Just a quick top 10 list:
1)
Both: Lack of peer-to-peer capabilities –server required
2)
MQTT: No formal way to implement any kind of federation of multiple MQTT server(s) or implement high availability
3)
MQTT : Lack of any kind of RPC mechanism (the IoT isn’t just about data – it’s also about services and events)
4)
Both: Lack of discovery of server(s), device(s) and device capabilities
5)
Both: No discoverable meta-model/weak data typing
6)
Both: no easy way to integrate from within browser-based apps without other “bridges” (undervalues the role of “people” in the IoT)
7)
Both: marginal security and privacy capabilities with very “chunky” granularity
8)
XMPP: QoS policies are very application-specific
9)
Both: scarcity of reference code and open source options for server implementation(s)
10)
MQTT: No support from the IETF – depends on your definition of the “internet” of thing whether or not this matters
I could probably keep going, but you get the idea.
From: m2m-iwg-bounces@xxxxxxxxxxx [mailto:m2m-iwg-bounces@xxxxxxxxxxx]
On Behalf Of James Butcher
Sent: Monday, June 24, 2013 9:43 AM
To: m2m-iwg@xxxxxxxxxxx
Subject: Re: [m2m-iwg] Beyond MQTT: A Cisco View on IoT Protocols
Hi
Its an interesting subject. I agree it will be a combination of these technologies that take the IoT forward and that the commercial drive for it will be key.
Rick - which specific limitations of XMPP and MQTT are you referring to there?
A technology that wasn't mentioned in the article is the OMG's Data Distribution Service (DDS). It addresses interoperability but also enables real-time scalable communication for both device-to-cloud and device-to-device in a pub/sub manner.
James
-
From: Rick Bullotta <rick.bullotta@xxxxxxxxxxxxx>
-
Date: Fri, 21 Jun 2013 11:53:50 +0000
-
Accept-language: en-US
-
Delivered-to: m2m-iwg@xxxxxxxxxxx
-
Thread-index: AQHObnX+r3x1g0jEU0265NvRaI/K3Q==
-
Thread-topic: [m2m-iwg] Beyond MQTT: A Cisco View on IoT Protocols
It will always be a case of "AND", not "OR". The protocols that solve actual interoperability and usage issues and make it easier to build useful applications will succeed.
Personally I don't think XMPP and MQTT will dominate long term because of fundamental limitations for usage in the IoT. COAP is interesting but the binary focus creates a ton of interop challenges.
Probably some evolution of COAP with richer metadata, multiple protocol bindings, and expanded data and service representations could succeed.
Very little if any new technology needs to be created to realize a solid solution. It's really more of a people and organizational challenge.
I don't think this has been shared on the list before, but here is [1] a nice reflection on protocols for M2M and IoT. It's interesting to see how CoAP, MQTT, XMPP, … compare with each other from a very high-level perspective. Also, interesting
to see that XMPP seems to be moving more and more towards IoT (extensions for provisioning, sensor data interchange, efficient XML interchange)…
"Connecting sensors and objects opens up an entirely new world of possible use cases—and it’s precisely those use cases that will determine when to use the right protocols for the right applications."
