X-Git-Url: http://plrg.eecs.uci.edu/git/?p=pingpong.git;a=blobdiff_plain;f=Code%2FProjects%2FSmartPlugDetector%2Fsrc%2Fmain%2Fjava%2Fedu%2Fuci%2Fiotproject%2FMain.java;h=55e57574ef956d38c2f937afe0bd64d44b72a087;hp=b1e56508245515da6259f279f16c77dfa72e2c7f;hb=f2adc4163688cb942f207fe6cf91aedc191d07e2;hpb=0402965f143de5ae5efd5f8899aa7d6a108194b4 diff --git a/Code/Projects/SmartPlugDetector/src/main/java/edu/uci/iotproject/Main.java b/Code/Projects/SmartPlugDetector/src/main/java/edu/uci/iotproject/Main.java index b1e5650..55e5757 100644 --- a/Code/Projects/SmartPlugDetector/src/main/java/edu/uci/iotproject/Main.java +++ b/Code/Projects/SmartPlugDetector/src/main/java/edu/uci/iotproject/Main.java @@ -1,15 +1,28 @@ package edu.uci.iotproject; -import edu.uci.iotproject.analysis.PcapPacketPair; -import edu.uci.iotproject.analysis.PcapProcessingPipeline; -import edu.uci.iotproject.analysis.TcpConversationUtils; +import static edu.uci.iotproject.analysis.UserAction.Type; + +import edu.uci.iotproject.analysis.*; +import edu.uci.iotproject.comparison.seqalignment.ExtractedSequence; +import edu.uci.iotproject.comparison.seqalignment.SequenceAlignment; +import edu.uci.iotproject.comparison.seqalignment.SequenceExtraction; +import edu.uci.iotproject.io.TriggerTimesFileReader; +import edu.uci.iotproject.util.PcapPacketUtils; +import edu.uci.iotproject.util.PrintUtils; +import org.apache.commons.math3.stat.clustering.Cluster; +import org.apache.commons.math3.stat.clustering.DBSCANClusterer; import org.pcap4j.core.*; +import org.pcap4j.packet.namednumber.DataLinkType; import java.io.EOFException; +import java.io.File; +import java.io.PrintWriter; import java.net.UnknownHostException; -import java.util.ArrayList; -import java.util.List; +import java.time.Instant; +import java.util.*; import java.util.concurrent.TimeoutException; +import java.util.stream.Collectors; +import java.util.stream.Stream; /** * This is a system that reads PCAP files to compare @@ -26,95 +39,520 @@ public class Main { public static void main(String[] args) throws PcapNativeException, NotOpenException, EOFException, TimeoutException, UnknownHostException { // ------------------------------------------------------------------------------------------------------------- - // Example/debug code for searching for a pattern at the MAC layer. -// String fileName = "./pcap/mac-tplink.local.pcapng"; -// PcapHandle handle; -// try { -// handle = Pcaps.openOffline(fileName, PcapHandle.TimestampPrecision.NANO); -// } catch (PcapNativeException pne) { -// handle = Pcaps.openOffline(fileName); -// } -// Arrays.asList(1590, 1590, 1590, 1001, 337, 197, 636, 1311, 177) // Full pattern (all non-zero payload packets). -// MacLayerFlowPattern pattern = new MacLayerFlowPattern("TP_LINK_LOCAL_OFF_MAC", "50:c7:bf:33:1f:09", Arrays.asList(637, 1312)); -// MacLayerFlowPatternFinder finder = new MacLayerFlowPatternFinder(handle, pattern); -// finder.findFlowPattern(); - // ------------------------------------------------------------------------------------------------------------- -// -// //final String fileName = args.length > 0 ? args[0] : "/home/rtrimana/pcap_processing/smart_home_traffic/Code/Projects/SmartPlugDetector/pcap/wlan1.local.dns.pcap"; -// final String fileName = args.length > 0 ? args[0] : "/scratch/June-2018/TPLink/wlan1/tplink.wlan1.local.pcap"; -// //final String fileName = args.length > 0 ? args[0] : "/scratch/June-2018/DLink/wlan1/dlink.wlan1.local.pcap"; -// final String trainingFileName = "./pcap/TP_LINK_LOCAL_ON_SUBSET.pcap"; -//// final String trainingFileName = "./pcap/TP_LINK_LOCAL_ON.pcap"; -//// -//// // ====== Debug code ====== -// PcapHandle handle; -// PcapHandle trainingPcap; -// try { -// handle = Pcaps.openOffline(fileName, PcapHandle.TimestampPrecision.NANO); -// trainingPcap = Pcaps.openOffline(trainingFileName, PcapHandle.TimestampPrecision.NANO); -// } catch (PcapNativeException pne) { -// handle = Pcaps.openOffline(fileName); -// trainingPcap = Pcaps.openOffline(trainingFileName); -// } -//// -//// // TODO: The followings are the way to extract multiple hostnames and their associated packet lengths lists -//// //List list = new ArrayList<>(); -//// //list.add("events.tplinkra.com"); -//// //FlowPattern fp = new FlowPattern("TP_LINK_LOCAL_ON", list, trainingPcap); -//// //List list2 = new ArrayList<>(); -//// //list2.add("devs.tplinkcloud.com"); -//// //list2.add("events.tplinkra.com"); -//// //FlowPattern fp3 = new FlowPattern("TP_LINK_REMOTE_ON", list2, trainingPcap); -//// -// FlowPattern