Favor quiet circuits when choosing which order to relay cells in.

Each circuit is ranked in terms of how many cells from it have been
relayed recently, using a time-weighted average.

This patch has been tested this on a private Tor network on PlanetLab,
and gotten improvements of 12-35% in time it takes to fetch a small
web page while there's a simultaneous large data transfer going on
simultaneously.

[Commit msg by nickm based on mail from Ian Goldberg.]

1 files changed, 12 insertions, 0 deletions

diff --git a/src/or/circuitlist.c b/src/or/circuitlist.c
index 2c949def00..eb8c90f460 100644
--- a/src/or/circuitlist.c
+++ b/src/or/circuitlist.c
@@ -385,6 +385,10 @@ init_circuit_base(circuit_t *circ)
   circ->package_window = circuit_initial_package_window();
   circ->deliver_window = CIRCWINDOW_START;
 
+  /* Initialize the cell_ewma_t structure */
+  circ->n_cell_ewma.last_cell_time = circ->highres_created;
+  circ->n_cell_ewma.cell_count = 0.0;
+
   circuit_add(circ);
 }
 
@@ -432,6 +436,14 @@ or_circuit_new(circid_t p_circ_id, or_connection_t *p_conn)
 
   init_circuit_base(TO_CIRCUIT(circ));
 
+  /* Initialize the cell_ewma_t structure */
+
+  /* Fetch the timeval that init_circuit_base filled in. */
+  circ->p_cell_ewma.last_cell_time = TO_CIRCUIT(circ)->highres_created;
+
+  /* Initialize the cell counts to 0 */
+  circ->p_cell_ewma.cell_count = 0.0;
+
   return circ;
 }