question about webtraf.*

["Guo, Liang" <guol@cs.bu.edu>]

I'm trying to understand the code for simulating webtraffic
(webcache/webtraf.*, tcl/webcache/webtraf.tcl). According to the SIGCOMM
paper, an http session should have the following hierarchy:
session arrivals follow the "InterSession Distribution", each session
has size of "SessionSize". Within each session, there might be several
pages required by the same user, but may be launched from different server
sites (however each page is loaded from the same server). Page arrivals
follow "InterPage Distribution", pages has "PageSize
(number of Objects, whose arrivals follow "InterObject
Distribution"). The object will have "ObjectSize". 

However, in webtraf.cc, procedure WebTrafSession:launchReq, it seems that
for each individual object, the session will randomly pick up a server
(dst) for it. From what I understood, this means that each object will be
loaded from different servers, so essentially here "Page" contribute
nothing to this simulation. (Although in the example ns gives, each page
contains only one object so essentially page and object are the same.)
Is this a bug or I'm missing something? Thanks in advance.

Regards.

Guo, Liang.