statistics.cpp 35 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807
  1. #include <iostream>
  2. #include <fstream>
  3. #include <vector>
  4. #include <math.h>
  5. #include "statistics.h"
  6. #include <sstream>
  7. #include <SQLiteCpp/SQLiteCpp.h>
  8. #include "statistics_db.h"
  9. #include "statistics.h"
  10. #include "utilities.h"
  11. using namespace Tins;
  12. /**
  13. * Checks if there is a payload and increments payloads counter.
  14. * @param pdu_l4 The packet that should be checked if it has a payload or not.
  15. */
  16. void statistics::checkPayload(const PDU *pdu_l4) {
  17. if(this->getDoExtraTests()) {
  18. // pdu_l4: Tarnsport layer 4
  19. int pktSize = pdu_l4->size();
  20. int headerSize = pdu_l4->header_size(); // TCP/UDP header
  21. int payloadSize = pktSize - headerSize;
  22. if (payloadSize > 0)
  23. payloadCount++;
  24. }
  25. }
  26. /**
  27. * Checks the correctness of TCP checksum and increments counter if the checksum was incorrect.
  28. * @param ipAddressSender The source IP.
  29. * @param ipAddressReceiver The destination IP.
  30. * @param tcpPkt The packet to get checked.
  31. */
  32. void statistics::checkTCPChecksum(const std::string &ipAddressSender, const std::string &ipAddressReceiver, TCP tcpPkt) {
  33. if(this->getDoExtraTests()) {
  34. if(check_tcpChecksum(ipAddressSender, ipAddressReceiver, tcpPkt))
  35. correctTCPChecksumCount++;
  36. else incorrectTCPChecksumCount++;
  37. }
  38. }
  39. /**
  40. * Calculates entropy of the source and destination IPs in a time interval.
  41. * @param intervalStartTimestamp The timstamp where the interval starts.
  42. * @return a vector: contains source IP entropy and destination IP entropy.
  43. */
  44. std::vector<float> statistics::calculateLastIntervalIPsEntropy(std::chrono::microseconds intervalStartTimestamp){
  45. if(this->getDoExtraTests()) {
  46. // TODO: change datastructures
  47. std::vector<int> IPsSrcPktsCounts;
  48. std::vector<int> IPsDstPktsCounts;
  49. std::vector<double> IPsSrcProb;
  50. std::vector<double> IPsDstProb;
  51. int pktsSent = 0, pktsReceived = 0;
  52. for (auto i = ip_statistics.begin(); i != ip_statistics.end(); i++) {
  53. int IPsSrcPktsCount = 0;
  54. // TODO: iter backwards and break
  55. for (auto j = i->second.pkts_sent_timestamp.begin(); j != i->second.pkts_sent_timestamp.end(); j++) {
  56. if(*j >= intervalStartTimestamp)
  57. IPsSrcPktsCount++;
  58. }
  59. if(IPsSrcPktsCount != 0) {
  60. IPsSrcPktsCounts.push_back(IPsSrcPktsCount);
  61. pktsSent += IPsSrcPktsCount;
  62. }
  63. int IPsDstPktsCount = 0;
  64. for (auto j = i->second.pkts_received_timestamp.begin(); j != i->second.pkts_received_timestamp.end(); j++) {
  65. if(*j >= intervalStartTimestamp)
  66. IPsDstPktsCount++;
  67. }
  68. if(IPsDstPktsCount != 0) {
  69. IPsDstPktsCounts.push_back(IPsDstPktsCount);
  70. pktsReceived += IPsDstPktsCount;
  71. }
  72. }
  73. for (auto i = IPsSrcPktsCounts.begin(); i != IPsSrcPktsCounts.end(); i++) {
  74. IPsSrcProb.push_back(static_cast<double>(*i) / static_cast<double>(pktsSent));
  75. }
  76. for (auto i = IPsDstPktsCounts.begin(); i != IPsDstPktsCounts.end(); i++) {
  77. IPsDstProb.push_back(static_cast<double>(*i) / static_cast<double>(pktsReceived));
  78. }
  79. // Calculate IP source entropy
  80. double IPsSrcEntropy = 0;
  81. for (unsigned i = 0; i < IPsSrcProb.size(); i++) {
  82. if (IPsSrcProb[i] > 0)
  83. IPsSrcEntropy += -IPsSrcProb[i] * log2(IPsSrcProb[i]);
  84. }
  85. // Calculate IP destination entropy
  86. double IPsDstEntropy = 0;
  87. for (unsigned i = 0; i < IPsDstProb.size(); i++) {
  88. if (IPsDstProb[i] > 0)
  89. IPsDstEntropy += -IPsDstProb[i] * log2(IPsDstProb[i]);
  90. }
  91. // FIXME: return doubles not floats
  92. std::vector<float> entropies = {static_cast<float>(IPsSrcEntropy), static_cast<float>(IPsDstEntropy)};
  93. return entropies;
  94. }
  95. else {
  96. return {-1, -1};
  97. }
  98. }
  99. /**
  100. * Calculates the cumulative entropy of the source and destination IPs, i.e., the entropy for packets from the beginning of the pcap file.
