D-AG0019 [RE: D-AG0007.1- defining reliable and stable WS ]

Hi Hugo,

-----Original Message-----
From: Hugo Haas [mailto:hugo@w3.org]
Sent: Wednesday, March 20, 2002 7:06 PM
[snip]

I think that you disagree with the "reliable" part of the proposed
requirement.  Hmmm... Rereading Suresh's email[1], reliability
(guaranteed access as I understand Suresh's wording) does seem
unreasonable; I had missed that in my first reading.

<sd>
My response to your comments on the definition. of reliability [1]
"Reliable - A WS is reliable as long as accessing (SD: a very "loaded" term
here) a WS is guaranteed under clearly specified conditions to all potential
users."

I was only providing a definition, not a requirement.
Is this definition acceptable?

Classical reliability definitions are always conditional. Phrases
like MTBF, and 99% reliable, etc are traditionally used to describe it.
In the WS case, we might say for example, 
The XYZ-WS is "guaranteed" to be accessible with  0.99 probability between
8AM - 2PM CDT
or such. In other words, "guarantee" does not mean 100%.
Again, the term "accessible" is not accessing a URI. It is the reliability
of messaging
used, and reliability of the binding of the message payload to the service,
and the service
itself (if one were to trisect).

As for the argument that web is a lossy media, I would say, so is any media
of communication. Lossy media does not remove the need for reliability defn.
or requirement.
E.g., we have requirements or at least expectations of reliability from our
phone companies.

Can WS-A define what constitutes reliability? Probably yes. Can WS-A require
certain reliability for all its implementations? I would say no.

Regards,
-Suresh

</sd>


  1. http://lists.w3.org/Archives/Public/www-ws-arch/2002Mar/0309.html

Received on Thursday, 21 March 2002 18:42:56 UTC