launchpad-dev team mailing list archive

[Robert Collins <robertc@xxxxxxxxxxxxxxxxx>]

On Tue, Jun 14, 2011 at 10:56 PM, Jonathan Lange <jml@xxxxxxxxxxxxx> wrote:
> On Tue, Jun 14, 2011 at 10:54 AM, Robert Collins
> <robertc@xxxxxxxxxxxxxxxxx> wrote:
> ...
>> That means that if we want to use a different mq to rabbit we need a
>> reasonably compelling reason behind that: ideally one which other
>> projects would agree with and eventually migrate.
>>
>> Now, when folk @ Canonical started deploying message queues, rabbitmq
>> was basically 'it' - the 0mq schism came along later.
>>
>> As I see it we have a few questions to answer:
>>  - should we invest $unknown_time in chasing this sporadic failure
>> down to ground
>>  - should we look at getting a rabbit expert to help us?
>>  - should we use rabbit?
>>  - and if not rabbit, what then [and what is compellingly different]?
>
> I don't have an opinion on any specific technology – I'm sadly too ignorant.

The first step to true enlightenment :)

> However, please please please take the development momentum of the
> tools into account when we choose a tool. Rabbit, say, might not have
> something we need today, but if they've got fifty responsive people
> committing and releasing, they probably will soon enough.

Both are enrolled in ohloh. http://www.ohloh.net/p/zeromq and
http://www.ohloh.net/p/rabbitmq
Both are active; rabbit appears to have more commiters (~10) vs 0mq
(~5). 0mq is a year younger than rabbit, and created by the folk that
created amqp; it is reasonable to characterise it as amqp 2.0 in the
sense that they have iterated on the basic ideas and concepts, though
the protocol is (AIUI) fairly different).

I don't think the project size difference is large enough to really be
a factor here.

Theres a lot of buzz and review and comparisons between mq's at the
moment. 0mq and rabbitmq both consistently show up.

-Rob