[rabbitmq-discuss] Pressured to move to AMQP 1.0

[DallasEng]

Having finally demonstrated to my company that RabbitMQ and AMQP allow us
to build kickass distributed solutions, I am now getting strong pressure to
move off RabbitMQ to "something more modern that supports AMQP 1.0". I
understand that 0.9.1 and 1.0 are completely different beasts, and that
there's a 1.0 protocol plugin for RabbitMQ, but that plugin has limitations
and things in 1.0 that it does not support. The Powers That Be at the
company look at IBM and Microsoft having AMQP 1.0 support and saying that's
clearly where we need to be if we want to (a) not be left behind, and (b)
advertise that we're using "the industry standard".

Since there's no indication that RabbitMQ will ever support a
fully-compliant 1.0 mode, do any of you have suggestions for how I can
demonstrate that staying on RabbitMQ for new development is a good idea?

Thanks!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140604/fcbe8233/attachment.html>

[Michael Klishin]

Post by DallasEng
Since there's no indication that RabbitMQ will ever support a
fully-compliant 1.0 mode

What kind of indication do you need?

So far RabbitMQ team received very little feedback on the plugin. Missing
features can be added given that

* There is enough interest
* AMQP 1.0 clients become stable and get support for the missing features

As for the question about staying with 0-9-1, at least when AMQP 1.0 plugin
was developed,
its ecosystem barely existed (or it existed behind closed doors). RabbitMQ
ecosystem is there and is vibrant.
Plus Rabbit supports STOMP, MQTT, and Web messaging.

[Michael Watson]

Thanks, MK.

The main problem is one of perception. "AMQP 1.0 is the standard.
It's ISO/IEC 19464 and we're chartered with using industry standards." I
can argue that 0.9.1, which has strong support from a massive community, is
the de facto standard, but "ISO/IEC" are hard TLAs to ignore.

Apache has a lot of respect internally, so Qpid is getting some strong
interest, and I haven't found any recent comparisons between RabbitMQ and
Qpid's Java or C++ implementations.

Re: The AMQP 1.0 plug-in, since we're launching creation of a large-scale
distributed platform project that needs high reliability, the fact that the
plugin still calls itself a prototype several years down the road is seen
internally as problematic.

There's some concern about 1.0's application-headers vs basic.properties
and the fact that the plugin doesn't transcode them, because we've found
those dictionaries to be useful.

MQTT is a compelling feature, thanks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140604/09adaad3/attachment.html>

[Gordon Sim]

"AMQP 1.0 is the standard. It's ISO/IEC 19464 and we're chartered with
using industry standards." I can argue that 0.9.1, which has strong
support from a massive community, is the de facto standard, but
"ISO/IEC" are hard TLAs to ignore.

In my opinion, as far as protocols go, what matters most is the extent
of support and the level of interoperability.

That's what makes TCP/IP (or HTTP,further up the stack) an industry
standard - everyone supports it.

I would argue that support for AMQP 1.0 is at least growing. It is
unlikely that the earlier iterations of the protocol see new
implementations, particularly on the broker side.

So 0.9.1 is to all intents and purposes the RabbitMQ protocol (much like
0-10 is just a Qpid protocol, or OpenWire is just an ActiveMQ protocol).

If interoperability and choice is a consideration (and I'd certainly
accept they aren't in every case) then AMQP 1.0, STOMP and MQTT are
better choices as they are supported by more than one broker
implementation (including RabbitMQ).

--Gordon.

[Simon MacMullen]

Post by DallasEng
Since there's no indication that RabbitMQ will ever support a
fully-compliant 1.0 mode, do any of you have suggestions for how I can
demonstrate that staying on RabbitMQ for new development is a good idea?

Ultimately, it's a market, not a planned economy :-)

RabbitMQ will support AMQP 1.0 to the extent that the market demands it.
You are right to say that the 1.0 plugin has seen little development
since its release: that's simply because we haven't heard back from
anyone who's using it. If AMQP 1.0 takes off, RabbitMQ users will not be
left behind.

(Is anyone out there using it? Please tell us if you are!)

Post by DallasEng
The main problem is one of perception. "AMQP 1.0 is the standard.
It's ISO/IEC 19464 and we're chartered with using industry standards." I can
argue that 0.9.1, which has strong support from a massive community, is
the de facto standard, but "ISO/IEC" are hard TLAs to ignore.

I would argue that "ISO standard" and "industry standard" are two
different things. TCP/IP is an industry standard, the OSI networking
stack is an ISO standard, even though no-one uses it.

Cheers, Simon

[rambocoder]

You can make a business case to your management that instead of spending
hundreds of thousands of dollars on IBM Websphere MQ to get started,and
then much more indefinitely for IBM support contract, you could hire 1
Erlang developer who would implement the missing AMQP 1.0 features in
RabbitMQ https://github.com/rabbitmq/rabbitmq-amqp1.0

That would create tremendous goodwill towards your company and make the
world a better place :)

Or the company could just pay huge monthly bills by using Microsoft Azure
Messaging http://azure.microsoft.com/en-us/pricing/details/service-bus/
because Microsoft charges for e-v-e-r-y-thing, even "Billable messages
include any "no message available" replies".

