IntSight

Abstract

Performance requirements for many of today's high-performance networks are expressed as service-level objectives (SLOs), i.e., precise guarantees, typically on latency and bandwidth, that a user can expect from the network. For network operators, monitoring their own SLO compliance, and quickly diagnosing any violations, is a critical element for effective operations. Unfortunately, existing network architectures are not engineered for this purpose; there is no mechanism, for example, for the operator to monitor the 95th percentile latency experienced by a customer. Data plane programmability has made per-packet measurements possible but brings the challenge of keeping the monitoring overhead low and practical. In this paper, we present IntSight, a system for highly accurate and fine-grained detection and diagnosis of SLO violations. The main contribution of IntSight is, building upon in-band telemetry, introducing path-wise computation of network metrics and selective generation of reports. We show the effectiveness of IntSight by way of two use cases. Our evaluation using real networks also shows that IntSight generates up to two orders of magnitude less monitoring traffic than state-of-the-art approaches. Furthermore, its processing and memory requirements are low and therefore compatible with currently existing programmable platforms.

HH-IPG: Leveraging Inter-Packet Gap Metrics in P4 Hardware for Heavy Hitter Detection

“…QoS-HH. After HH detection, different traffic management and control actions are typically carried at different time scales, e.g., DCTCP [28], DDoS attack [29], IntSight [30], H-UPF [31]. To analyze the proposed approach, we evaluate the Flow Completion Time of the pipeline configured to send all IPG detected HHs to lower priority queues.…”

Advancing Network Monitoring and Operation with In-band Network Telemetry and Data Plane Programmability

“…One of our major achievements was receiving the Best Paper Award for our paper on INTO [Marques and Gaspary 2018] at the Brazilian Symposium on Computer Networks and Distributed Systems (SBRC). Another achievement was getting IntSight [Marques et al 2020] published at ACM International Conference on Emerging Networking Experiments and Technologies (CoNEXT), considering its rigorous review process and small acceptance rate. There, we also earned the ACM reproducibility badges due to our artifacts being not only publicly available, but also functional and reusable.…”

Desenvolvendo o Código da Internet do Futuro

“…Exemplos de metadados que podem ser coletados em cada switch são o ID do switch, o ID da porta de entrada/saída, o carimbo de data/hora, a contagem de bytes, a contagem de descartes e a utilização do enlace, bem como um ID de fila, ocupação e estado de congestionamento. Diversos desafios vêm sendo abordados pelo Grupo de Pesquisa em Redes de Computadores no âmbito de INT [Marques et al 2019, Marques et al 2020, Vassoler et al 2023. A seguir, a título de exemplo motivador, descreve-se um desses trabalhos.…”

“…Capitalizando sobre as oportunidades habilitadas por INT, projetou-se e desenvolveu-se IntSight [Marques et al 2020], um sistema para detectar e diagnosticar violações de SLO, com foco em atraso fim-a-fim e largura de banda. Em contraste com INT clássico, Int-Sight usa cabeçalhos de telemetria como espaços de trabalho, onde os dispositivos de encaminhamento calculam metadados path-wise (por exemplo, IDs de caminho, pontos de contenção e atrasos fim-a-fim) progressivamente.…”

“…Em seguida, N x cria e insere um cabeçalho de telemetria no pacote contendo campos como EEDelay (Etapa 2b). A lista completa de campos presentes nos cabeçalhos de telemetria é apresentada em [Marques et al 2020].…”