Exercise 1:

- Comments:

* A lot of you seem to wonder why there are so many flows 
from a web-server compared to less flows to a web-server. 
Please, think about the nature of a typical www-document. 
And combine this nature of web document structure to the 
definition of a flow. You could also try to find out how 
http 1.0 and http 1.1. differ from each other.

* uniq is nice command, but it has its limitations. You should 
always discover the weaknesses of your tools.

* Please, take care of the presentation. Structurize your 
reports so that results are clear and visible. Put code into
appendixes. Explain the code in the text and make reference 
to it. Try to present results clearly.

* Nice catch to use bit shifting in determining the combinations. Props.

Results (f=fail, p=pass , d=pass with distinction 
and bonus for the final exam):

Opkno	Ex1
55645R	p
54281V	p
51418E	p
55612W	p
54691L	d
55563J	p
49782C	p
51538M	p
54736U	d
54271J	p
49714M	p
49909T	p
51360D	p
49528S	d
49929U	p
49799A	p
49794S	p
49720U	p
55383V	p
76494A	p
55533T	d
58055M	p

Exercise 2:

- Comments:

* Distributions seem to be really hard to understand. For instance, marking means, 
vars, etc. goes far better by making them vertical to the appropriate bin location. 
Not horizontal, please.

* If packet arrivals are sampled every 10 seconds, for instance, then the results, 
say 10000 packet arrivals, needs to be divided by 10 (seconds) to get pkts/s values. 
Really disappointing how many got this wrong. 

* Funet-traces (and flows) completely misunderstood. Please, think when a flow starts, 
when it ends and when it is written to a trace file. How does this effect to one 
trace file and flow arrivals there? How accurate is the characterization of the flow
arrival process? Also very disappointed for the lack of thinking here. However, probably
also my bad for not explaining the flow gathering process with enough details. 

* The down side of "code sharing" is now visible. One makes bad code, (almost) everybody uses it, and
loses possibility to get extra points from this exercise. Do not just copy, think about the results and
figures. Do they make sense? Are you able to understand them? Are the figures clear and to the point? 
Are the axis-values explained?     

* It's perfectly ok to use Matlab's EWMA for plotting "a moving mean". If you do this, please explain 
how it works and how it differs (if at all) from the EMA given in "Chapter 2".

* Although you're allowed to work in groups, please do not stop thinking for yourself.

* It seems that a lot of people have put a lot of time into this exercise. Sad to say, 
but most of the time your efforts have not produced quality.


Cumulative Results (1st exercise included) (f=fail, p=pass , d=pass with distinction 
and bonus for the final exam):

Opkno	Ex1	Ex2
55645R	p	p
54281V	p	p
51418E	p	d
55612W	p	p
54691L	d	p
55563J	p	p
49782C	p	-
51538M	p	p
54736U	d	p
54271J	p	p
49714M	p	p
49909T	p	p
51360D	p	p
49528S	d	p
49929U	p	p
49799A	p	d
49794S	p	p
49720U	p	d
55383V	p	p
76494A	p	p
55533T	d	p
58055M	p	-