Publications

Accurate and Efficient Monitoring for Virtualized SDN in Clouds

IEEE Transactions on Cloud Computing, JCR 2020 IF 5.938 Top 9.545%, vol. 11, no. 1, pp. 229-246, 1 Jan.-March 2023, doi: 10.1109/TCC.2021.3089225.

Gyeongsik Yang, Yeonho Yoo, Minkoo Kang, Heesang Jin, Chuck Yoo

Download PDF

Abstract

This article presents V-Sight, a network monitoring framework for programmable virtual networks in clouds. Network virtualization based on software-defined networking (SDN-NV) in clouds makes it possible to realize programmable virtual networks; consequently, this technology offers many benefits to cloud services for tenants. However, to the best of our knowledge, network monitoring, which is a prerequisite for managing and optimizing virtual networks, has not been investigated in the context of SDN-NV systems. As the first framework for network monitoring in SDN-NV, we identify three challenges: non-isolated and inaccurate statistics, high monitoring delay, and excessive control channel consumption for gathering statistics. To address these challenges, V-Sight introduces three key mechanisms: 1) statistics virtualization for isolated statistics, 2) transmission disaggregation for reduced transmission delay, and 3) pCollector aggregation for efficient control channel consumption. The evaluation results reveal that V-Sight successfully provides accurate and isolated statistics while reducing the monitoring delay and control channel consumption in orders of magnitude. We also show that V-Sight can achieve a data plane throughput close to that of non-virtualized SDN.

Paper summary

Give each virtual network its own timely statistics

A tenant needs measurements of its own network, even when physical resources are shared.

Physical switch counters can mix traffic from several virtual networks. V-Sight translates those counters into tenant-specific statistics while addressing the delay and control traffic caused by collecting them.

Translate statistics and coordinate collection

Statistics virtualization uses the virtual-to-physical resource map to calculate the requested counters. Transmission disaggregation prepares physical statistics in a cache before requests arrive. Collector aggregation combines overlapping collection work and adjusts its timing.

Physical counters flow through shared collectors and a cache before being translated into tenant statistics.View full-size figure
Physical counters flow through shared collectors and a cache before being translated into tenant statistics.

Check accuracy, delay and collection cost

The journal study evaluates statistics accuracy, collection delay and control-channel consumption, as well as cache time skew and CPU and memory use. It also examines effects on tenant traffic, reporting monitoring behavior and data-plane throughput close to the non-virtualized SDN comparison.

Three mechanisms connect tenant-specific counters to shared collection work.
MechanismWhat it changes
Statistics virtualizationTranslate physical counters into tenant statistics
Transmission disaggregationPrepare counters in a cache before requests arrive
Collector aggregationCombine overlapping collection and adjust its timing

Key takeaways

  1. Tenant isolation extends to observability. The virtualization mapping translates physical counters into meaningful tenant statistics.
  2. Independent monitoring requests can share physical work. Reusing collection reduces redundant traffic while keeping the returned statistics tenant-specific.