Re: [Ice] Peter's review of ICEbis- Why do we nominate by putting the valid pair in the triggered check queue? (7.1.1)

Christer Holmberg <christer.holmberg@ericsson.com> Mon, 29 May 2017 06:38 UTC

From: Christer Holmberg <christer.holmberg@ericsson.com>
To: Peter Thatcher <pthatcher@google.com>, "ice@ietf.org" <ice@ietf.org>
Thread-Topic: [Ice] Peter's review of ICEbis- Why do we nominate by putting the valid pair in the triggered check queue? (7.1.1)
Thread-Index: AQHS2EYvl710MdTO5kmqLa/o+V6VaQ==
Date: Mon, 29 May 2017 06:38:29 +0000
Message-ID: <D55198C5.1D37D%christer.holmberg@ericsson.com>
Accept-Language: en-US
Content-Language: en-US
user-agent: Microsoft-MacOutlook/14.7.4.170508
Content-Type: multipart/alternative; boundary="_000_D55198C51D37Dchristerholmbergericssoncom_"
MIME-Version: 1.0
Archived-At: <https://mailarchive.ietf.org/arch/msg/ice/C3Aw75FxIfm3RkP2jRJYm9a517o>
Subject: Re: [Ice] Peter's review of ICEbis- Why do we nominate by putting the valid pair in the triggered check queue? (7.1.1)
Precedence: list

Hi,

> - Why do we nominate by putting the valid pair in the triggered check queue?  Why not just say "send a binding request"?  In particular reason?
> Do we *want* it to be delayed if there are many things in the front of the queue?

I wonder whether the reason is to maintain the pacing among the CHECK LISTs, i.e., to make sure that the nomination request transaction for CHECK LIST X is not initiated at the same time as a connectivity check transaction for CHECK LIST Y?

But, even if we put the pair in the TRIGGERED CHECK LIST, we need to make sure it becomes the top-most pair in the list, because we don’t want to test other pairs that may still be i the list.

Regards,

Christer

Re: [Ice] Peter's review of ICEbis- Why do we nom… Christer Holmberg
Re: [Ice] Peter's review of ICEbis- Why do we nom… Peter Thatcher
Re: [Ice] Peter's review of ICEbis- Why do we nom… Christer Holmberg