fp = new FlowPattern("TP_LINK_LOCAL_ON", "events.tplinkra.com", trainingPcap); -// //FlowPattern fp = new FlowPattern("DLINK_LOCAL_ON", "rfe-us-west-1.dch.dlink.com", trainingPcap); -// FlowPatternFinder fpf = new FlowPatternFinder(handle, fp); -// fpf.start(); -//// -//// // ======================== + // ------------ # Code for extracting traffic generated by a device within x seconds of a trigger # ------------ + // Paths to input and output files (consider supplying these as arguments instead) and IP of the device for + // which traffic is to be extracted: + String path = "/scratch/July-2018"; // Rahmadi +// String path = "/Users/varmarken/temp/UCI IoT Project/experiments"; // Janus + boolean verbose = true; + final String onPairsPath = "/scratch/July-2018/on.txt"; + final String offPairsPath = "/scratch/July-2018/off.txt"; - /* - PcapReader pcapReader = new PcapReader(args[0]); - PcapProcessingPipeline pipeline = new PcapProcessingPipeline(pcapReader); - TcpReassembler tcpReassembler = new TcpReassembler(); - pipeline.addPcapPacketConsumer(tcpReassembler); - pipeline.executePipeline(); - System.out.println("Pipeline terminated"); + // 1) D-Link July 26 experiment +// final String inputPcapFile = path + "/2018-07/dlink/dlink.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-07/dlink/dlink-processed.pcap"; +// final String triggerTimesFile = path + "/2018-07/dlink/dlink-july-26-2018.timestamps"; +// final String deviceIp = "192.168.1.199"; // .246 == phone; .199 == dlink plug? + // Actual training +// final String inputPcapFile = path + "/2018-10/dlink-plug/dlink-plug.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/dlink-plug/dlink-plug-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/dlink-plug/dlink-plug-oct-17-2018.timestamps"; +// final String deviceIp = "192.168.1.199"; // .246 == phone; .199 == dlink plug? + // TODO: EXPERIMENT - November 7, 2018 +// final String inputPcapFile = path + "/experimental_result/standalone/dlink-plug/wlan1/dlink-plug.wlan1.local.pcap"; +// final String outputPcapFile = path + "/experimental_result/standalone/dlink-plug/wlan1/dlink-plug-processed.pcap"; +// final String triggerTimesFile = path + "/experimental_result/standalone/dlink-plug/timestamps/dlink-plug-nov-7-2018.timestamps"; +// //final String deviceIp = "192.168.1.199"; // .246 == phone; .199 == dlink plug? +// final String deviceIp = "192.168.1.246"; // .246 == phone; .199 == dlink plug? - List> pairs = new ArrayList<>(); - for (Conversation c : tcpReassembler.getTcpConversations()) { - pairs.add(TcpConversationUtils.extractPacketPairs(c)); - } - */ + // 2) TP-Link July 25 experiment +// final String inputPcapFile = path + "/2018-07/tplink/tplink.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-07/tplink/tplink-processed.pcap"; +// final String triggerTimesFile = path + "/2018-07/tplink/tplink-july-25-2018.timestamps"; +// final String deviceIp = "192.168.1.159"; + // Actual training +// final String inputPcapFile = path + "/2018-10/tplink-plug/tplink-plug.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/tplink-plug/tplink-plug-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/tplink-plug/tplink-plug-oct-17-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .159 == tplink plug + // TODO: EXPERIMENT - November 8, 2018 +// final String inputPcapFile = path + "/experimental_result/standalone/tplink-plug/wlan1/tplink-plug.wlan1.local.pcap"; +// final String outputPcapFile = path + "/experimental_result/standalone/tplink-plug/wlan1/tplink-plug-processed.pcap"; +// final String triggerTimesFile = path + "/experimental_result/standalone/tplink-plug/timestamps/tplink-plug-nov-8-2018.timestamps"; +// final String deviceIp = "192.168.1.159"; // .246 == phone; .159 == tplink plug +//// final String deviceIp = "192.168.1.246"; // .246 == phone; .159 == tplink plug - // -------- 07-17-2018 -------- - // Only consider packets to/from the TP-Link plug. - PcapReader pcapReader = new PcapReader(args[0], "ip host 192.168.1.159"); - TcpReassembler tcpReassembler = new TcpReassembler(); - PcapPacket packet; - while((packet = pcapReader.readNextPacket()) != null) { - tcpReassembler.consumePacket(packet); + // 2b) TP-Link July 25 experiment TRUNCATED: + // Only contains "true local" events, i.e., before the behavior changes to remote-like behavior. + // Last included event is at July 25 10:38:11; file filtered to only include packets with arrival time <= 10:38:27. +// final String inputPcapFile = path + "/2018-07/tplink/tplink.wlan1.local.truncated.pcap"; +// final String outputPcapFile = path + "/2018-07/tplink/tplink-processed.truncated.pcap"; +// final String triggerTimesFile = path + "/2018-07/tplink/tplink-july-25-2018.truncated.timestamps"; +// final String deviceIp = "192.168.1.159"; + + // 3) SmartThings Plug July 25 experiment +// final String inputPcapFile = path + "/2018-07/stplug/stplug.