  101. * @return a vector: contains the cumulative entropies of source and destination IPs
  102. */
  103. std::vector<float> statistics::calculateIPsCumEntropy(){
  104. if(this->getDoExtraTests()) {
  105. std::vector <std::string> IPs;
  106. std::vector <float> IPsSrcProb;
  107. std::vector <float> IPsDstProb;
  108. for (auto i = ip_statistics.begin(); i != ip_statistics.end(); i++) {
  109. IPs.push_back(i->first);
  110. IPsSrcProb.push_back((float)i->second.pkts_sent/packetCount);
  111. IPsDstProb.push_back((float)i->second.pkts_received/packetCount);
  112. }
  113. // Calculate IP source entropy
  114. float IPsSrcEntropy = 0;
  115. for(unsigned i=0; i < IPsSrcProb.size();i++){
  116. if (IPsSrcProb[i] > 0)
  117. IPsSrcEntropy += - IPsSrcProb[i]*log2(IPsSrcProb[i]);
  118. }
  119. // Calculate IP destination entropy
  120. float IPsDstEntropy = 0;
  121. for(unsigned i=0; i < IPsDstProb.size();i++){
  122. if (IPsDstProb[i] > 0)
  123. IPsDstEntropy += - IPsDstProb[i]*log2(IPsDstProb[i]);
  124. }
  125. std::vector<float> entropies = {IPsSrcEntropy, IPsDstEntropy};
  126. return entropies;
  127. }
  128. else {
  129. return {-1, -1};
  130. }
  131. }
  132. /**
  133. * Calculates sending packet rate for each IP in a time interval. Finds min and max packet rate and adds them to ip_statistics map.
  134. * @param intervalStartTimestamp The timstamp where the interval starts.
  135. */
  136. void statistics::calculateIPIntervalPacketRate(std::chrono::duration<int, std::micro> interval, std::chrono::microseconds intervalStartTimestamp){
  137. for (auto i = ip_statistics.begin(); i != ip_statistics.end(); i++) {
  138. int IPsSrcPktsCount = 0;
  139. for (auto j = i->second.pkts_sent_timestamp.begin(); j != i->second.pkts_sent_timestamp.end(); j++) {
  140. if(*j >= intervalStartTimestamp)
  141. IPsSrcPktsCount++;
  142. }
  143. float interval_pkt_rate = (float) IPsSrcPktsCount * 1000000 / interval.count(); // used 10^6 because interval in microseconds
  144. i->second.interval_pkt_rate.push_back(interval_pkt_rate);
  145. if(interval_pkt_rate > i->second.max_interval_pkt_rate || i->second.max_interval_pkt_rate == 0)
  146. i->second.max_interval_pkt_rate = interval_pkt_rate;
  147. if(interval_pkt_rate < i->second.min_interval_pkt_rate || i->second.min_interval_pkt_rate == 0)
  148. i->second.min_interval_pkt_rate = interval_pkt_rate;
  149. }
  150. }
  151. /**
  152. * Registers statistical data for a time interval.
  153. * @param intervalStartTimestamp The timstamp where the interval starts.
  154. * @param intervalEndTimestamp The timstamp where the interval ends.
  155. * @param previousPacketCount The total number of packets in last interval.
  156. */
  157. void statistics::addIntervalStat(std::chrono::duration<int, std::micro> interval, std::chrono::microseconds intervalStartTimestamp, std::chrono::microseconds intervalEndTimestamp){
  158. // Add packet rate for each IP to ip_statistics map
  159. calculateIPIntervalPacketRate(interval, intervalStartTimestamp);
  160. std::vector<float> ipEntopies = calculateLastIntervalIPsEntropy(intervalStartTimestamp);
  161. std::vector<float> ipCumEntopies = calculateIPsCumEntropy();
  162. std::string lastPktTimestamp_s = std::to_string(intervalEndTimestamp.count());
  163. std::string intervalStartTimestamp_s = std::to_string(intervalStartTimestamp.count());
  164. // The intervalStartTimestamp_s is the previous interval lastPktTimestamp_s
  165. // TODO: check with carlos if first and last packet timestamps are alright
  166. interval_statistics[lastPktTimestamp_s].start = std::to_string(intervalStartTimestamp.count());
  167. interval_statistics[lastPktTimestamp_s].end = std::to_string(intervalEndTimestamp.count());
  168. interval_statistics[lastPktTimestamp_s].pkts_count = packetCount - intervalCumPktCount;
  169. interval_statistics[lastPktTimestamp_s].pkt_rate = static_cast<float>(interval_statistics[lastPktTimestamp_s].pkts_count) / (static_cast<double>(interval.count()) / 1000000);
  170. interval_statistics[lastPktTimestamp_s].kbytes = static_cast<float>(sumPacketSize - intervalCumSumPktSize) / 1024;
