Bandwidth and Latency Optimization

Address Your Bandwidth Exhaustion Challenges

• Kubernetes Networking Limitations

  • Kubernetes does not provide any native traffic control capabilities.
  • Traffic Rate-Limiting is essential to provide optimal resource consumption and to avoid bandwidth exhaustion.
  • Kubernetes’ native Bandwidth Rate-Limiting is still experimental and has detrimental effects to latency.
  • Additionally, users may have poor experience when connecting to external-facing Kubernetes clusters as most TCP congestion protocols are not designed to work well in today’s diverse networks

• Cilium Bandwidth Management

  • Cilium’s Bandwidth Manager can enforce rate-limiting per Pod with a single line of YAML.
  • Bandwidth Manager provides a 4x reduction in latency compared to alternatives.
  • Cilium is the first Cloud Native networking platform to support BBR (Bottleneck Bandwidth and Round-trip propagation time), a congestion control algorithm designed by Google.
  • When Google tested BBR, they observed up to 2,700x improvement in throughput.

Optimal Bandwidth Management

• Unlike alternatives, Cilium’s rate limiting isn’t detrimental to performances and is designed for multi-queue and multi-core NICs.
• Cilium’s BBR provides exceptional improvements in throughput and latency for external-facing applications, offering consumers a fantastic user experience.
• Extremely simple to configure with a single line of configuration

Intuitive and simple integration

A single line of YAML can enforce traffic limits on your Pod.

Demo

In this short video, Senior Technical Marketing Engineer Nico Vibert walks you through how to use Cilium Bandwidth Manager to rate-limit the traffic sent by your Kubernetes Pods. Great to address potential contention issues!

• Pod Traffic Rate Limiting with Cilium Bandwidth Manager youtube

Better Bandwidth Management with eBPF

Kubernetes provides many knobs for managing common system resources such as vCPUs and memory limits per Pod, but often forgotten is the effect of unbounded network communication in a cluster. A large churn of packets from several services can starve bandwidth for other services.

In this KubeCon EU 2022 talk, Daniel Borkmann & Christopher M. Luciano explore how eBPF can be leveraged to dynamically insert logic for flexible, efficient and scalable rate limiting and bandwidth management on a per-Pod basis.

• Better Bandwidth Management with eBPF - Daniel Borkmann & Christopher M. Luciano, Isovalent Youtube