Cheers,

-rambocoder

Post by DallasEng
Having finally demonstrated to my company that RabbitMQ and AMQP allow us
to build kickass distributed solutions, I am now getting strong pressure to
move off RabbitMQ to "something more modern that supports AMQP 1.0". I
understand that 0.9.1 and 1.0 are completely different beasts, and that
there's a 1.0 protocol plugin for RabbitMQ, but that plugin has limitations
and things in 1.0 that it does not support. The Powers That Be at the
company look at IBM and Microsoft having AMQP 1.0 support and saying that's
clearly where we need to be if we want to (a) not be left behind, and (b)
advertise that we're using "the industry standard".
Since there's no indication that RabbitMQ will ever support a
fully-compliant 1.0 mode, do any of you have suggestions for how I can
demonstrate that staying on RabbitMQ for new development is a good idea?
Thanks!

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140609/0abef3b1/attachment.html>

[Michael Watson]

Thanks, ramoboder.

I'm working through the issues now. Have had some success, but decision
makers still somewhat hung up on the damned versioning. It's hard to claim
that we're using an industry standard in AMQP 0.9.1, even though many
consider it the defacto standard. There's business value in being able to
claim compatibility with the version that Microsoft, IBM, and Apache are
all supporting.

Michael

Post by rambocoder
You can make a business case to your management that instead of spending
hundreds of thousands of dollars on IBM Websphere MQ to get started,and
then much more indefinitely for IBM support contract, you could hire 1
Erlang developer who would implement the missing AMQP 1.0 features in
RabbitMQ https://github.com/rabbitmq/rabbitmq-amqp1.0
That would create tremendous goodwill towards your company and make the
world a better place :)
Or the company could just pay huge monthly bills by using Microsoft Azure
Messaging http://azure.microsoft.com/en-us/pricing/details/service-bus/
because Microsoft charges for e-v-e-r-y-thing, even "Billable messages
include any "no message available" replies".
Cheers,
-rambocoder

Post by DallasEng
Having finally demonstrated to my company that RabbitMQ and AMQP allow us
to build kickass distributed solutions, I am now getting strong pressure to
move off RabbitMQ to "something more modern that supports AMQP 1.0". I
understand that 0.9.1 and 1.0 are completely different beasts, and that
there's a 1.0 protocol plugin for RabbitMQ, but that plugin has limitations
and things in 1.0 that it does not support. The Powers That Be at the
company look at IBM and Microsoft having AMQP 1.0 support and saying that's
clearly where we need to be if we want to (a) not be left behind, and (b)
advertise that we're using "the industry standard".
Since there's no indication that RabbitMQ will ever support a
fully-compliant 1.0 mode, do any of you have suggestions for how I can
demonstrate that staying on RabbitMQ for new development is a good idea?
Thanks!

_______________________________________________
rabbitmq-discuss mailing list
rabbitmq-discuss at lists.rabbitmq.com
https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140609/668230b5/attachment.html>

[Tony Garnock-Jones]

Based on the state of interop as discussed the other day, I reckon the
RabbitMQ team should declare the 1.0 plugin "production". It doesn't seem
to be in any worse of a state than any other AMQP 1.0 implementation. It's
as compatible as anyone else, by the sounds of it.

If that were done, the boxes would all get ticked and people would be
reassured they could migrate to AMQP 1.0 if and when they felt like it.

Post by Michael Watson
Thanks, ramoboder.
I'm working through the issues now. Have had some success, but decision
makers still somewhat hung up on the damned versioning. It's hard to claim
that we're using an industry standard in AMQP 0.9.1, even though many
consider it the defacto standard. There's business value in being able to
claim compatibility with the version that Microsoft, IBM, and Apache are
all supporting.
Michael

Post by rambocoder
You can make a business case to your management that instead of spending
hundreds of thousands of dollars on IBM Websphere MQ to get started,and
then much more indefinitely for IBM support contract, you could hire 1
Erlang developer who would implement the missing AMQP 1.0 features in
RabbitMQ https://github.com/rabbitmq/rabbitmq-amqp1.0
That would create tremendous goodwill towards your company and make the
world a better place :)
Or the company could just pay huge monthly bills by using Microsoft Azure
Messaging http://azure.microsoft.com/en-us/pricing/details/service-bus/
because Microsoft charges for e-v-e-r-y-thing, even "Billable messages
include any "no message available" replies".
Cheers,
-rambocoder

Post by DallasEng
Having finally demonstrated to my company that RabbitMQ and AMQP allow
us to build kickass distributed solutions, I am now getting strong pressure
to move off RabbitMQ to "something more modern that supports AMQP 1.0". I
understand that 0.9.1 and 1.0 are completely different beasts, and that
there's a 1.0 protocol plugin for RabbitMQ, but that plugin has limitations
and things in 1.0 that it does not support. The Powers That Be at the
company look at IBM and Microsoft having AMQP 1.0 support and saying that's
clearly where we need to be if we want to (a) not be left behind, and (b)
advertise that we're using "the industry standard".
Since there's no indication that RabbitMQ will ever support a
fully-compliant 1.0 mode, do any of you have suggestions for how I can
demonstrate that staying on RabbitMQ for new development is a good idea?
Thanks!

_______________________________________________
rabbitmq-discuss mailing list
rabbitmq-discuss at lists.rabbitmq.com
https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss

_______________________________________________
rabbitmq-discuss mailing list
rabbitmq-discuss at lists.rabbitmq.com
https://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss