Orquestração escalável da coleta de dados de telemetria In-band em planos de dados programáveis

Abstract

In-band network telemetry (INT) is an emerging network monitoring paradigm. By collecting low-level telemetry items in real time, INT can substantially enhance network-wide visibility - allowing, for example, timely detection problems such as micro-burst. Recent studies have focused on (i) developing INT mechanisms to increase network-wide visibility; and (ii) to design new monitoring solutions. However, little has been done to coordinate the process of collecting telemetry items in this new paradigm. This is particularly challenging because depending on which network telemetry items are collected, it might degrade network-wide visibility in terms of consistency/freshness. Previous researches have focused on optimally orchestrating the collection of in-band network telemetry statistics from the network, providing little or none scalability to be applied in realistic environments. In this work, we (i) theoretically formalize the In-band Network Telemetry Orchestration Plan Problem and propose a machine learning based orchestration model and (ii) tackled the scalability limitation of existing approaches by proposing an iterated local search procedure. Results show that our machine learning based orchestration model approach outperforms state-of-the-art heuristics by up a factor of 8x with respect to the number of network anomalies identified, for instance. At least, our algorithm outper-forms state-of-the-art orchestration solutions by a factor of 2 with respect to the number monitoring applications requirements satisfied, while keeping the solution close to the optimum (less than 5%) – demanding a few seconds to execute. Key-words: Network Monitoring. In-band Network Telemetry. Telemetry Orchestration Plan Problem.