  171. interval_statistics[lastPktTimestamp_s].kbyte_rate = interval_statistics[lastPktTimestamp_s].kbytes / (static_cast<double>(interval.count()) / 1000000);
  172. interval_statistics[lastPktTimestamp_s].payload_count = payloadCount - intervalPayloadCount;
  173. interval_statistics[lastPktTimestamp_s].incorrect_tcp_checksum_count = incorrectTCPChecksumCount - intervalIncorrectTCPChecksumCount;
  174. interval_statistics[lastPktTimestamp_s].correct_tcp_checksum_count = correctTCPChecksumCount - intervalCorrectTCPChecksumCount;
  175. interval_statistics[lastPktTimestamp_s].novel_ip_count = ip_statistics.size() - intervalCumNovelIPCount;
  176. interval_statistics[lastPktTimestamp_s].novel_ttl_count = ttl_values.size() - intervalCumNovelTTLCount;
  177. interval_statistics[lastPktTimestamp_s].novel_win_size_count = win_values.size() - intervalCumNovelWinSizeCount;
  178. interval_statistics[lastPktTimestamp_s].novel_tos_count = tos_values.size() - intervalCumNovelToSCount;
  179. interval_statistics[lastPktTimestamp_s].novel_mss_count = mss_values.size() - intervalCumNovelMSSCount;
  180. interval_statistics[lastPktTimestamp_s].novel_port_count = port_values.size() - intervalCumNovelPortCount;
  181. intervalPayloadCount = payloadCount;
  182. intervalIncorrectTCPChecksumCount = incorrectTCPChecksumCount;
  183. intervalCorrectTCPChecksumCount = correctTCPChecksumCount;
  184. intervalCumPktCount = packetCount;
  185. intervalCumSumPktSize = sumPacketSize;
  186. intervalCumNovelIPCount = ip_statistics.size();
  187. intervalCumNovelTTLCount = ttl_values.size();
  188. intervalCumNovelWinSizeCount = win_values.size();
  189. intervalCumNovelToSCount = tos_values.size();
  190. intervalCumNovelMSSCount = mss_values.size();
  191. intervalCumNovelPortCount = port_values.size();
  192. if(ipEntopies.size()>1){
  193. interval_statistics[lastPktTimestamp_s].ip_src_entropy = ipEntopies[0];
  194. interval_statistics[lastPktTimestamp_s].ip_dst_entropy = ipEntopies[1];
  195. }
  196. if(ipCumEntopies.size()>1){
  197. interval_statistics[lastPktTimestamp_s].ip_src_cum_entropy = ipCumEntopies[0];
  198. interval_statistics[lastPktTimestamp_s].ip_dst_cum_entropy = ipCumEntopies[1];
  199. }
  200. }
  201. /**
  202. * Registers statistical data for a sent packet in a given conversation (two IPs, two ports).
  203. * Increments the counter packets_A_B or packets_B_A.
  204. * Adds the timestamp of the packet in pkts_A_B_timestamp or pkts_B_A_timestamp.
  205. * @param ipAddressSender The sender IP address.
  206. * @param sport The source port.
  207. * @param ipAddressReceiver The receiver IP address.
  208. * @param dport The destination port.
  209. * @param timestamp The timestamp of the packet.
  210. */
  211. void statistics::addConvStat(const std::string &ipAddressSender,int sport,const std::string &ipAddressReceiver,int dport, std::chrono::microseconds timestamp){
  212. conv f1 = {ipAddressReceiver, dport, ipAddressSender, sport};
  213. conv f2 = {ipAddressSender, sport, ipAddressReceiver, dport};
  214. // if already exist A(ipAddressReceiver, dport), B(ipAddressSender, sport) conversation
  215. if (conv_statistics.count(f1)>0){
  216. conv_statistics[f1].pkts_count++;
  217. if(conv_statistics[f1].pkts_count<=3)
  218. conv_statistics[f1].interarrival_time.push_back(std::chrono::duration_cast<std::chrono::microseconds> (timestamp - conv_statistics[f1].pkts_timestamp.back()));
  219. conv_statistics[f1].pkts_timestamp.push_back(timestamp);
  220. }
  221. // Add new conversation A(ipAddressSender, sport), B(ipAddressReceiver, dport)
  222. else{
  223. conv_statistics[f2].pkts_count++;
  224. if(conv_statistics[f2].pkts_timestamp.size()>0 && conv_statistics[f2].pkts_count<=3 )
  225. conv_statistics[f2].interarrival_time.push_back(std::chrono::duration_cast<std::chrono::microseconds> (timestamp - conv_statistics[f2].pkts_timestamp.back()));
  226. conv_statistics[f2].pkts_timestamp.push_back(timestamp);
  227. }
  228. }
  229. /**
  230. * Registers statistical data for a sent packet in a given extended conversation (two IPs, two ports, protocol).
  231. * Increments the counter packets_A_B or packets_B_A.
