TCP Exclusivity

[martinthomson]

While protocol portability is widely regarded as a good thing, the idea that HTTP/2.0 might use a different substrate than TCP is a dangerous one. It is likely that many design choices will be made based on the assumption that the protocol runs atop TCP and not some other transport protocol.

We should decide if the HTTP/2.0 specification is written exclusively for TCP or whether we want to pay lip service to protocol agnosticism. Note that exclusivity does not preclude the use of other protocols in a later specification, it just removes one axis of freedom in the design.

[mnot]

The line we need to walk is to allow the use of other transport protocols (or application protocols that masquerade as one) that provide similar guarantees to TCP.

Thus, the profile that we identify what we're doing will exclusively and directly define the layering of HTTP/2.0 on TCP, and also upon TLS upon TCP. Other profiles might define other layerings of the semantics across other transport protocols, but we won't define them.

This is alluded to in the charter, and was discussed at length in the chartering process.

So, I see changes that need to happen in the specs to accommodate this approach, but I think they're already covered by other issues; is there any reason to keep this issue open?

[martinthomson]

I think that I'd like a conclusion that says that it's OK not to pay lip service to transport portability, so that I can change statements like:

"HTTP/2.0 runs over a reliable substrate, such as TCP."

to

"HTTP/2.0 runs over TCP."

That doesn't preclude a future effort to define the use of other transport protocols (even unreliable ones, for the truly mad).

In the same vein, I'd also like to eliminate the pretense that the framing layer is reusable for other protocols. Other efforts that decide that the framing layer works for them can reuse it, but reusability shouldn't be our goal (and our problem).

[mnot]

Again, let's see how the negotiation text fits in (#1, #12) and sort out our nomenclature for the framing layer vs. the HTTP layer; then this should become clear.

[mnot]

Ted volunteering to produce a proposal.

[mnot]

Discussed in Zurich; closed without action.