How Our SREs Safeguard Nanosecond Performance—at Scale—in an Environment Built to Fail
summary
Optiver describes a low-latency, on-prem trading stack that moves 230 TB/day across ~1,500 servers to connect to 60+ exchanges using FPGAs and fiber/microwave links. Control over changes is explicit: human-reviewed deployments, an assisted-deployment tool, a Housekeeper validator app, and physical and virtual kill switches. The firm runs ~14,000 releases and facilitates ~35,000 changes/year across 600 binaries and 10,000+ customized components, favoring minimal complexity and a “fail hard” stop-on-error policy.
AI-written summary · about
chapters (8)
- 0:00Intro
- 1:24Agenda
- 1:50How Trading Works
- 3:35The Trading Environment
- 6:05Night Capital
- 8:47SRE Principles
- 9:10SRE Control
- 10:59Automation
related talks
35mSRE -- culture and strategy @ TrustingSocial
11mBuilding a system that never stops: New Relic at scale
21mHigh Network Reliability & Availability in FE & BE for Scalable Training Solutions
46mChaos Experiments - Datacenter Stress Testing
31mScience Reliability Engineering for High Performance Computing
56mGetting all the 99.99(9) you always wanted
43mFrom "Keeping the Lights On" to "Designing the LEDs": A Detailed Review...
32mReliability in Finance: Delivering Quality Market Data at Scale