  232. * Adds the timestamp of the packet in pkts_A_B_timestamp or pkts_B_A_timestamp.
  233. * Updates all other statistics of conv_statistics_extended
  234. * @param ipAddressSender The sender IP address.
  235. * @param sport The source port.
  236. * @param ipAddressReceiver The receiver IP address.
  237. * @param dport The destination port.
  238. * @param protocol The used protocol.
  239. * @param timestamp The timestamp of the packet.
  240. */
  241. void statistics::addConvStatExt(const std::string &ipAddressSender,int sport,const std::string &ipAddressReceiver,int dport,const std::string &protocol, std::chrono::microseconds timestamp){
  242. if(this->getDoExtraTests()) {
  243. convWithProt f1 = {ipAddressReceiver, dport, ipAddressSender, sport, protocol};
  244. convWithProt f2 = {ipAddressSender, sport, ipAddressReceiver, dport, protocol};
  245. convWithProt f;
  246. // if there already exists a communication interval for the specified conversation
  247. if (conv_statistics_extended.count(f1) > 0 || conv_statistics_extended.count(f2) > 0){
  248. // find out which direction of conversation is contained in conv_statistics_extended
  249. if (conv_statistics_extended.count(f1) > 0)
  250. f = f1;
  251. else
  252. f = f2;
  253. // increase pkts count and check on delay
  254. conv_statistics_extended[f].pkts_count++;
  255. if (conv_statistics_extended[f].pkts_timestamp.size()>0 && conv_statistics_extended[f].pkts_count<=3)
  256. conv_statistics_extended[f].interarrival_time.push_back(std::chrono::duration_cast<std::chrono::microseconds> (timestamp - conv_statistics_extended[f].pkts_timestamp.back()));
  257. conv_statistics_extended[f].pkts_timestamp.push_back(timestamp);
  258. // if the time difference has exceeded the threshold, create a new interval with this message
  259. if (timestamp - conv_statistics_extended[f].comm_intervals.back().end > (std::chrono::microseconds) ((unsigned long) COMM_INTERVAL_THRESHOLD)) { // > or >= ?
  260. commInterval new_interval = {timestamp, timestamp, 1};
  261. conv_statistics_extended[f].comm_intervals.push_back(new_interval);
  262. }
  263. // otherwise, set the time of the last interval message to the current timestamp and increase interval packet count by 1
  264. else{
  265. conv_statistics_extended[f].comm_intervals.back().end = timestamp;
  266. conv_statistics_extended[f].comm_intervals.back().pkts_count++;
  267. }
  268. }
  269. // if there does not exist a communication interval for the specified conversation
  270. else{
  271. // add initial interval entry for this conversation
  272. commInterval initial_interval = {timestamp, timestamp, 1};
  273. entry_convStatExt entry;
  274. entry.comm_intervals.push_back(initial_interval);
  275. entry.pkts_count = 1;
  276. entry.pkts_timestamp.push_back(timestamp);
  277. conv_statistics_extended[f2] = entry;
  278. }
  279. }
  280. }
  281. /**
  282. * Aggregate the collected information about all communication intervals within conv_statistics_extended of every conversation.
  283. * Do this by computing the average packet rate per interval and the average time between intervals.
  284. * Also compute average interval duration and total communication duration (i.e. last_msg.time - first_msg.time)
  285. */
  286. void statistics::createCommIntervalStats(){
  287. // iterate over all <convWithProt, entry_convStatExt> pairs
  288. for (auto &cur_elem : conv_statistics_extended) {
  289. entry_convStatExt &entry = cur_elem.second;
  290. std::vector<commInterval> &intervals = entry.comm_intervals;
  291. // if there is only one interval, the time between intervals cannot be computed and is therefore set to 0
  292. if (intervals.size() == 1){
  293. double interval_duration = (double) (intervals[0].end - intervals[0].start).count() / (double) 1e6;
  294. entry.avg_int_pkts_count = (double) intervals[0].pkts_count;
  295. entry.avg_time_between_ints = (double) 0;
  296. entry.avg_interval_time = interval_duration;
  297. }
  298. // If there is more than one interval, compute the specified averages
  299. else if (intervals.size() > 1){
  300. long summed_pkts_count = intervals[0].pkts_count;
  301. std::chrono::microseconds time_between_ints_sum = (std::chrono::microseconds) 0;
  302. std::chrono::microseconds summed_int_duration = intervals[0].end - intervals[0].start;
  303. for (std::size_t i = 1; i < intervals.size(); i++) {
  304. summed_pkts_count += intervals[i].pkts_count;
  305. summed_int_duration += intervals[i].end - intervals[i].start;
  306. time_between_ints_sum += intervals[i].start - intervals[i - 1].end;
  307. }
  308. entry.avg_int_pkts_count = summed_pkts_count / ((double) intervals.size());
  309. entry.avg_time_between_ints = (time_between_ints_sum.count() / (double) (intervals.size() - 1)) / (double) 1e6;
  310. entry.avg_interval_time = (summed_int_duration.count() / (double) intervals.size()) / (double) 1e6;
  311. }
  312. entry.total_comm_duration = (double) (entry.pkts_timestamp.back() - entry.pkts_timestamp.front()).count() / (double) 1e6;
  313. }
  314. }
  315. /**
  316. * Increments the packet counter for the given IP address and MSS value.