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-07/stplug/stplug-processed.pcap"; +// final String triggerTimesFile = path + "/2018-07/stplug/smartthings-july-25-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!) + // October 18 +// final String inputPcapFile = path + "/2018-10/st-plug/st-plug.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/st-plug/st-plug-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/st-plug/st-plug-oct-18-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!) + + // 4) Wemo July 30 experiment +// final String inputPcapFile = path + "/2018-07/wemo/wemo.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-07/wemo/wemo-processed.pcap"; +// final String triggerTimesFile = path + "/2018-07/wemo/wemo-july-30-2018.timestamps"; +// final String deviceIp = "192.168.1.145"; // .246 == phone; .145 == WeMo + + // 5) Wemo Insight July 31 experiment +// final String inputPcapFile = path + "/2018-07/wemoinsight/wemoinsight.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-07/wemoinsight/wemoinsight-processed.pcap"; +// final String triggerTimesFile = path + "/2018-07/wemoinsight/wemo-insight-july-31-2018.timestamps"; +// final String deviceIp = "192.168.1.135"; + + // 6) TP-Link Bulb August 1 experiment +// final String inputPcapFile = path + "/2018-08/tplink-bulb/tplinkbulb.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/tplink-bulb/tplinkbulb-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/tplink-bulb/tplink-bulb-aug-3-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .140 == TP-Link bulb + + // 7) Kwikset Doorlock August 6 experiment +// final String inputPcapFile = path + "/2018-08/kwikset-doorlock/kwikset-doorlock.data.wlan1.pcap"; +//// final String inputPcapFile = path + "/2018-08/kwikset-doorlock/kwikset-doorlock.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/kwikset-doorlock/kwikset-doorlock-processed.pcap"; +//// final String triggerTimesFile = path + "/2018-08/kwikset-doorlock/kwikset-doorlock-aug-6-2018.timestamps"; +// final String triggerTimesFile = path + "/2018-08/kwikset-doorlock/kwikset-doorlock-8hr-data-oct-11-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!) + // TODO: EXPERIMENT - November 10, 2018 + final String inputPcapFile = path + "/experimental_result/standalone/kwikset-doorlock/wlan1/kwikset-doorlock.wlan1.local.pcap"; + final String outputPcapFile = path + "/experimental_result/standalone/kwikset-doorlock/wlan1/kwikset-doorlock-processed.pcap"; +// final String inputPcapFile = path + "/experimental_result/standalone/kwikset-doorlock/eth1/kwikset-doorlock.eth1.local.pcap"; +// final String outputPcapFile = path + "/experimental_result/standalone/kwikset-doorlock/eth1/kwikset-doorlock-processed.pcap"; + final String triggerTimesFile = path + "/experimental_result/standalone/kwikset-doorlock/timestamps/kwikset-doorlock-nov-10-2018.timestamps"; +// final String deviceIp = "192.168.1.183"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!) + final String deviceIp = "192.168.1.246"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!) + + // September 12, 2018 - includes both wlan1 and eth1 interfaces +// final String inputPcapFile = path + "/2018-08/kwikset-doorlock/kwikset3.wlan1.local.pcap"; +// //final String inputPcapFile = path + "/2018-08/kwikset-doorlock/kwikset3.eth1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/kwikset-doorlock/kwikset3-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/kwikset-doorlock/kwikset-doorlock-sept-12-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!) + + // 8) Hue Bulb August 7 experiment +// final String inputPcapFile = path + "/2018-08/hue-bulb/hue-bulb.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/hue-bulb/hue-bulb-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/hue-bulb/hue-bulb-aug-7-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; + // October 30 experiment +// final String inputPcapFile = path + "/2018-10/hue-bulb/hue-bulb.