Everything You Wanted to Know About High Availability (But Were Afraid to Ask)
We all want our applications and databases online at all times, but that’s much easier said than done. In order to determine which tools will best accomplish that, we need to know what High Availability actually means. Consensus, Quorum, Split Brain, CAP Theory, Failover, CDN, Database-backed Caching, API queues, and other terms get thrown around liberally. But what are the underlying situations that spawned these concepts? How can we combine them to keep our application stack online, even if several parts of it are malfunctioning or simply broken? For those who are interested, this talk can help clear everything up!
Exposure to management of a major stack resource (app, data, cache, infrastructure, etc.) is a great start, since it gives some insight into what it takes to keep it online, or the role it may play. Many of the concepts we'll discuss are universal and apply to topology considerations in general and can be applied in many contexts.
- What each layer of a HA stack contributes
- Fundamental HA terminology
- Architecture distribution strategies
- Necessary design compromises
- Application interactivity under duress