  317. * @param ipAddress The IP address whose MSS packet counter should be incremented.
  318. * @param mssValue The MSS value of the packet.
  319. */
  320. void statistics::incrementMSScount(const std::string &ipAddress, int mssValue) {
  321. mss_values[mssValue]++;
  322. mss_distribution[{ipAddress, mssValue}]++;
  323. }
  324. /**
  325. * Increments the packet counter for the given IP address and window size.
  326. * @param ipAddress The IP address whose window size packet counter should be incremented.
  327. * @param winSize The window size of the packet.
  328. */
  329. void statistics::incrementWinCount(const std::string &ipAddress, int winSize) {
  330. win_values[winSize]++;
  331. win_distribution[{ipAddress, winSize}]++;
  332. }
  333. /**
  334. * Increments the packet counter for the given IP address and TTL value.
  335. * @param ipAddress The IP address whose TTL packet counter should be incremented.
  336. * @param ttlValue The TTL value of the packet.
  337. */
  338. void statistics::incrementTTLcount(const std::string &ipAddress, int ttlValue) {
  339. ttl_values[ttlValue]++;
  340. ttl_distribution[{ipAddress, ttlValue}]++;
  341. }
  342. /**
  343. * Increments the packet counter for the given IP address and ToS value.
  344. * @param ipAddress The IP address whose ToS packet counter should be incremented.
  345. * @param tosValue The ToS value of the packet.
  346. */
  347. void statistics::incrementToScount(const std::string &ipAddress, int tosValue) {
  348. tos_values[tosValue]++;
  349. tos_distribution[{ipAddress, tosValue}]++;
  350. }
  351. /**
  352. * Increments the protocol counter for the given IP address and protocol.
  353. * @param ipAddress The IP address whose protocol packet counter should be incremented.
  354. * @param protocol The protocol of the packet.
  355. */
  356. void statistics::incrementProtocolCount(const std::string &ipAddress, const std::string &protocol) {
  357. protocol_distribution[{ipAddress, protocol}].count++;
  358. }
  359. /**
  360. * Returns the number of packets seen for the given IP address and protocol.
  361. * @param ipAddress The IP address whose packet count is wanted.
  362. * @param protocol The protocol whose packet count is wanted.
  363. */
  364. int statistics::getProtocolCount(const std::string &ipAddress, const std::string &protocol) {
  365. return protocol_distribution[{ipAddress, protocol}].count;
  366. }
  367. /**
  368. * Increases the byte counter for the given IP address and protocol.
  369. * @param ipAddress The IP address whose protocol byte counter should be increased.
  370. * @param protocol The protocol of the packet.
  371. * @param byteSent The packet's size.
  372. */
  373. void statistics::increaseProtocolByteCount(const std::string &ipAddress, const std::string &protocol, long bytesSent) {
  374. protocol_distribution[{ipAddress, protocol}].byteCount += bytesSent;
  375. }
  376. /**
  377. * Returns the number of bytes seen for the given IP address and protocol.
  378. * @param ipAddress The IP address whose byte count is wanted.
  379. * @param protocol The protocol whose byte count is wanted.
  380. * @return a float: The number of bytes
  381. */
  382. float statistics::getProtocolByteCount(const std::string &ipAddress, const std::string &protocol) {
  383. return protocol_distribution[{ipAddress, protocol}].byteCount;
  384. }
  385. /**
  386. * Increments the packet counter for
  387. * - the given sender IP address with outgoing port and
  388. * - the given receiver IP address with incoming port.
  389. * @param ipAddressSender The IP address of the packet sender.
  390. * @param outgoingPort The port used by the sender.
  391. * @param ipAddressReceiver The IP address of the packet receiver.
  392. * @param incomingPort The port used by the receiver.
  393. */
  394. void statistics::incrementPortCount(const std::string &ipAddressSender, int outgoingPort, const std::string &ipAddressReceiver,
  395. int incomingPort, const std::string &protocol) {
  396. port_values[outgoingPort]++;
  397. port_values[incomingPort]++;
  398. ip_ports[{ipAddressSender, "out", outgoingPort, protocol}].count++;
  399. ip_ports[{ipAddressReceiver, "in", incomingPort, protocol}].count++;
  400. }
  401. /**
  402. * Increases the packet byte counter for
  403. * - the given sender IP address with outgoing port and
  404. * - the given receiver IP address with incoming port.
  405. * @param ipAddressSender The IP address of the packet sender.
  406. * @param outgoingPort The port used by the sender.
  407. * @param ipAddressReceiver The IP address of the packet receiver.
