Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [paho-dev] On Paho's opinionated URI validation

Since the Paho project is going to host some interop tests and is looking to be a reference implementation (AIUI, and I'm one of the leads *grin*) then I certainly wouldn't see it as an issue for us to propose a short spec on this and have others sign up to it if they feel able to. I think I know at least 2 other projects who might want to support this :-)

Still a bit concerned about the "MQTT-S" vs mqtts:// thing but if it is well documented (and -S becomes -SN) then we can make that work too.

On Tue, Nov 5, 2013 at 11:16 AM, Paul Fremantle <paul.fremantle@xxxxxxxxxx> wrote:
I agree with Michael this would be a very useful enhancement to the standardization. In a lot of situations ESBs (like Apache Synapse, Mule, Camel, etc) use the URI prefix as a decider for which protocol to use, and the generic tcp:// prefix is obviously overloaded. The benefit of a dedicated scheme is that it simplifies the configuration of these systems, and it also means that as soon as you see a URI in code or in a config file, you have an idea of the protocol. In other words its helpful to both the machine and the developer.

Dave - yes I agree this is outside the OASIS charter. I personally think this *could* be done in OASIS, but it would require some painful work in the OASIS group to re-charter. The other approach (option #2) would be for the main participants in MQTT to publish a very short joint spec, or alternatively for us to document this under Eclipse Paho and hope that it is adopted. The final approach (option #3) would be to choose a third-party standards body, but that seems a bit weird to me. 

Of course if we do a de-facto approach (option #2) then that doesn't preclude the OASIS group taking that on board in the future if it seems a good idea.


On 5 November 2013 10:31, Michael Klishin <michael.s.klishin@xxxxxxxxx> wrote:
2013/11/5 Dave Locke <locke@xxxxxxxxxx>
The approach taken in the Paho clients has been to assume that the protocol is MQTT and then the URI is the transport that MQTT will flow across e.g. TCP://, SSL://  WS://, WSS://     If this was standardised is this enough or is there a need to not assume anything and have a uri's like MQTT://  (for TCP)  MQTTS:// (TCP with SSL/TLS)   MQTTWS://  (Websockets ...)  ?


Having mqtt://, mqtts:// (over TLS), mqttws:// and mqttsws:// sounds reasonable.
It would largely solve the problem for the PaaS providers and make MQTT much closer to the convention
HTTP and AMQP have.

paho-dev mailing list

Paul Fremantle
Part-time PhD student - School of Computing
twitter: pzfreo / skype: paulfremantle / blog:
CTO and Co-Founder, WSO2
OASIS WS-RX TC Co-chair, Apache Member
07740 199 729

paho-dev mailing list

Andy Piper | Kingston upon Thames, London (UK)
blog:   |   skype: andypiperuk
twitter: @andypiper  |  images:

Back to the top