Explore the world of microservices scheduling and security in this 45-minute conference talk from GOTO London 2016. Dive into the intricacies of cluster scheduling with Nomad, a powerful tool developed by HashiCorp. Learn about job specification, decoupling benefits, and various workload types supported by Nomad. Discover how to implement service discovery, distribute secrets securely, and handle stateful applications. Gain insights into Nomad's architecture, inspired by Google's Borg and Omega systems, and understand its optimistic scheduling approach. Explore the challenges of scaling microservices and learn from the experiences of building Console, a large-scale cluster management tool. Understand the importance of schedulers in modern infrastructure and how Nomad addresses the complexities of deploying and managing microservices at scale.
Scheduling & Securing Microservices with Nomad
Overview
Syllabus
Introduction
Tools
Nomad
What is a scheduler
What are schedulers used for
Why do we use schedulers
Decoupling benefits
Better quality of service
Cluster schedulers
Nomad Cluster Scheduler
Job Specification
Job Decoupling
Workload Types
Job Files
Service Discovery Index
Registering Services
Secret Distribution
Security Challenges
Vault
Login Flow
Declarative Block
Vault Token
Compliance disks
Stateful applications
Hard applications
Nomad hints
Besteffort replication
Making deployments easy
Nomad is a single binary
Surf is a cluster management tool
Surf has been proven at very large scale
Learnings from Console
HighLevel Operational Goals
Building for Scale
Building with Surf Console
Research
Borg
Architecture
Isolation
Regions
Google Omega
Nomad Architecture
System Architecture
Optimistic Scheduling
Nomad Million Container Challenge
Failures
Why would you need a scheduler
Citadel
What is Nomad
Taught by
GOTO Conferences