  408. * @param incomingPort The port used by the receiver.
  409. * @param byteSent The packet's size.
  410. */
  411. void statistics::increasePortByteCount(const std::string &ipAddressSender, int outgoingPort, const std::string &ipAddressReceiver,
  412. int incomingPort, long bytesSent, const std::string &protocol) {
  413. ip_ports[{ipAddressSender, "out", outgoingPort, protocol}].byteCount += bytesSent;
  414. ip_ports[{ipAddressReceiver, "in", incomingPort, protocol}].byteCount += bytesSent;
  415. }
  416. /**
  417. * Increments the packet counter for
  418. * - the given sender MAC address and
  419. * - the given receiver MAC address.
  420. * @param srcMac The MAC address of the packet sender.
  421. * @param dstMac The MAC address of the packet receiver.
  422. * @param typeNumber The payload type number of the packet.
  423. */
  424. void statistics::incrementUnrecognizedPDUCount(const std::string &srcMac, const std::string &dstMac, uint32_t typeNumber,
  425. const std::string &timestamp) {
  426. unrecognized_PDUs[{srcMac, dstMac, typeNumber}].count++;
  427. unrecognized_PDUs[{srcMac, dstMac, typeNumber}].timestamp_last_occurrence = timestamp;
  428. }
  429. /**
  430. * Creates a new statistics object.
  431. */
  432. statistics::statistics(std::string resourcePath) {;
  433. this->resourcePath = resourcePath;
  434. }
  435. /**
  436. * Stores the assignment IP address -> MAC address.
  437. * @param ipAddress The IP address belonging to the given MAC address.
  438. * @param macAddress The MAC address belonging to the given IP address.
  439. */
  440. void statistics::assignMacAddress(const std::string &ipAddress, const std::string &macAddress) {
  441. ip_mac_mapping[ipAddress] = macAddress;
  442. }
  443. /**
  444. * Registers statistical data for a sent packet. Increments the counter packets_sent for the sender and
  445. * packets_received for the receiver. Adds the bytes as kbytes_sent (sender) and kybtes_received (receiver).
  446. * @param ipAddressSender The IP address of the packet sender.
  447. * @param ipAddressReceiver The IP address of the packet receiver.
  448. * @param bytesSent The packet's size.
  449. */
  450. void statistics::addIpStat_packetSent(const std::string &ipAddressSender, const std::string &ipAddressReceiver, long bytesSent, std::chrono::microseconds timestamp) {
  451. // Adding IP as a sender for first time
  452. if(ip_statistics[ipAddressSender].pkts_sent==0){
  453. // Add the IP class
  454. ip_statistics[ipAddressSender].ip_class = getIPv4Class(ipAddressSender);
  455. }
  456. // Adding IP as a receiver for first time
  457. if(ip_statistics[ipAddressReceiver].pkts_received==0){
  458. // Add the IP class
  459. ip_statistics[ipAddressReceiver].ip_class = getIPv4Class(ipAddressReceiver);
  460. }
  461. // Update stats for packet sender
  462. ip_statistics[ipAddressSender].kbytes_sent += (float(bytesSent) / 1024);
  463. ip_statistics[ipAddressSender].pkts_sent++;
  464. ip_statistics[ipAddressSender].pkts_sent_timestamp.push_back(timestamp);
  465. // Update stats for packet receiver
  466. ip_statistics[ipAddressReceiver].kbytes_received += (float(bytesSent) / 1024);
  467. ip_statistics[ipAddressReceiver].pkts_received++;
  468. ip_statistics[ipAddressReceiver].pkts_received_timestamp.push_back(timestamp);
  469. if(this->getDoExtraTests()) {
  470. // Increment Degrees for sender and receiver, if Sender sends its first packet to this receiver
  471. std::unordered_set<std::string>::const_iterator found_receiver = contacted_ips[ipAddressSender].find(ipAddressReceiver);
  472. if(found_receiver == contacted_ips[ipAddressSender].end()){
  473. // Receiver is NOT contained in the List of IPs, that the Sender has contacted, therefore this is the first packet in this direction
  474. ip_statistics[ipAddressSender].out_degree++;
  475. ip_statistics[ipAddressReceiver].in_degree++;
  476. // Increment overall_degree only if this is the first packet for the connection (both directions)
  477. // Therefore check, whether Receiver has contacted Sender before
  478. std::unordered_set<std::string>::const_iterator sender_contacted = contacted_ips[ipAddressReceiver].find(ipAddressSender);
  479. if(sender_contacted == contacted_ips[ipAddressReceiver].end()){
  480. ip_statistics[ipAddressSender].overall_degree++;
  481. ip_statistics[ipAddressReceiver].overall_degree++;
  482. }
  483. contacted_ips[ipAddressSender].insert(ipAddressReceiver);
  484. }
  485. }
  486. }
  487. /**
  488. * Setter for the timestamp_firstPacket field.
