W3C home > Mailing lists > Public > www-ws-arch@w3.org > May 2003

WSA architectural concepts and relationsihips related to WS, SOA , and the Web

From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
Date: Tue, 6 May 2003 13:22:31 -0600
Message-ID: <9A4FC925410C024792B85198DF1E97E40598580D@usmsg03.sagus.com>
To: www-ws-arch@w3.org



> -----Original Message-----
> From: Newcomer, Eric [mailto:Eric.Newcomer@iona.com]
> Sent: Tuesday, May 06, 2003 1:49 PM
> To: Newcomer, Eric; Baker, Mark
> Cc: www-ws-arch@w3.org
> Subject: RE: WS, SOA, and the Web
> 
> 
> 
> Ok, on re-reading this I can see it's getting a bit out there 
> into theory.  The only real point I'm trying to make is that 
> at this stage in WSA, we need to concentrate on defining the 
> relationships among architectural elements.

With great trepidation, I take a stab at this, in partial fufillment of my
editors action item to rework section 1.6.3 on SOA and REST Architectures.
Consider this a line cast into the trout pond; if any big ugly brutes snap
at it I'll not bother the list with refinements it right now, but if we can
get some clearer text in the WD along these lines, that would be good:
(terms in all capital letters are "concepts" that should be in the concepts
section and glossary if we were to accept this).

DISTRIBUTED SYSTEM -- Everthing we talk about IS-A distributed system, that
is one made up of discrete software AGENTS that must work together to
implement some functionality.  Furthermore, the AGENTS in a DISTRIBUTED
SYSTEM do not operate in the same processing environment, so they must
communicate by hardware/software protocol stacks that are intrinsically less
reliable that direct code invocation and shared memory.

SERVICE -- A SERVICE IS-A  type of software AGENT that is a coarse-grained
module  that performs some well-defined operation and does not depend on the
context of a larger application. 

SERVICE ORIENTED ARCHITECTURE (SOA) -- A DISTRIBUTED SYSTEM HAS-A of a
collection of SERVICES invoked by software AGENTS.

The WORLD WIDE WEB (WWW) IA-A particular SOA that operates as a networked
information system in which AGENTS manipulate "resources" identified by a
UNIFORM RESOURCE IDENTIFIER (URI) via protocols that use URIs to directly or
indirectly communicate over the internet.

The REST WEB (need a better word, or more concepts to build up to this) IS-A
subset of the WWW in which SERVICES implement a mimimal set of operations
(create-read-update-delete or PUT-GET-POST-DELETE) and resources are
manipulated by the exchange of "representations" using these basic
operations.  [I'm sure this could be reworked using the concepts here).

[trial balloon ]The scope of the WSA (I hesitate to say WEB SERVICES for
obvious reasons!) consists of SOA that communicate via the WWW using XML.
That is, distributed systems in which the services communicate via protocols
that use URIs to directly or indirectly communicate over the internet, and
the definitions of the services and protocols [are | can be] given in an
XML-based language.

We can identify two major classes of [web services | thingies in scope for
the WSA]: [DIRECT SOA | RESTFUL WEB SERVICES] in which the primary purpose
of the service is to manipulate XML representations of Web resources using
the [minimal|CRUD|HTTP] operations, and those {INDIRECT SOA ?}in which the
primary purpose of the service is to perform an arbitrarily complex set of
operations on resources that may not be "on the Web", and the XML
[representations | messages] contain the data needed to invoke those
operations.  In other words, DIRECT SOA services are implemented by web
servers that manipulate data directly, and INDIRECT SOA services are
external code resources that are invoked via messages to web servers.


So, is the beginning of a resolution or the beginning of a feeding frenzy at
the trout pond?  Obviously there is great room for refinement and
improvement and "friendly amendments" are solicited, but if this isn't even
close to acceptable to you, please just say "Trout" and we can get on with
something for which consensus is more likely.
Received on Tuesday, 6 May 2003 15:22:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:18 GMT