Re: [quicwg/base-drafts] Clarify path validation and connection migration (#4102)

[David Schinazi <[e-mail]>]

@DavidSchinazi commented on this pull request.

Overall I think this PR clarifies thing (and is purely editorial). I've added some minor comments inline.

> +entropy and might be spoofed. No method is provided to establish return
+reachability, as endpoints independently determine reachability on each
+direction of a path.

I found `No method is provided to establish return reachability` confusing. How about:

```suggestion
entropy and might be spoofed. Endpoints independently determine reachability
on each direction of a path, and therefore return reachability can only be
established by the peer.
```

> +peer. Path validation ensures that packets received from a migrating peer do
+not carry a spoofed source address.

I personally found `is used to ensure` clearer than `ensures`

```suggestion
peer. Path validation is used to ensure that packets received
from a migrating peer do not carry a spoofed source address.
```

> @@ -2152,35 +2155,40 @@ receive packets without first having sent a packet on that path. Effective NAT
 traversal needs additional synchronization mechanisms that are not provided
 here.
 
-An endpoint MAY include PATH_CHALLENGE and PATH_RESPONSE frames that are used
-for path validation with other frames.  In particular, an endpoint can pad a
-packet carrying a PATH_CHALLENGE for Path Maximum Transfer Unit (PMTU)
-discovery (see {{pmtud}}), or an endpoint can include a PATH_RESPONSE with its
-own PATH_CHALLENGE.
+An endpoint MAY include other frames with the PATH_CHALLENGE and PATH_RESPONSE
+frames used for path validation.  In particular, an endpoint can include
+PADDING with a PATH_CHALLENGE for Path Maximum Transfer Unit (PMTU) discovery
+(see {{pmtud}}), and/or include a PATH_RESPONSE with its own PATH_CHALLENGE.

This last bit sounded backwards to me in the original text. I think as a server I expect to commonly receive an unprompted PATH_CHALLENGE, and will therefore have to send PATH_RESPONSE. The performance optimization is to include a PATH_CHALLENGE with that PATH_RESPONSE, not the other way around.

```suggestion
(see {{pmtud}}), and/or include a PATH_CHALLENGE with its own PATH_RESPONSE.
```

>  
 When probing a new path, an endpoint might want to ensure that its peer has an
 unused connection ID available for responses. The endpoint can send
 NEW_CONNECTION_ID and PATH_CHALLENGE frames in the same packet. This ensures
 that an unused connection ID will be available to the peer when sending a
 response.
 
+The number of new paths an endpoint can probe is limited by the number of
+extra Connection IDs the peer has previously supplied.

I would add the following clarification:

```suggestion
extra Connection IDs the peer has previously supplied. Supplying a peer with enough
Connection IDs allows it to simultaneously probe multiple paths.
```

>  In response to such a packet, an endpoint MUST start sending subsequent packets
 to the new peer address and MUST initiate path validation ({{migrate-validate}})
-to verify the peer's ownership of the unvalidated address.
+to verify the peer's ownership of the address if it has not already.

This sounds backwards to me. Don't we need to validate the path *before* sending all subsequent packets on it?

How about:

```
In response to such a packet, an endpoint MUST initiate path validation ({{migrate-validate}})
to verify the peer's ownership of the address if it has not already. Once the address is
validated, the endpoint MUST start sending subsequent packets to the new peer address.
```

(sorry the GitHub suggestion UI seems to have choked on this one)

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/pull/4102#pullrequestreview-490069491