  489. * @param ts The timestamp of the first packet in the PCAP file.
  490. */
  491. void statistics::setTimestampFirstPacket(Tins::Timestamp ts) {
  492. timestamp_firstPacket = ts;
  493. }
  494. /**
  495. * Setter for the timestamp_lastPacket field.
  496. * @param ts The timestamp of the last packet in the PCAP file.
  497. */
  498. void statistics::setTimestampLastPacket(Tins::Timestamp ts) {
  499. timestamp_lastPacket = ts;
  500. }
  501. /**
  502. * Getter for the timestamp_firstPacket field.
  503. */
  504. Tins::Timestamp statistics::getTimestampFirstPacket() {
  505. return timestamp_firstPacket;
  506. }
  507. /**
  508. * Getter for the timestamp_lastPacket field.
  509. */
  510. Tins::Timestamp statistics::getTimestampLastPacket() {
  511. return timestamp_lastPacket;
  512. }
  513. /**
  514. * Getter for the packetCount field.
  515. */
  516. int statistics::getPacketCount() {
  517. return packetCount;
  518. }
  519. /**
  520. * Getter for the sumPacketSize field.
  521. */
  522. int statistics::getSumPacketSize() {
  523. return sumPacketSize;
  524. }
  525. /**
  526. * Returns the average packet size.
  527. * @return a float indicating the average packet size in kbytes.
  528. */
  529. float statistics::getAvgPacketSize() const {
  530. // AvgPktSize = (Sum of all packet sizes / #Packets)
  531. return (sumPacketSize / packetCount) / 1024;
  532. }
  533. /**
  534. * Adds the size of a packet (to be used to calculate the avg. packet size).
  535. * @param packetSize The size of the current packet in bytes.
  536. */
  537. void statistics::addPacketSize(uint32_t packetSize) {
  538. sumPacketSize += ((float) packetSize);
  539. }
  540. /**
  541. * Setter for the doExtraTests field.
  542. */
  543. void statistics::setDoExtraTests(bool var) {
  544. doExtraTests = var;
  545. }
  546. /**
  547. * Getter for the doExtraTests field.
  548. */
  549. bool statistics::getDoExtraTests() {
  550. return doExtraTests;
  551. }
  552. /**
  553. * Calculates the capture duration.
  554. * @return a formatted string HH:MM:SS.mmmmmm with
  555. * HH: hour, MM: minute, SS: second, mmmmmm: microseconds
  556. */
  557. std::string statistics::getCaptureDurationTimestamp() const {
  558. // Calculate duration
  559. timeval fp, lp, d;
  560. fp.tv_sec = timestamp_firstPacket.seconds();
  561. fp.tv_usec = timestamp_firstPacket.microseconds();
  562. lp.tv_sec = timestamp_lastPacket.seconds();
  563. lp.tv_usec = timestamp_lastPacket.microseconds();
  564. timersub(&lp, &fp, &d);
  565. long int hour = d.tv_sec / 3600;
  566. long int remainder = (d.tv_sec - hour * 3600);
  567. long int minute = remainder / 60;
  568. long int second = (remainder - minute * 60) % 60;
  569. long int microseconds = d.tv_usec;
  570. // Build desired output format: YYYY-mm-dd hh:mm:ss
  571. char out[64];
  572. sprintf(out, "%02ld:%02ld:%02ld.%06ld ", hour, minute, second, microseconds);
  573. return std::string(out);
  574. }
  575. /**
  576. * Calculates the capture duration.
  577. * @return a formatted string SS.mmmmmm with
  578. * S: seconds (UNIX time), mmmmmm: microseconds
  579. */
  580. float statistics::getCaptureDurationSeconds() const {
  581. timeval fp, lp, d;
  582. fp.tv_sec = timestamp_firstPacket.seconds();
  583. fp.tv_usec = timestamp_firstPacket.microseconds();
  584. lp.tv_sec = timestamp_lastPacket.seconds();
  585. lp.tv_usec = timestamp_lastPacket.microseconds();
  586. timersub(&lp, &fp, &d);
  587. char buf[64];
  588. snprintf(buf, sizeof(buf), "%u.%06u", static_cast<uint>(d.tv_sec), static_cast<uint>(d.tv_usec));
  589. return std::stof(std::string(buf));
  590. }
  591. /**
  592. * Creates a timestamp based on a time_t seconds (UNIX time format) and microseconds.
  593. * @param seconds
  594. * @param microseconds
  595. * @return a formatted string Y-m-d H:M:S.m with
  596. * Y: year, m: month, d: day, H: hour, M: minute, S: second, m: microseconds
  597. */
  598. std::string statistics::getFormattedTimestamp(time_t seconds, suseconds_t microseconds) const {
  599. timeval tv;
  600. tv.tv_sec = seconds;
  601. tv.tv_usec = microseconds;
  602. char tmbuf[20], buf[64];
  603. auto nowtm = gmtime(&(tv.tv_sec));
  604. strftime(tmbuf, sizeof(tmbuf), "%Y-%m-%d %H:%M:%S", nowtm);
  605. snprintf(buf, sizeof(buf), "%s.%06u", tmbuf, static_cast<uint>(tv.tv_usec));
  606. return std::string(buf);
  607. }
  608. /**
  609. * Calculates the statistics for a given IP address.
