| • Introduction to Spotlight series on Node.js | |
| • Interview with James Snell about H2 implementation for Node.js | |
| • Discussion of current state and challenges of implementing H2 in Node.js | |
| • Debate over whether H2 should be a module or part of Node's core | |
| • Importance of keeping Node's core small and focused | |
| • Defining what belongs in Node Core vs. separate modules | |
| • Importance of keeping Node Core small and focused on web fundamentals | |
| • URL parsing as an example of a functionality that could be moved to a separate module | |
| • Discussion on H2 protocol and its differences from H1, including binary framing, stateful header compression, and implications for web application design and performance | |
| • Security as a top priority for the next year | |
| • H2 protocol will be designed to be highly spec-compliant and prioritizes security over performance | |
| • H1 was not fully compliant with the specification, which led to several security issues | |
| • H2 servers will require TLS connections by default due to browser implementation requirements | |
| • Node's reputation as a TLS terminator is a limitation that needs improvement for better performance | |
| • H2 enables new extensibility models and possibilities for new kinds of protocols | |
| • New proposals for other layered protocols are emerging within the working group | |
| • WebSockets' relationship with H1 is being considered as a precedent | |
| • The framing model in H2 will allow for experimentation without the pain of introducing WebSockets like before | |
| • Potential innovations could come out of it, but require collective experience and testing | |
| • Criticisms of the H2P protocol include: | |
| • Inefficient header compression (staple header compression) | |
| • Repetitive headers leading to waste in transmission | |
| • HPAC's stateful compression model adds complexity | |
| • Maintaining state tables can be resource-intensive | |
| • Additional complexity in H2 compared to H1 | |
| • Flow control and prioritization in H2 | |
| • Impact on proxy software vendors and protocol criticism | |
| • Server affinity issue with H2 | |
| • Performance benefits of using H2 efficiently (e.g. better request handling) | |
| • Memory usage trade-off for increased performance | |
| • H2 protocol offers a significant savings of header bytes, potentially saving users money on bandwidth costs | |
| • The adoption of H2 requires careful consideration due to trade-offs in security and API compatibility with H1 | |
| • Status messages are removed in H2, which may break existing applications that rely on them | |
| • Designing an application specifically for H2 is necessary to take advantage of its capabilities | |
| • H2 is more suitable for internal server-to-server communication rather than user-facing web servers | |
| • The H2 project is currently being worked on by a single developer with plans to grow the team and accept open-source contributions | |
| • H2 protocol implementation and integration into Node.js | |
| • Discussion about deprecating the H1 protocol in favor of H2 | |
| • Decision to offer both H1 and H2 as options for developers | |
| • Implementation challenges and future plans for HTTP development in Node.js | |
| • Consideration for a module-based approach for H2 implementation | |
| • Discussing the size and complexity of an ecosystem | |
| • Need for input on code direction | |
| • Best ways to provide feedback and get involved (repo, open issues, pull requests) | |
| • Code stabilization and distinct areas needing help (tests, performance benchmarks) | |
| • Linking repo in show notes | |
| • Closing down Node Interactive |