Update the ratio calculation to prefer faster measurements.

Also: simplify complicated filtering steps, always take the
most recent measurement, and use slightly smaller file sizes
based on measurements.
This commit is contained in:
Mike Perry 2009-07-14 20:47:51 -07:00
parent 011b732436
commit 3a1b9526f8

View File

@ -45,12 +45,11 @@ Status: Open
file is downloaded via these circuits. The file sizes are set based
on node percentile rank as follows:
0-10: 4M
10-20: 2M
20-30: 1M
30-50: 512k
50-75: 256k
75-100: 128k
0-10: 2M
10-20: 1M
20-30: 512k
30-50: 256k
50-100: 128k
These sizes are based on measurements performed during test scans.
@ -72,8 +71,8 @@ Status: Open
with capacity of one standard deviation below a node's average
are also removed.
The final ratio result will be the unfiltered ratio if it is
close to 1.0, otherwise it will be the filtered ratio.
The final ratio result will be greater of the unfiltered ratio
and the filtered ratio.
6. Pseudocode for Ratio Calculation Algorithm
@ -88,11 +87,8 @@ Status: Open
BW_measured(N) = MEAN(b | b is bandwidth of a stream through N)
Bw_stddev(N) = STDDEV(b | b is bandwidth of a stream through N)
Bw_avg(S) = MEAN(b | b = BW_measured(N) for all N in S)
Normal_Routers(S) = {N | Bw_measured(N)/Bw_avg(S) > 0.5 }
for N in S:
Normal_Streams(N) =
{stream via N | all nodes in stream not in {Normal_Routers(S)-N}
and bandwidth > BW_measured(N)-Bw_stddev(N)}
Normal_Streams(N) = {stream via N | bandwidth >= BW_measured(N)}
BW_Norm_measured(N) = MEAN(b | b is a bandwidth of Normal_Streams(N))
Bw_net_avg(Slices) = MEAN(BW_measured(N) for all N in Slices)
@ -100,9 +96,9 @@ Status: Open
for N in all Slices:
Bw_net_ratio(N) = Bw_measured(N)/Bw_net_avg(Slices)
Bw_Norm_net_ratio(N) = Bw_measured2(N)/Bw_Norm_net_avg(Slices)
Bw_Norm_net_ratio(N) = BW_Norm_measured(N)/Bw_Norm_net_avg(Slices)
ResultRatio(N) = ClosestToOne(Bw_net_ratio(N), Bw_Norm_net_ratio(N))
ResultRatio(N) = MAX(Bw_net_ratio(N), Bw_Norm_net_ratio(N))
7. Security implications
@ -142,8 +138,8 @@ Status: Open
by another script that uses them to produce network-wide averages
and calculate ratios as per the algorithm in section 6. Because nodes
may shift in capacity, they may appear in more than one slice and/or
appear more than once in the file set. The line that yields a ratio
closest to 1.0 will be chosen in this case.
appear more than once in the file set. The most recently measured
line will be chosen in this case.
9. Integration with Proposal 160