Re: [quicwg/base-drafts] 5tuple routing (#3536)

[Martin Thomson <[e-mail]>]

@martinthomson commented on this pull request.

I'm sure that this is now basically unrecognizable...

> @@ -1156,6 +1156,34 @@ SHOULD ignore any such packets.
 
 Servers MUST drop incoming packets under all other circumstances.
 
+### Considerations for 5-tuple routing architectures

```suggestion
### Considerations for Simple Load Balancers
```

> +* If servers can use other dedicated server IP addresses or ports than
+the one that the client is initiating connections to, they could use the
+preferred_address transport parameter to request that clients move
+connections to these dedicated addresses. Note that clients could
+choose not to use the preferred address.

```suggestion
* If servers can use a dedicated server IP address or port, other than the one
that the client is initially connects to, they could use the preferred_address
transport parameter to request that clients move connections to a dedicated
address. Note that clients could choose not to use the preferred address.
```

> +could result in packets being forwarded to the wrong server. Such a server
+deployment could use one of the following methods for connection continuity
+when a client's address changes.
+
+* Servers could use an out-of-band mechanism to forward packets to the correct
+server based on Connection ID.
+
+* If servers can use other dedicated server IP addresses or ports than
+the one that the client is initiating connections to, they could use the
+preferred_address transport parameter to request that clients move
+connections to these dedicated addresses. Note that clients could
+choose not to use the preferred address.
+
+A server in a deployment that does not implement a solution to
+maintain connection continuity during connection migration
+SHOULD disallow it via the disable_active_migration transport

```suggestion
SHOULD disallow migration using the disable_active_migration transport
```

> +Server deployments that use such load balancing might allow for
+creation of a stateless reset oracle, and should therefore follow the
+guidelines in {{reset-oracle}}.

Might as well just reiterate the existing requirement instead:

```suggestion
Server deployments that use this simple form of load balancing MUST avoid the
creation of a stateless reset oracle; see {{reset-oracle}}.
```

Fewer words, too.

-- 
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/3536#pullrequestreview-402270209