Tuning Your Cloud: Improve Global Network Performance for Application
summary
Shows concrete steps and measured gains: an experimental setup with two peered VPCs (us-east-1 and ap-southeast-1) running C5n.9xlarge on Amazon Linux 2 uses iPerf3 to transfer 1 GB over ~220 ms RTT. Tuning Linux TCP socket buffers under TCP Cubic cuts transfer time from ~150 s to ~90 s and raises throughput from ~55 Mbps to ~90 Mbps at the cost of ~1,000 lost packets versus zero with defaults. Also: CloudFront edges reduce user RTT from ~150 ms to ~30 ms, shrinking handshake delay from ~300 ms to ~60 ms.
AI-written summary · about
chapters (12)
- 0:00Intro
- 0:39Topics Understanding Application Performance: Why TCP Matters Choosing the right Cloud Architecture
- 1:43Mice: Short Connections
- 3:10Elephants: Long Connections
- 3:52The Transmission Control Protocol (TCP): Startup
- 5:51The Transmission Control Protocol (TCP): Growth
- 7:38The Transmission Control Protocol (TCP): Recovery
- 10:23Solutions: Three things you can influence 1 Latency from your applications to your users: Service Architecture Region selection, use of edge services 2. Throughput from your infrastructure: Infrastructure Design
- 11:29Amazon CloudFront: Improving Latency to Users
- 14:55Why Packet Throughput Matters
- 16:56The Nitro System enables performance improvements
- 23:45Amazon Linux kernel TCP Tuning: 1GB Transfer
related talks
26mTuning Your Cloud: Improve Global Network Performance for Applications
31mA Practical, Research‑Backed Introduction To Cloud‑Native Kernel‑Bypass Networking
31mTuning Linux TCP for data-center networks
45mReducing Network Latency: Innovations for a Faster Internet
24mBreaking the TCP Barrier: Accelerated I/O for S3 With RDMA
16m"Performance Testing: ..."
22mAdapting TCP for Reconfigurable Datacenter Networks
34mComprehensive Performance Benchmark on Various Well-Known CNI Plugins