Improving the Reliability of Cable Broadband Networks via Proactive Network Maintenance

TL;DR

This work tackles the reliability problem in cable broadband networks by introducing CableMon, a public-domain system that learns network fault criteria from PNM time-series data guided by customer trouble tickets and uses unsupervised clustering to distinguish network faults from subscriber-premise faults. The approach mitigates noise in PNM data through time-series features and derives fault thresholds via maximizing a ticketing-rate ratio, while clustering devices to categorize fault types. Evaluation on eight months of ISP data shows CableMon achieves superior ticket-level detection and accurate fault-type discrimination (Rand Index ≈ 0.91) compared with public-domain tools and supervised baselines. The findings demonstrate that coupling ticket-derived supervision with PNM analytics enables effective proactive maintenance with reduced dispatch errors and have potential applicability to other networks such as cellular and Wi-Fi.

Abstract

Cable broadband networks are one of the few "last-mile" broadband technologies widely available in the U.S. Unfortunately, they have poor reliability after decades of deployment. The cable industry proposed a framework called Proactive Network Maintenance (PNM) to diagnose the cable networks. However, there is little public knowledge or systematic study on how to use these data to detect and localize cable network problems. Existing tools in the public domain have prohibitive high false-positive rates. In this paper, we propose CableMon, the first public-domain system that applies machine learning techniques to PNM data to improve the reliability of cable broadband networks. CableMon tackles two key challenges faced by cable ISPs: accurately detecting failures, and distinguishing whether a failure occurs within a network or at a subscriber's premise. CableMon uses statistical models to generate features from time series data and uses customer trouble tickets as hints to infer abnormal/failure thresholds for these generated features. Further, CableMon employs an unsupervised learning model to group cable devices sharing similar anomalous patterns and effectively identify impairments that occur inside a cable network and impairments occur at a subscriber's premise, as these two different faults require different types of technical personnel to repair them. We use eight months of PNM data and customer trouble tickets from an ISP and experimental deployment to evaluate CableMon's performance. Our evaluation results show that CableMon can effectively detect and distinguish failures from PNM data and outperforms existing public-domain tools.

Figures (11)

• Figure 1: An overview of the Hybrid Fiber Coaxial (HFC) architecture.
• Figure 2: This figure shows how the customer ticketing rate varies with the values of SNR. Ticketing rate tends to increase when SNR values are low.
• Figure 3: Figure (a) shows how the transmission powers of several cable devices in the same fiber optical node fluctuate over time. Orange dots are devices that show the same anomalous transmission power patterns. Green triangles are devices that show normal patterns. Red squares are devices that show distinct anomalous patterns. Figure (b) shows the locations of the cable devices using the same colored icons.
• Figure 4: Analysis of ticketing rate ratio.
• Figure 5: This figure explains the sliding window algorithm. When the number of abnormal points within a sliding window exceeds a threshold, the window is considered to be abnormal. An abnormal event is given by merging the abnormal windows.