Merge branch 'master' of https://github.uci.edu/rtrimana/smart_home_traffic
[pingpong.git] / Code / Projects / SmartPlugDetector / src / main / java / edu / uci / iotproject / Main.java
1 package edu.uci.iotproject;
2
3 import static edu.uci.iotproject.analysis.UserAction.Type;
4
5 import edu.uci.iotproject.analysis.TcpConversationUtils;
6 import edu.uci.iotproject.analysis.TrafficLabeler;
7 import edu.uci.iotproject.analysis.TriggerTrafficExtractor;
8 import edu.uci.iotproject.analysis.UserAction;
9 import edu.uci.iotproject.io.TriggerTimesFileReader;
10 import org.pcap4j.core.*;
11 import org.pcap4j.packet.namednumber.DataLinkType;
12
13 import java.io.EOFException;
14 import java.net.UnknownHostException;
15 import java.time.Instant;
16 import java.util.*;
17 import java.util.concurrent.TimeoutException;
18
19 /**
20  * This is a system that reads PCAP files to compare
21  * patterns of DNS hostnames, packet sequences, and packet
22  * lengths with training data to determine certain events
23  * or actions for smart home devices.
24  *
25  * @author Janus Varmarken
26  * @author Rahmadi Trimananda (rtrimana@uci.edu)
27  * @version 0.1
28  */
29 public class Main {
30
31
32     public static void main(String[] args) throws PcapNativeException, NotOpenException, EOFException, TimeoutException, UnknownHostException {
33         // -------------------------------------------------------------------------------------------------------------
34         // ------------ # Code for extracting traffic generated by a device within x seconds of a trigger # ------------
35         // Paths to input and output files (consider supplying these as arguments instead) and IP of the device for
36         // which traffic is to be extracted:
37         String path = "/scratch/July-2018"; // Rahmadi
38         //String path = "/Users/varmarken/temp/UCI IoT Project/experiments"; // Janus
39
40         // D-Link July 26 experiment
41 //        final String inputPcapFile = path + "/2018-07/dlink/dlink.wlan1.local.pcap";
42 //        final String outputPcapFile = path + "/2018-07/dlink/dlink-processed.pcap";
43 //        final String triggerTimesFile = path + "/2018-07/dlink/dlink-july-26-2018.timestamps";
44 //        final String deviceIp = "192.168.1.246"; // .246 == phone; .199 == dlink plug?
45
46         // TP-Link July 25 experiment
47         final String inputPcapFile = path + "/2018-07/tplink/tplink.wlan1.local.pcap";
48         final String outputPcapFile = path + "/2018-07/tplink/tplink-processed.pcap";
49         final String triggerTimesFile = path + "/2018-07/tplink/tplink-july-25-2018.timestamps";
50         final String deviceIp = "192.168.1.159";
51
52         // SmartThings Plug July 25 experiment
53 //        final String inputPcapFile = path + "/2018-07/stplug/stplug.wlan1.local.pcap";
54 //        final String outputPcapFile = path + "/2018-07/stplug/stplug-processed.pcap";
55 //        final String triggerTimesFile = path + "/2018-07/stplug/smartthings-july-25-2018.timestamps";
56 //        final String deviceIp = "192.168.1.246"; // .246 == phone; .142 == SmartThings Hub (note: use eth0 capture for this!)
57
58         // Wemo July 30 experiment
59 //        final String inputPcapFile = path + "/2018-07/wemo/wemo.wlan1.local.pcap";
60 //        final String outputPcapFile = path + "/2018-07/wemo/wemo-processed.pcap";
61 //        final String triggerTimesFile = path + "/2018-07/wemo/wemo-july-30-2018.timestamps";
62 //        final String deviceIp = "192.168.1.145";
63
64         // Wemo Insight July 31 experiment
65 //        final String inputPcapFile = path + "/2018-07/wemoinsight/wemoinsight.wlan1.local.pcap";
66 //        final String outputPcapFile = path + "/2018-07/wemoinsight/wemoinsight-processed.pcap";
67 //        final String triggerTimesFile = path + "/2018-07/wemoinsight/wemo-insight-july-31-2018.timestamps";
68 //        final String deviceIp = "192.168.1.135";
69
70         // TP-Link BULB August 1 experiment
71 //        final String inputPcapFile = path + "/2018-08/tplink-bulb/tplinkbulb.wlan1.local.pcap";
72 //        final String outputPcapFile = path + "/2018-08/tplink-bulb/tplinkbulb-processed.pcap";
73 //        final String triggerTimesFile = path + "/2018-08/tplink-bulb/tplink-bulb-aug-3-2018.timestamps";
74 //        final String deviceIp = "192.168.1.140";
75
76         TriggerTimesFileReader ttfr = new TriggerTimesFileReader();
77         List<Instant> triggerTimes = ttfr.readTriggerTimes(triggerTimesFile, false);
78         // Tag each trigger with "ON" or "OFF", assuming that the first trigger is an "ON" and that they alternate.
