[tor-commits] [torflow/master] Tweaks to try to make PID look more like other bw auths.

[mikeperry at torproject.org]

commit 335ce85c130c5ffef8525afb56f77ad423e12f9b
Author: Mike Perry <mikeperry-git at fscked.org>
Date:   Thu Nov 10 14:25:11 2011 -0800

    Tweaks to try to make PID look more like other bw auths.
    
    Just to make sure something crazy is not going on, let's try eliminating the
    circ failure dampening, and re-instate the "larger ratio" idea for PID
    control. With these both re-enabled, PID "feedback" should look more similar
    to the other bandwidth auths, because with just one auth running the code,
    there is no feedback.
---
 NetworkScanners/BwAuthority/aggregate.py |   18 ++++++++++++------
 1 files changed, 12 insertions(+), 6 deletions(-)

diff --git a/NetworkScanners/BwAuthority/aggregate.py b/NetworkScanners/BwAuthority/aggregate.py
index be9c595..d1c8b45 100755
--- a/NetworkScanners/BwAuthority/aggregate.py
+++ b/NetworkScanners/BwAuthority/aggregate.py
@@ -139,7 +139,8 @@ class Node:
       self.filt_bw = line.filt_bw
       self.ns_bw = line.ns_bw
       self.desc_bw = line.desc_bw
-      self.circ_fail_rate = line.circ_fail_rate
+      # XXX: Temporary test
+      self.circ_fail_rate = 0.0 #line.circ_fail_rate
       self.strm_fail_rate = line.strm_fail_rate
 
 class Line:
@@ -360,6 +361,8 @@ def main(argv):
     # socket exhaustion issues..
     true_filt_avg = sum(map(lambda n: n.filt_bw*(1.0-n.circ_fail_rate),
                          nodes.itervalues()))/float(len(nodes))
+    true_strm_avg = sum(map(lambda n: n.strm_bw*(1.0-n.circ_fail_rate),
+                         nodes.itervalues()))/float(len(nodes))
   else:
     true_filt_avg = sum(map(lambda n: n.filt_bw,
                          nodes.itervalues()))/float(len(nodes))
@@ -396,13 +399,17 @@ def main(argv):
   tot_net_bw = 0
   for n in nodes.itervalues():
     n.fbw_ratio = n.filt_bw/true_filt_avg
-
-    # Always use filtered bandwidths for feedback
-    n.ratio = n.fbw_ratio
+    n.sbw_ratio = n.strm_bw/true_strm_avg
 
     if cs_junk.bwauth_pid_control:
       # Penalize nodes for circ failure rate
-      n.pid_error = (n.filt_bw*(1.0-n.circ_fail_rate) - true_filt_avg)/true_filt_avg
+      # n.pid_error = (n.filt_bw*(1.0-n.circ_fail_rate) - true_filt_avg)/true_filt_avg
+
+      # FIXME: For now, let's try this larger-ratio thing..
+      if n.sbw_ratio > n.fbw_ratio:
+        n.pid_error = (n.strm_bw*(1.0-n.circ_fail_rate) - true_strm_avg)/true_strm_avg
+      else:
+        n.pid_error = (n.filt_bw*(1.0-n.circ_fail_rate) - true_filt_avg)/true_filt_avg
 
       if n.idhex in prev_votes.vote_map:
         # If there is a new sample, let's use it for all but guards
@@ -449,7 +456,6 @@ def main(argv):
         plog("INFO", "No prev vote for node "+n.nick+": Consensus feedback")
     else: # No PID feedback
       # Choose the larger between sbw and fbw
-      n.sbw_ratio = n.strm_bw/true_strm_avg
       if n.sbw_ratio > n.fbw_ratio:
         n.ratio = n.sbw_ratio
       else: