Questions
1. The snort stats look a bit screwie.
   1. Cpu stats look funny, not even consistant funny realtek/ intel look rather different and don't jive with the os's version of things.
   2. Avg_bytes, is inconsistant with known ethernet specs. Avg bytes of 1800 bytes per packet is probibly incorrect.
   3. Drops is zero... so what is the definition of 'drops' in this case?
   4. open_sessions and new_sessions are seemingly the same. it's probibly okay, but worth thinking about.

This is a work in progress, as such it's not quite perfect. Here is a partial list of imperfections. Tell me about more and I'll add them, and maybe try to fix them.

1. There seem to be an abnormal number of context switches on the DUTs durring the early portions of the mtu slide traffic pattern. It is reproducable, not sure what's going on. This may be caused by the early parts of this traffic profile being generation bound, thus there are task switches to the idle process.
2. snort_inline.all seems to have some of it's data dissapeared, yet the DUT data seems to be fine.
3. why is all.preproc with no rules, is slower than all which has all the processors and the rules?
4. The y_scale is incorrect on the avgbytes / packet v time grpahs, fixed pending verification.
5. need cpu states v time.
6. The rtt times for ganamead/ir_wm_snort_inline.all/all.totaltimes.vtime.png, look to be flat... interesting.
7. Like wise the the snort.stats have extra datapoints where i'm not sure they should.
8. Maxintime lines of the rooster graphs can be radicly wrong when the old average is >> new average, this dosn't matter to much of the annalysis because durring these times the study is focusing on another dut, but it is visualy distracting.