CHASM: A High Level Analysis of Temporal's New State Machine SDK
summary
CHASM is a new Temporal programming model that exposes reusable, active application state machines (called cats) — queues, semaphores, maps, bloom filters and more — with the same identity/lifetime model as workflows and interoperable references. Promises include selectable durability, higher throughput primitives (notably a schedule ported to CHASM with measurable speedups), composable collections of workflows, and planned rate-limiting cats. A web crawler example reduces an ops-heavy queue+Kubernetes design to a three-activity workflow (fetch, parse, store) and works with fewer moving parts.
AI-written summary · about
chapters (18)
- 0:00Introduction and speaker background
- 1:08Talk scope and audience focus
- 2:32Limitations of current workflows
- 7:05Birth of Chasm and core principles
- 8:10Chasm architecture and cats
- 11:24Promises and capabilities of Chasm
- 15:20Identity and interoperability in Chasm
- 17:27Cooperation and collections in Chasm
- 18:34Existing Chasm components
- 20:23Rate limiting and future cats
- 21:45Chasm as the foundation of EOS
- 24:16EOS vision and ecosystem
- 26:01State machines in real-world modeling
- 28:08Design principles for Chasm systems
- 30:55Business modeling with Chasm
- 33:24Logical vs physical design
- 33:46Web crawler demo simulation
- 37:57AI guidance and summary
related talks
29mQ: A Sound Verification Framework for Statecharts and Their Implementations
40mAn Interdisciplinary View on Prototypes - Towards Executable Models That Can Be...
46mState Machines (The Good, The Bad and The Ugly)
21mDDD and FSM: tackling complexity with state machines
36mState Machine in Elixir with Machinery
15mAbstract Interpretation of Temporal Safety Effects of Higher Order Programs
29mCascade: a Meta-Language for Change, Cause and Effect
41mWhat If We Implemented a State Machine to Help the Little Mouse?