79         List<UserAction> userActions = new ArrayList<>();
80         for (int i = 0; i < triggerTimes.size(); i++) {
81             userActions.add(new UserAction(i % 2 == 0 ? Type.TOGGLE_ON : Type.TOGGLE_OFF, triggerTimes.get(i)));
82         }
83         TriggerTrafficExtractor tte = new TriggerTrafficExtractor(inputPcapFile, triggerTimes, deviceIp);
84         final PcapDumper outputter = Pcaps.openDead(DataLinkType.EN10MB, 65536).dumpOpen(outputPcapFile);
85         DnsMap dnsMap = new DnsMap();
86         TcpReassembler tcpReassembler = new TcpReassembler();
87         TrafficLabeler trafficLabeler = new TrafficLabeler(userActions);
88         tte.performExtraction(pkt -> {
89             try {
90                 outputter.dump(pkt);
91             } catch (NotOpenException e) {
92                 e.printStackTrace();
93             }
94         }, dnsMap, tcpReassembler, trafficLabeler);
95         outputter.flush();
96         outputter.close();
97
98         if (tte.getPacketsIncludedCount() != trafficLabeler.getTotalPacketCount()) {
99             // Sanity/debug check
100             throw new AssertionError(String.format("mismatch between packet count in %s and %s",
101                     TriggerTrafficExtractor.class.getSimpleName(), TrafficLabeler.class.getSimpleName()));
102         }
103
104         // Extract all conversations present in the filtered trace.
105         List<Conversation> allConversations = tcpReassembler.getTcpConversations();
106         // Group conversations by hostname.
107         Map<String, List<Conversation>> convsByHostname = TcpConversationUtils.groupConversationsByHostname(allConversations, dnsMap);
108         System.out.println("Grouped conversations by hostname.");
109         // For each hostname, count the frequencies of packet lengths exchanged with that hostname.
110         final Map<String, Map<Integer, Integer>> pktLenFreqsByHostname = new HashMap<>();
111         convsByHostname.forEach((host, convs) -> pktLenFreqsByHostname.put(host, TcpConversationUtils.countPacketLengthFrequencies(convs)));
112         System.out.println("Counted frequencies of packet lengths exchanged with each hostname.");
113         // For each hostname, count the frequencies of packet sequences (i.e., count how many conversations exchange a
114         // sequence of packets of some specific lengths).
115         final Map<String, Map<String, Integer>> pktSeqFreqsByHostname = new HashMap<>();
116         convsByHostname.forEach((host, convs) -> pktSeqFreqsByHostname.put(host, TcpConversationUtils.countPacketSequenceFrequencies(convs)));
117         System.out.println("Counted frequencies of packet sequences exchanged with each hostname.");
118         // For each hostname, count frequencies of packet pairs exchanged with that hostname across all conversations
119         final Map<String, Map<String, Integer>> pktPairFreqsByHostname =
120                 TcpConversationUtils.countPacketPairFrequenciesByHostname(allConversations, dnsMap);
121         System.out.println("Counted frequencies of packet pairs per hostname");
122         // For each user action, reassemble the set of TCP connections occurring shortly after
123         final Map<UserAction, List<Conversation>> userActionToConversations = trafficLabeler.getLabeledReassembledTcpTraffic();
124         final Map<UserAction, Map<String, List<Conversation>>> userActionsToConvsByHostname = trafficLabeler.getLabeledReassembledTcpTraffic(dnsMap);
125         System.out.println("Reassembled TCP conversations occurring shortly after each user event");
126
127
128
129         // Contains all ON events: hostname -> sequence identifier -> list of conversations with that sequence
130         Map<String, Map<String, List<Conversation>>> ons = new HashMap<>();
131         // Contains all OFF events: hostname -> sequence identifier -> list of conversations with that sequence
132         Map<String, Map<String, List<Conversation>>> offs = new HashMap<>();
133         userActionsToConvsByHostname.forEach((ua, hostnameToConvs) -> {
134             Map<String, Map<String, List<Conversation>>> outer = ua.getType() == Type.TOGGLE_ON ? ons : offs;
135             hostnameToConvs.forEach((host, convs) -> {
136                 Map<String, List<Conversation>> seqsToConvs = TcpConversationUtils.
137                         groupConversationsByPacketSequence(convs);
138                 outer.merge(host, seqsToConvs, (oldMap, newMap) -> {
139                     newMap.forEach((sequence, cs) -> oldMap.merge(sequence, cs, (list1, list2) -> {
140                         list1.addAll(list2);
141                         return list1;
142                     }));
143                     return oldMap;
144                 });
145             });
146         });
147
148         System.out.println("");
149
150         // -------------------------------------------------------------------------------------------------------------
151         // -------------------------------------------------------------------------------------------------------------
152     }
153
154 }
155
156
157 // TP-Link MAC 50:c7:bf:33:1f:09 and usually IP 192.168.1.159 (remember to verify per file)
158 // frame.len >= 556 && frame.len <= 558 && ip.addr == 192.168.1.159