent.