This is still a draft based on a personal understanding. Selenium RC developers should confirm this before it is final.

Parts Involved

There are four main processes involved in the run of a Selenium RC test, test code (using Ruby as example here), selenium server instance, testing browser instance, and web application under test

Code

@sel = Selenium::SeleniumDriver.new("localhost", 4444, "*chrome", "http://www.google.com", 15000)
@sel.start
@sel.open("http://www.google.com")
@sel.type("q", "hello world")
...
@sel.stop

Sequence Chart

Here is a poor man's version of the sequence diagram for a Selenium test run in Ruby

Ruby test code	Data Flow	Selenium Server	Data Flow	Browser
		running
@sel = ...new..
@sel.start	--->
		starts browser	--->
		waits		starts
			<---	connects to selenium server
	<---	returns new session id
(stores session id)
@sel.open
(sends open command with session id)	-->
		forwards command as reply to browser	---->
				selenium core processes command, performs the action
			<---	sends result as another request in the same session
		processes result
	<---	sends result back to test process
@sel.type
(sends type command with session id and arguments)	-->
		forwards command as reply to browser	---->
				selenium core processes command, performs the action
			<---	sends result as another request
		processes result
	<---	sends result back to test process
...
...
@sel.stop
(sends type command with session id)	-->
		shuts down browser
	<---	ends sessions
...