Re: [sca-assembly] some of our concerns with the eventing proposal - Pub/Sub

[Mike Edwards <mike_edwards@uk.ibm.com>]

Yours,  Mike.

Strategist - Emerging Technologies, SCA & SDO.
Co Chair OASIS SCA Assembly TC.
IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.
Phone & FAX: +44-1962-818014    Mobile: +44-7802-467431  
Email:  mike_edwards@uk.ibm.com

Danny van der Rijn <dannyv@tibco.com> wrote
on 09/09/2009 22:40:11:

> [image removed] 
> 
> Re: [sca-assembly] some of our concerns with the eventing proposal
> 
> Danny van der Rijn 
> 
> to:
> 
> Anish Karmarkar
> 
> 09/09/2009 22:40
> 
> Cc:
> 
> Jim Marino, Martin Chapman, Scott Vorthmann, OASIS Assembly
> 
> 
> 
> Anish Karmarkar wrote:
> [full conversation below]
> One of the mismatches:
>  From a pub-sub dev POV, as a publisher I don't want to be aware
of all
> the subscribers and vice versa. In the current model (for Java POJO)
> what I get on the reference side is an array of references -- which
> breaks the model. I then also have to iterate over the array and do
a
> "publish" for each member of the array. That makes no sense
from a
> pub-sub POV. I want to one publish per event/message. One-way is not
> pub=sub.
> 
> I wouldn't go so far as to say that the current model's x..n wirings
> is anything more than a poor-man's hack to approximate pub/sub 
> (perhaps *this* is Martin's chocolate teapot).  
> 
> However, using a binding such as JMS with a x..1 binding comes a lot
> closer to pub/sub.  One-way IS pub/sub IF your transport has
pub/sub
> qualities.  
> 
> Squeezing pub/sub out of a fundamentally point-to-point transport
> like HTTP is going to take a much larger miracle than the current
proposal.
> 

It isn't a miracle at all.  WS-Eventing and WS-Notification
show how to do it.
And there are implementations of both of them.


<snip>

---

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU