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

Re: Proposed text on reliability in the web services architecture

From: Paul Prescod <paul@prescod.net>
Date: Fri, 10 Jan 2003 17:53:42 -0800
Message-ID: <3E1F7926.1080703@prescod.net>
To: Assaf Arkin <arkin@intalio.com>, www-ws-arch@w3.org
CC: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, bhaugen <linkage@interaccess.com>

Assaf Arkin wrote:

> Let's assume for the second that all SQL upates are idempotent. (Not
> necessarily, but it simplifies the discussion) And SQL selects don't 
> have any side effects (not that I know of). So mapping SQL statements 
> to HTTP GET/PUT would make my database RESTful (correct?).

I don't think it is useful to think of it as mapping SQL statements to 
HTTP GET/PUT. You have abstractions you want to expose to your 
customers. You develop a set of resources that are a logical, secure and 
abstract view on those resources. Some of the data underneath those 
resources may be expressed as relational tables. Other bits of it may 
come from anywhere else.

> Persumably, if I expose my database as a Web service I could build a 
> RESTful service. And it will scale as well as the database does, which 
> is not a bad proposition (afterall most applications simply add 
> overhead to the database).

A REST web service will scale as well as it is architected to. The 
chances of it scaling in exactly the same fashion as some particular 
data source seems vanishingly small.

Let's not get confused about the relationship between HTTP methods and 
relational data. The point is NOT that there is a straightforward 
mapping which allows the removal of an intermediate layer.

> putting a middle layer between the client and the database.

Mike Champion is correct that it takes effort to insert a REST layer 
precisely because it _is_ a layer. And because conforming to standards 
(whether architectural standards or syntactic ones) always takes more 
effort than just doing what you were doing before the standard came along.

>  Instead of
> performing database operations, the client performs application 
> operations which then become sequences of database operations. One you 
> sequence a select and an update, or two updates, with some logic, 
> you're no longer 100% idempotent  How much, I never bothered to 
> measure. 20%? 40%? 60%

I have no idea what this means.

> You don't have to do it. You can build a system out of a set of small
> idempotent operations.

No you cannot. In general, you cannot build most useful systems out of 
only idempotent operations. That's why HTTP does not provide only 
idempotent operations.

  Paul Prescod
Received on Friday, 10 January 2003 20:54:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:13 GMT