More cloud storage performance-measurement scripts.

[bluesky.git] / parsetrace / split-trace.py

diff --git a/parsetrace/split-trace.py b/parsetrace/split-trace.py
index e1ff68a..caf77ee 100755 (executable)
--- a/parsetrace/split-trace.py
+++ b/parsetrace/split-trace.py
@@ -7,7 +7,8 @@ import impacket, itertools, pcapy, re, socket, subprocess, sys
 import impacket.ImpactDecoder, impacket.ImpactPacket
 
 # Domain names for cloud service providers, whose traces we want to pull out.
-DOMAINS = ['.amazon.com', '.core.windows.net']
+DOMAINS = ['.amazon.com', '.amazonaws.com', '.core.windows.net',
+           '204.246.162.', '87.238.86.']
 
 # The collection of flows we've seen.  The value associated with each flow is a
 # sequence number indicating in what order we saw the flows in the trace.
@@ -53,7 +54,10 @@ for file in sys.argv[1:]:
         matches = False
         for d in DOMAINS:
             if name.endswith(d): matches = True
-        if not matches: continue
+            if name.startswith(d): matches = True
+        if not matches:
+            print "Host", name, "not recognized, skipping"
+            continue
 
         filter = "tcp and (host %s and host %s) and (port %d and port %d)" \
             % (src, dst, sport, dport)