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/hue-bulb/hue-bulb-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/hue-bulb/hue-bulb-oct-30-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .100 == Hue hub + + // 9) Lifx Bulb August 8 experiment +// final String inputPcapFile = path + "/2018-08/lifx-bulb/lifx-bulb.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/lifx-bulb/lifx-bulb-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/lifx-bulb/lifx-bulb-aug-8-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .231 == Lifx + // October 18 +// final String inputPcapFile = path + "/2018-10/lifx-bulb/lifx-bulb.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/lifx-bulb/lifx-bulb-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/lifx-bulb/lifx-bulb-oct-18-2018.timestamps"; +// final String deviceIp = "192.168.1.231"; // .246 == phone; .231 == Lifx + // November 1 +// final String inputPcapFile = path + "/2018-10/lifx-bulb/lifx-bulb.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/lifx-bulb/lifx-bulb-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/lifx-bulb/lifx-bulb-nov-1-2018.timestamps"; +// final String deviceIp = "192.168.1.231"; // .246 == phone; .231 == Lifx + + // 10) Amcrest Camera August 9 experiment +// final String inputPcapFile = path + "/2018-08/amcrest-camera/amcrest-camera.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/amcrest-camera/amcrest-camera-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/amcrest-camera/amcrest-camera-aug-9-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .235 == camera + + // 11) Arlo Camera August 10 experiment +// final String inputPcapFile = path + "/2018-08/arlo-camera/arlo-camera.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/arlo-camera/arlo-camera-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/arlo-camera/arlo-camera-aug-10-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .140 == camera + + // 12) Blossom sprinkler August 13 experiment +// final String inputPcapFile = path + "/2018-08/blossom/blossom.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/blossom/blossom-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/blossom/blossom-aug-13-2018.timestamps"; +// final String deviceIp = "192.168.1.229"; // .246 == phone; .229 == sprinkler + // 2 November +// final String inputPcapFile = path + "/2018-10/blossom-sprinkler/blossom-sprinkler.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-10/blossom-sprinkler/blossom-sprinkler-processed.pcap"; +// final String triggerTimesFile = path + "/2018-10/blossom-sprinkler/blossom-sprinkler-nov-2-2018.timestamps"; +// final String deviceIp = "192.168.1.229"; // .246 == phone; .229 == sprinkler + +// // 13) DLink siren August 14 experiment +// final String inputPcapFile = path + "/2018-08/dlink-siren/dlink-siren.wlan1.local.pcap"; +// //final String inputPcapFile = path + "/evaluation/dlink-siren/dlink-siren.data.wlan1.pcap"; +// final String outputPcapFile = path + "/2018-08/dlink-siren/dlink-siren-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/dlink-siren/dlink-siren-oct-12-2018.timestamps"; +// //final String triggerTimesFile = path + "/2018-08/dlink-siren/dlink-siren-aug-14-2018.timestamps"; +// //final String triggerTimesFile = path + "/actual/timestamps/dlink-siren-8hr-data-oct-10-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .183 == siren + // TODO: EXPERIMENT - November 9, 2018 +// final String inputPcapFile = path + "/experimental_result/standalone/dlink-siren/wlan1/dlink-siren.wlan1.local.pcap"; +// final String outputPcapFile = path + "/experimental_result/standalone/dlink-siren/wlan1/dlink-siren-processed.pcap"; +// final String triggerTimesFile = path + "/experimental_result/standalone/dlink-siren/timestamps/dlink-siren-nov-9-2018.timestamps"; +// final String deviceIp = "192.168.1.183"; // .246 == phone; .183 == siren +//// final String deviceIp = "192.168.1.246"; // .246 == phone; .183 == siren + + // 14) Nest thermostat August 15 experiment +// final String inputPcapFile = path + "/2018-08/nest/nest.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/nest/nest-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/nest/nest-aug-15-2018.timestamps"; +// final String deviceIp = "192.168.1.246"; // .246 == phone; .127 == Nest thermostat + + // 15) Alexa August 16 experiment +// final String inputPcapFile = path + "/2018-08/alexa/alexa.