Anatomy of a web service

Results 1 to 2 of 2

Thread: Anatomy of a web service

  1. #1
    Join Date
    Dec 1969

    Default Anatomy of a web service

    I would like to develop a web service to pull data from disparate databases at different physical locations and make the data searchable from the web.<BR>Here is the scenario: Many car dealers - three different inventory systems. Say 3 have systemA, 2 have systemB, 2 have systemC. I need to pull data from systemA, systemB and systemC so all dealers can check stock at all dealerships at once from the internet. This sounds like a perfect project for a web service. I am gray on a few things..<BR><BR>1. Is this indeed a good project for a web service?<BR>2. Do I need a web service for each database? ie systemA, systemB, systemC?<BR>3. The three systems will have different tables and field names.. where in the web service process does the data in the source database get mapped to generic names that a web app may consume for the end user to search on?<BR>4. Any suggested articles outlining this kind of project?<BR><BR>I am slowly comming up to speed on web services.. a few key pieces still missing in my mind.. Thanks for any input on this.<BR><BR>Steve Knowles<BR>Sytems Engineering.

  2. #2
    Join Date
    Dec 1969

    Default RE: Anatomy of a web service

    1. Sure, why not.<BR><BR>2. You have to get the dispersed data *somehow*. You could use something other than a web service (heck, you could just open up a hole in SQL Server that allowed direct network queries from one certain IP address and only using certain SPs), but what&#039;s wrong with simply having a web service that aggregates the other services?<BR><BR>3. I&#039;d do it in the individual web services. Have them serve of the same content regardless of what&#039;s backing them up. Allows you to publish a spec so that anybody else interested in being aggregated simply creates a matching service and says to you "here I am!"<BR><BR>COMMENTARY: You might want to aggressively cache the inventory on the main server. That way if a connection to one of the external services is unavailable you still have a reasonably accurate view of their inventory. In fact, given how seldom a car dealership&#039;s inventory changes (no more than a few times a day, right?), you might even think about a "push" system instead of a "pull". That is, the external sites *tell* you when their inventory changes, instead of you continually asking them.<BR><BR>I&#039;m not sure you can make a really bad choice, though, give that you will only have 3 external sites. If it were 300...<BR><BR>

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts