If small footprint is your primary aim you might want to consider using the Stomp protocol which is extremely easy to write a client for (at the expense of some performance for high volume clients).
http://stomp.codehaus.org/Protocol e.g. the entire Ruby client is a page or two of code (and its way more advanced than you probably need) - or the .Net client is 2 trivial classes. http://svn.codehaus.org/stomp/trunk/ruby/lib/stomp.rb You can generally write a Stomp client from scratch in a couple of hours assuming you can figure out how to use a socket and parse Strings. On 8/4/06, Gordon Sim <[EMAIL PROTECTED]> wrote:
Good question, though one I do not have as yet an answer for. However, the classes used purely for the encoding/decoding in java are currently about 150K unoptimised. They do also rely on Mina at present which is 278K. There would be a small amount on top of this most to tie this all into a minimal api (the current jms adaptor layer is 136K on top of this as a rough guide, excluding some of the external dependencies like log4j & apache commons, none of which are currently critical). To this point, no attempt has been made to specifically keep the size of the implementation below a particular threshold and further optimisation could probably bring these numbers down (a less generic io framework than mina for example could be used and by being more limited in scope would bring the size down quite a bit I expect). At runtime, the protocol allows negotiation of the maximum framesize per connection, the minimum acceptable value being 4096 bytes. So a client could operate within a restricted memory environment if required. (Again though, the current java client has not been developed with this goal in mind). I hope that helps a little bit, though its rather imprecise(?) Coach Wei wrote: > Per Cliff Schmidt's suggestion, reposting the question here instead - so > please respond using this new thread: > > What is the possible (estimated) minimum implementation footprint (in > term of kilobytes or megabytes) to support AMQP network wire-level > protocol? I am asking this thinking of the possibility of using AMQP > protocol in mobile applications such as J2ME. > > ---Coach Wei > > > >> From: Cliff Schmidt [mailto:[EMAIL PROTECTED] >> Sent: Thursday, August 03, 2006 3:13 PM >> To: general@incubator.apache.org >> Subject: Re: [VOTE] Accept Glasgow into Incubator >> >> Coach, >> >> If you don't view your question as related to the vote, would you mind >> reposting it to a separate or an existing thread about Glasgow? >> >> Maybe it's just my personal preference, but I like to keep vote >> > threads to > >> just votes and critical questions that were missed in the prior >> > discussion > >> (which, I admit, often doesn't happen). >> >> Thanks, >> Cliff >> >> > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- James ------- http://radio.weblogs.com/0112098/ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
