Compare: Livestats 6/ webtrends/Hitslink

How heavy can the load be on Hitslink?

[Receptional]

We have a client with over 300,000 unique visitors per month load balanced on two servers. (I wish I had that many...)

As a result we get two logfiles which take many hours to process using Webtrends and then we still have two webtrends reports!(so we process one and double the numbers to get an estimate!)

We are thinking that using a cookie based system like hitslink will fix a few issues, but would it be a big mistake? My assumptions are:
1) Webtrends gets it very wrong due to AOL users and the like being seen as many people each session (AOL uses multiple IP numbers every time a user logs on, hence the huge Virginia statistic)
2) A cookie based tracking system would get around the AOL issue
3) Multiple servers would not be an issue using Hitslink.
4) Because ISPs cache data at their server level, the log files for the server are unreliable anyway.

BUT...
Would a site of this size struggle to display data using something like?

I know that the new Livestats (Deepmetrix) can handle up to 8 servers but I am still very wary of log based stats due to caching and this difficulty in identifying a user throughout the session. It also crashed our server when we tried to install it, which did not make me happy.

Are here any other issues I have missed?

[TRACEYR]

I have been looking at this and the Files from the server and measuring hits etc takes forever to digest. Depending on what you want to do, i.e
[xav.com...] that does the tracking with Perl and you can then run this on your own server and see the results, it also logs cached pages.

It takes a while to install, but i was pretty impressed and it is FREE, you can pay $20 to support it.