Web WMS-client: 2-tier vs 3-tier architecture

Roland G. roland at GISPLAN.COM.BR
Thu Nov 25 15:03:39 EST 2004


Hello,

sorry if the following has already been discussed in this list, but I
just wanted to reword the whole problem in a clear manner in the hope of
obtaining additional input. I want to know which are the specific
advantages and disadvantages of the 2-tier vs the 3-tier architecture,
read on:

We are considering one of the several available Web WMS clients
available. I define a Web WMS client as a program that runs in a
browser, using JavaScript and HTML only(we want a client without applets
or plugins).

Since you need a Web server to serve the WMS client from the following
entities will be involved in all cases:

1) The Web server(this is from where you download the WMS client to your
browser)
2) The WMS server.
3) The WMS client(which will be downloaded from the web server and run
in your browser)

Possible architectures:

2-tier: After the WMS client has been downloaded to the browser it
doesn't interact with the Web server anymore. From then on all the
interaction will go on between the WMS client and the WMS-server only.

Characteristics:
-Lots of JavaScript. More processing on the client side.
-Static client, won't be updated after initial download.
-Few processing on the web-server(only serves intial download). Good if
many simultaneous users since it relieves the server.
-genuine WMS CLIENT


3-tier: After the WMS client has been downloaded to the browser, it
sends all it's requests to the Web-server, which will forward them to
the WMS server. The WMS server will send the response back to the web
server. The web server can then process the data and send the final
response back to the WMS client. In this case the WMS client can also be
altered after each request to the web server.

Characteristics:
-Fewer JavaScript, simpler client
-More processing on the web-server(if lots of users danger of
overloading the server)
-more powerful processing possible since not limited to JavaScript on
the browser
-Client not static, can be altered during the process.
-spurious WMS client(doesn't interact with a WMS server directly)

Any additonal thoughts on the characteristics/advantages/disadvantages
of each architecture?

Thanks,

Roland



More information about the mapserver-dev mailing list