[OASIS Issue Tracker] (MQTT-320) Expectations of timing accuracy in MQTT implementations

[OASIS Issues Tracker <workgroup_mailer@lists.oasis-open.org>]

    [ https://issues.oasis-open.org/browse/MQTT-320?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=64088#comment-64088 ] 

Ken Borgendale commented on MQTT-320:
-------------------------------------

There are multiple issues involved in the timer accuracy:
- Clock drift
- Periodic processing of timeouts
- administrative override of requested time intervals

In the case of message expiration you also have the issue that queue and transport times are not accounted for.

I would rather not do too much definition of this other than to note the issues.  The 1.5x of the keepalive became normative text and causes more problems than if it was just not there.

> Expectations of timing accuracy in MQTT implementations
> -------------------------------------------------------
>
>                 Key: MQTT-320
>                 URL: https://issues.oasis-open.org/browse/MQTT-320
>             Project: OASIS Message Queuing Telemetry Transport (MQTT) TC
>          Issue Type: Improvement
>          Components: core
>    Affects Versions: 5
>            Reporter: Ian Craggs
>            Priority: Minor
>
> There are a number of aspects of MQTT which involve timings, starting with the keepalive interval, and some new features introduced in MQTT 5 such as message expiration.
> These timers are denoted in seconds (I don't think we have any exceptions to that), but the implementation of those timers in both servers and clients may not actually be at the resolution of small numbers of seconds.  I suggest that we include some wording to limit the expectation of accuracy when a small number of seconds is used on such a timer, where "small" is to be defined, but could be less than 10 for instance.



--
This message was sent by Atlassian JIRA
(v6.2.2#6258)