[net2-wg] CTP experiments

[Omprakash Gnawali]

>  From a few experiments, it became pretty clear that the current  
> switch threshold is *way* too high. I think that to start to really  
> nail down what these values should be, I need to run a long series of  
> simulation experiments with different values, which we can then  
> compare with real network results.
> 
> For now, though, I found that changing the parameters to
> 
> switch threshold = 0.2
> alpha = 0.5
> data window = 5
> 
> was much more effective.
> 

Alpha and data window are related parameters because they both control
how the packet arrival pattern is summarized. I wrote a simple script
to experiment with switching threshold, alpha, and data window
sizes. It seems that a large alpha AND a large data window is not
useful in the long run.

import random
from sys import *


# 50% pkt success
# regular pattern of 1, 0, 1, 0, ...
def pkt_arrival_on_off(n):
    for i in range(n):
        if i % 2 == 0:
            yield True
        else:
            yield False


# 80% pkt success
# regular pattern of four 1's and one 0, ...
def pkt_arrival_1(n):
    for i in range(n):
        if i % 5 == 0:
            yield False
        else:
            yield True

#80% pkt success
#random number generator
def pkt_arrival_2(n):
    for i in range(n):
        if random.random() > 0.05:
            yield True
        else:
            yield False

nval = pkt_arrival_2(500)

window_size = 3
alpha = 0.2
est = newest = 10
threshold = 1.5
pktbuf = []

print "est, curest"
for v in nval:
    pktbuf.append(v)
    if len(pktbuf) == window_size:
        if sum(pktbuf) == 0:
            curest = len(pktbuf)
        else:
            curest = 1.0 * len(pktbuf) / sum(pktbuf)
        newest = alpha * est + (1-alpha)*curest
        if est - newest > threshold:
            print "parent switched"
        est = newest
        print "%0.03f, %.03f" % (est, curest)
        pktbuf = []


- om_p