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/alexa/alexa-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/alexa/alexa-aug-16-2018.timestamps"; +// final String deviceIp = "192.168.1.225"; // .246 == phone; .225 == Alexa + // August 17 +// final String inputPcapFile = path + "/2018-08/alexa/alexa2.wlan1.local.pcap"; +// final String outputPcapFile = path + "/2018-08/alexa/alexa2-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/alexa/alexa-aug-17-2018.timestamps"; +// final String deviceIp = "192.168.1.225"; // .246 == phone; .225 == Alexa + + // September 17 +// final String inputPcapFile = path + "/2018-08/noise/noise.eth1.pcap"; +// final String outputPcapFile = path + "/2018-08/noise/noise-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/noise/noise-sept-17-2018.timestamps"; +// final String deviceIp = "192.168.1.142"; // .142 == SmartThings Hub; .199 == dlink plug; .183 == siren + // September 26 - D-Link noise +// final String inputPcapFile = path + "/2018-08/noise/noise.dlink.wlan1.pcap"; +// final String outputPcapFile = path + "/2018-08/noise/noise-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/noise/dlink-noise-sept-26-2018.timestamps"; +// final String deviceIp = "192.168.1.183"; // .199 == dlink plug; .183 == siren + // September 27 - Kwikset noise +// final String inputPcapFile = path + "/2018-08/noise/noise.kwikset.eth1.pcap"; +// final String outputPcapFile = path + "/2018-08/noise/noise-processed.pcap"; +// final String triggerTimesFile = path + "/2018-08/noise/kwikset-doorlock-noise-sept-27-2018.timestamps"; +// final String deviceIp = "192.168.1.142"; // .142 == SmartThings Hub; + + TriggerTimesFileReader ttfr = new TriggerTimesFileReader(); + List triggerTimes = ttfr.readTriggerTimes(triggerTimesFile, false); + // Tag each trigger with "ON" or "OFF", assuming that the first trigger is an "ON" and that they alternate. + List userActions = new ArrayList<>(); + for (int i = 0; i < triggerTimes.size(); i++) { + userActions.add(new UserAction(i % 2 == 0 ? Type.TOGGLE_ON : Type.TOGGLE_OFF, triggerTimes.get(i))); } - // Now we have a set of reassembled TCP conversations. - List conversations = tcpReassembler.getTcpConversations(); - for(Conversation c : conversations) { - List pairs = TcpConversationUtils.extractPacketPairs(c); - for (PcapPacketPair pair : pairs) { - // TODO ... - // 1. discard packets that are not within X seconds after trigger time - // 2. conversations may be (are) with different servers - so need to plot in different plots, one per hostname? + TriggerTrafficExtractor tte = new TriggerTrafficExtractor(inputPcapFile, triggerTimes, deviceIp); + final PcapDumper outputter = Pcaps.openDead(DataLinkType.EN10MB, 65536).dumpOpen(outputPcapFile); + DnsMap dnsMap = new DnsMap(); + TcpReassembler tcpReassembler = new TcpReassembler(); + TrafficLabeler trafficLabeler = new TrafficLabeler(userActions); + tte.performExtraction(pkt -> { + try { + outputter.dump(pkt); + } catch (NotOpenException e) { + e.printStackTrace(); } + }, dnsMap, tcpReassembler, trafficLabeler); + outputter.flush(); + outputter.close(); + + if (tte.getPacketsIncludedCount() != trafficLabeler.getTotalPacketCount()) { + // Sanity/debug check + throw new AssertionError(String.format("mismatch between packet count in %s and %s", + TriggerTrafficExtractor.class.getSimpleName(), TrafficLabeler.class.getSimpleName())); } - // ---------------------------- + // Extract all conversations present in the filtered trace. + List allConversations = tcpReassembler.getTcpConversations(); + // Group conversations by hostname. + Map> convsByHostname = TcpConversationUtils.groupConversationsByHostname(allConversations, dnsMap); + System.out.println("Grouped conversations by hostname."); + // For each hostname, count the frequencies of packet lengths exchanged with that hostname. + final Map> pktLenFreqsByHostname = new HashMap<>(); + convsByHostname.forEach((host, convs) -> pktLenFreqsByHostname.put(host, TcpConversationUtils.countPacketLengthFrequencies(convs))); + System.out.println("Counted frequencies of packet lengths exchanged with each hostname."); + // For each hostname, count the frequencies of packet sequences (i.e., count how many conversations exchange a + // sequence of packets of some specific lengths). + final Map> pktSeqFreqsByHostname = new HashMap<>(); + convsByHostname.forEach((host, convs) -> pktSeqFreqsByHostname.put(host, TcpConversationUtils.countPacketSequenceFrequencies(convs))); + System.out.println("Counted frequencies of packet sequences exchanged with each hostname."); + // For each hostname, count frequencies of packet pairs exchanged with that hostname across all conversations + final Map> pktPairFreqsByHostname = + TcpConversationUtils.countPacketPairFrequenciesByHostname(allConversations, dnsMap); + System.out.println("Counted frequencies of packet pairs per hostname"); + // For each user action, reassemble the set of TCP connections occurring shortly after + final Map> userActionToConversations = trafficLabeler.getLabeledReassembledTcpTraffic(); + final Map>> userActionsToConvsByHostname = trafficLabeler.getLabeledReassembledTcpTraffic(dnsMap); + System.out.println("Reassembled TCP conversations occurring shortly after each user event"); + /* + * NOTE: no need to generate these more complex on/off maps that also contain mappings from hostname and + * sequence identifiers as we do not care about hostnames and sequences during clustering. + * We can simply use the UserAction->List map to generate ON/OFF groupings of conversations. + */ + /* + // Contains all ON events: hostname -> sequence identifier -> list of conversations with that sequence + Map>> ons = new HashMap<>(); + // Contains all OFF events: hostname -> sequence identifier -> list of conversations with that sequence + Map>> offs = new HashMap<>(); + userActionsToConvsByHostname.forEach((ua, hostnameToConvs) -> { + Map>> outer = ua.getType() == Type.TOGGLE_ON ? ons : offs; + hostnameToConvs.forEach((host, convs) -> { + Map> seqsToConvs = TcpConversationUtils. + groupConversationsByPacketSequence(convs, verbose); + outer.merge(host, seqsToConvs, (oldMap, newMap) -> { + newMap.forEach((sequence, cs) -> oldMap.merge(sequence, cs, (list1, list2) -> { + list1.addAll(list2); + return list1; + })); + return oldMap; + }); + }); + }); + */ + + // ================================================ CLUSTERING ================================================ + // Note: no need to use the more convoluted on/off maps; can simply use the UserAction->List map + // when don't care about hostnames and sequences (see comment earlier). + List onConversations = userActionToConversations.entrySet().stream(). + filter(e -> e.getKey().getType() == Type.TOGGLE_ON). // drop all OFF events from stream + map(e -> e.getValue()). // no longer interested in the UserActions + flatMap(List::stream). // flatten List> to a List + collect(Collectors.toList()); + List offConversations = userActionToConversations.entrySet().stream(). + filter(e -> e.getKey().getType() == Type.TOGGLE_OFF). + map(e -> e.getValue()). + flatMap(List::stream). + collect(Collectors.toList()); + //Collections.sort(onConversations, (c1, c2) -> c1.getPackets().) + + List onPairs = onConversations.stream(). + map(c -> c.isTls() ? TcpConversationUtils.extractTlsAppDataPacketPairs(c) : + TcpConversationUtils.extractPacketPairs(c)). + flatMap(List::stream). // flatten List> to List<> + collect(Collectors.toList()); + List offPairs = offConversations.stream(). + map(c -> c.isTls() ? TcpConversationUtils.extractTlsAppDataPacketPairs(c) : + TcpConversationUtils.extractPacketPairs(c)). + flatMap(List::stream). // flatten List> to List<> + collect(Collectors.toList()); + // Note: need to update the DnsMap of all PcapPacketPairs if we want to use the IP/hostname-sensitive distance. + Stream.concat(Stream.of(onPairs), Stream.of(offPairs)).flatMap(List::stream).forEach(p -> p.setDnsMap(dnsMap)); + // Perform clustering on conversation logged as part of all ON events. + DBSCANClusterer onClusterer = new DBSCANClusterer<>(10.0, 45); + List> onClusters = onClusterer.cluster(onPairs); + // Perform clustering on conversation logged as part of all OFF events. + DBSCANClusterer offClusterer = new DBSCANClusterer<>(10.0, 45); + List> offClusters = offClusterer.cluster(offPairs); + // Sort the conversations as reference + List sortedAllConversation = TcpConversationUtils.sortConversationList(allConversations); + // Output clusters + System.out.println("========================================"); + System.out.println(" Clustering results for ON "); + System.out.println(" Number of clusters: " + onClusters.size()); + int count = 0; + List>> ppListOfListReadOn = new ArrayList<>(); + List>> ppListOfListListOn = new ArrayList<>(); + for (Cluster c : onClusters) { + System.out.println(String.format("<<< Cluster #%02d (%03d points) >>>", ++count, c.getPoints().size())); + System.out.print(PrintUtils.toSummaryString(c)); + if(c.getPoints().size() > 45 && c.getPoints().size() < 55) { + // Print to file + List> ppListOfList = PcapPacketUtils.clusterToListOfPcapPackets(c); + ppListOfListListOn.add(ppListOfList); + } + } + // TODO: Merging test + ppListOfListListOn = PcapPacketUtils.mergeSignatures(ppListOfListListOn, sortedAllConversation); + ppListOfListListOn = PcapPacketUtils.sortSignatures(ppListOfListListOn); + // TODO: Need to remove sequence number 2 for ST plug since it is not a good signature! + //PcapPacketUtils.removeSequenceFromSignature(ppListOfListListOn, 2); + // TODO: Need to remove sequence number 0 for TP-Link plug since it is not a good signature! + // TODO: This sequence actually belongs to the local communication between the plug and the phone + //PcapPacketUtils.removeSequenceFromSignature(ppListOfListListOn, 0); + PcapPacketUtils.printSignatures(ppListOfListListOn); + //count = 0; + /*for (List> ll : ppListOfListListOn) { + PrintUtils.serializeClustersIntoFile("./onSignature" + ++count + ".sig", ll); + ppListOfListReadOn.add(PrintUtils.deserializeClustersFromFile("./onSignature" + count + ".sig")); + }*/ + PrintUtils.serializeSignatureIntoFile("./onSignature.sig", ppListOfListListOn); + ppListOfListReadOn = PrintUtils.deserializeSignatureFromFile("./onSignature.sig"); + + System.out.println("========================================"); + System.out.println(" Clustering results for OFF "); + System.out.println(" Number of clusters: " + offClusters.size()); + count = 0; + List>> ppListOfListReadOff = new ArrayList<>(); + List>> ppListOfListListOff = new ArrayList<>(); + for (Cluster c : offClusters) { + System.out.println(String.format("<<< Cluster #%03d (%06d points) >>>", ++count, c.getPoints().size())); + System.out.print(PrintUtils.toSummaryString(c)); + if(c.getPoints().size() > 45 && c.getPoints().size() < 55) { + // Print to file + List> ppListOfList = PcapPacketUtils.clusterToListOfPcapPackets(c); + ppListOfListListOff.add(ppListOfList); + } + } + // TODO: Merging test + ppListOfListListOff = PcapPacketUtils.mergeSignatures(ppListOfListListOff, sortedAllConversation); + ppListOfListListOff = PcapPacketUtils.sortSignatures(ppListOfListListOff); + // TODO: Need to remove sequence number 1 for Nest Thermostat since it is not a good signature! + //PcapPacketUtils.removeSequenceFromSignature(ppListOfListListOff, 1); + // TODO: Need to remove sequence number 0 for Arlo Camera since it is not a good signature! + //PcapPacketUtils.removeSequenceFromSignature(ppListOfListListOff, 0); + // TODO: Need to remove sequence number 2 for ST plug since it is not a good signature! + //PcapPacketUtils.removeSequenceFromSignature(ppListOfListListOff, 2); + // TODO: Need to remove sequence number 0 for TP-Link plug since it is not a good signature! + // TODO: This sequence actually belongs to the local communication between the plug and the phone + //PcapPacketUtils.removeSequenceFromSignature(ppListOfListListOff, 0); + PcapPacketUtils.printSignatures(ppListOfListListOff); + //count = 0; + /*for (List> ll : ppListOfListListOff) { + PrintUtils.serializeClustersIntoFile("./offSignature" + ++count + ".sig", ll); + ppListOfListReadOff.add(PrintUtils.deserializeClustersFromFile("./offSignature" + count + ".sig")); + }*/ + PrintUtils.serializeSignatureIntoFile("./offSignature.sig", ppListOfListListOff); + ppListOfListReadOff = PrintUtils.deserializeSignatureFromFile("./offSignature.sig"); + System.out.println("========================================"); + // ============================================================================================================ + + /* + System.out.println("==== ON ===="); + // Print out all the pairs into a file for ON events + File fileOnEvents = new File(onPairsPath); + PrintWriter pwOn = null; + try { + pwOn = new PrintWriter(fileOnEvents); + } catch(Exception ex) { + ex.printStackTrace(); + } + for(Map.Entry>> entry : ons.entrySet()) { + Map> seqsToConvs = entry.getValue(); + for(Map.Entry> entryConv : seqsToConvs.entrySet()) { + List listConv = entryConv.getValue(); + // Just get the first Conversation because all Conversations in this group + // should have the same pairs of Application Data. + for(Conversation conv : listConv) { + // Process only if it is a TLS packet + if (conv.isTls()) { + List tlsAppDataList = TcpConversationUtils.extractTlsAppDataPacketPairs(conv); + for(PcapPacketPair pair: tlsAppDataList) { + System.out.println(PrintUtils.toCsv(pair, dnsMap)); + pwOn.println(PrintUtils.toCsv(pair, dnsMap)); + } + } else { // Non-TLS conversations + List packetList = TcpConversationUtils.extractPacketPairs(conv); + for(PcapPacketPair pair: packetList) { + System.out.println(PrintUtils.toCsv(pair, dnsMap)); + pwOn.println(PrintUtils.toCsv(pair, dnsMap)); + } + } + } + } + } + pwOn.close(); + System.out.println("==== OFF ===="); + // Print out all the pairs into a file for ON events + File fileOffEvents = new File(offPairsPath); + PrintWriter pwOff = null; + try { + pwOff = new PrintWriter(fileOffEvents); + } catch(Exception ex) { + ex.printStackTrace(); + } + for(Map.Entry>> entry : offs.entrySet()) { + Map> seqsToConvs = entry.getValue(); + for(Map.Entry> entryConv : seqsToConvs.entrySet()) { + List listConv = entryConv.getValue(); + // Just get the first Conversation because all Conversations in this group + // should have the same pairs of Application Data. + for(Conversation conv : listConv) { + // Process only if it is a TLS packet + if (conv.isTls()) { + List tlsAppDataList = TcpConversationUtils.extractTlsAppDataPacketPairs(conv); + for(PcapPacketPair pair: tlsAppDataList) { + System.out.println(PrintUtils.toCsv(pair, dnsMap)); + pwOff.println(PrintUtils.toCsv(pair, dnsMap)); + } + } else { // Non-TLS conversations + List packetList = TcpConversationUtils.extractPacketPairs(conv); + for (PcapPacketPair pair : packetList) { + System.out.println(PrintUtils.toCsv(pair, dnsMap)); + pwOff.println(PrintUtils.toCsv(pair, dnsMap)); + } + } + } + } + } + pwOff.close(); + */ +// // ================================================================================================ +// // <<< Some work-in-progress/explorative code that extracts a "representative" sequence >>> +// // +// // Currently need to know relevant hostname in advance :( +// String hostname = "events.tplinkra.com"; +//// String hostname = "rfe-us-west-1.dch.dlink.com"; +// // Conversations with 'hostname' for ON events. +// List onsForHostname = new ArrayList<>(); +// // Conversations with 'hostname' for OFF events. +// List offsForHostname = new ArrayList<>(); +// // "Unwrap" sequence groupings in ons/offs maps. +// ons.get(hostname).forEach((k,v) -> onsForHostname.addAll(v)); +// offs.get(hostname).forEach((k,v) -> offsForHostname.addAll(v)); +// +// +// Map> onsForHostnameGroupedByTlsAppDataSequence = TcpConversationUtils.groupConversationsByTlsApplicationDataPacketSequence(onsForHostname); +// +// +// // Extract representative sequence for ON and OFF by providing the list of conversations with +// // 'hostname' observed for each event type (the training data). +// SequenceExtraction seqExtraction = new SequenceExtraction(); +//// ExtractedSequence extractedSequenceForOn = seqExtraction.extract(onsForHostname); +//// ExtractedSequence extractedSequenceForOff = seqExtraction.extract(offsForHostname); +// +// ExtractedSequence extractedSequenceForOn = seqExtraction.extractByTlsAppData(onsForHostname); +// ExtractedSequence extractedSequenceForOff = seqExtraction.extractByTlsAppData(offsForHostname); +// +// // Let's check how many ONs align with OFFs and vice versa (that is, how many times an event is incorrectly +// // labeled). +// int onsLabeledAsOff = 0; +// Integer[] representativeOnSeq = TcpConversationUtils.getPacketLengthSequence(extractedSequenceForOn.getRepresentativeSequence()); +// Integer[] representativeOffSeq = TcpConversationUtils.getPacketLengthSequence(extractedSequenceForOff.getRepresentativeSequence()); +// SequenceAlignment seqAlg = seqExtraction.getAlignmentAlgorithm(); +// for (Conversation c : onsForHostname) { +// Integer[] onSeq = TcpConversationUtils.getPacketLengthSequence(c); +// if (seqAlg.calculateAlignment(representativeOffSeq, onSeq) <= extractedSequenceForOff.getMaxAlignmentCost()) { +// onsLabeledAsOff++; +// } +// } +// int offsLabeledAsOn = 0; +// for (Conversation c : offsForHostname) { +// Integer[] offSeq = TcpConversationUtils.getPacketLengthSequence(c); +// if (seqAlg.calculateAlignment(representativeOnSeq, offSeq) <= extractedSequenceForOn.getMaxAlignmentCost()) { +// offsLabeledAsOn++; +// } +// } +// System.out.println(""); +// // ================================================================================================ +// +// +// // ------------------------------------------------------------------------------------------------------------- +// // ------------------------------------------------------------------------------------------------------------- } } -// TP-Link MAC 50:c7:bf:33:1f:09 and usually IP 192.168.1.159 (remember to verify per file) \ No newline at end of file +// TP-Link MAC 50:c7:bf:33:1f:09 and usually IP 192.168.1.159 (remember to verify per file) +// frame.len >= 556 && frame.len <= 558 && ip.addr == 192.168.1.159 \ No newline at end of file