  610. * @param ipAddress The IP address whose statistics should be calculated.
  611. * @return a ip_stats struct containing statistical data derived by the statistical data collected.
  612. */
  613. ip_stats statistics::getStatsForIP(const std::string &ipAddress) {
  614. float duration = getCaptureDurationSeconds();
  615. entry_ipStat ipStatEntry = ip_statistics[ipAddress];
  616. ip_stats s;
  617. s.bandwidthKBitsIn = (ipStatEntry.kbytes_received / duration) * 8;
  618. s.bandwidthKBitsOut = (ipStatEntry.kbytes_sent / duration) * 8;
  619. s.packetPerSecondIn = (ipStatEntry.pkts_received / duration);
  620. s.packetPerSecondOut = (ipStatEntry.pkts_sent / duration);
  621. s.AvgPacketSizeSent = (ipStatEntry.kbytes_sent / ipStatEntry.pkts_sent);
  622. s.AvgPacketSizeRecv = (ipStatEntry.kbytes_received / ipStatEntry.pkts_received);
  623. return s;
  624. }
  625. int statistics::getDefaultInterval() {
  626. return this->default_interval;
  627. }
  628. void statistics::setDefaultInterval(int interval) {
  629. this->default_interval = interval;
  630. }
  631. /**
  632. * Increments the packet counter.
  633. */
  634. void statistics::incrementPacketCount() {
  635. packetCount++;
  636. }
  637. /**
  638. * Prints the statistics of the PCAP and IP specific statistics for the given IP address.
  639. * @param ipAddress The IP address whose statistics should be printed. Can be empty "" to print only general file statistics.
  640. */
  641. void statistics::printStats(const std::string &ipAddress) {
  642. std::stringstream ss;
  643. ss << std::endl;
  644. ss << "Capture duration: " << getCaptureDurationSeconds() << " seconds" << std::endl;
  645. ss << "Capture duration (HH:MM:SS.mmmmmm): " << getCaptureDurationTimestamp() << std::endl;
  646. ss << "#Packets: " << packetCount << std::endl;
  647. ss << std::endl;
  648. // Print IP address specific statistics only if IP address was given
  649. if (ipAddress != "") {
  650. entry_ipStat e = ip_statistics[ipAddress];
  651. ss << "\n----- STATS FOR IP ADDRESS [" << ipAddress << "] -------" << std::endl;
  652. ss << std::endl << "KBytes sent: " << e.kbytes_sent << std::endl;
  653. ss << "KBytes received: " << e.kbytes_received << std::endl;
  654. ss << "Packets sent: " << e.pkts_sent << std::endl;
  655. ss << "Packets received: " << e.pkts_received << "\n\n";
  656. ip_stats is = getStatsForIP(ipAddress);
  657. ss << "Bandwidth IN: " << is.bandwidthKBitsIn << " kbit/s" << std::endl;
  658. ss << "Bandwidth OUT: " << is.bandwidthKBitsOut << " kbit/s" << std::endl;
  659. ss << "Packets per second IN: " << is.packetPerSecondIn << std::endl;
  660. ss << "Packets per second OUT: " << is.packetPerSecondOut << std::endl;
  661. ss << "Avg Packet Size Sent: " << is.AvgPacketSizeSent << " kbytes" << std::endl;
  662. ss << "Avg Packet Size Received: " << is.AvgPacketSizeRecv << " kbytes" << std::endl;
  663. }
  664. std::cout << ss.str();
  665. }
  666. /**
  667. * Derives general PCAP file statistics from the collected statistical data and
  668. * writes all data into a SQLite database, located at database_path.
  669. * @param database_path The path of the SQLite database file ending with .sqlite3.
  670. */
  671. void statistics::writeToDatabase(std::string database_path, std::vector<std::chrono::duration<int, std::micro>> timeIntervals, bool del) {
  672. // Generate general file statistics
  673. float duration = getCaptureDurationSeconds();
  674. long sumPacketsSent = 0, senderCountIP = 0;
  675. float sumBandwidthIn = 0.0, sumBandwidthOut = 0.0;
  676. for (auto i = ip_statistics.begin(); i != ip_statistics.end(); i++) {
  677. sumPacketsSent += i->second.pkts_sent;
  678. // Consumed bandwith (bytes) for sending packets
  679. sumBandwidthIn += (i->second.kbytes_received / duration);
  680. sumBandwidthOut += (i->second.kbytes_sent / duration);
  681. senderCountIP++;
  682. }
  683. float avgPacketRate = (packetCount / duration);
  684. long avgPacketSize = getAvgPacketSize();
  685. if(senderCountIP>0) {
  686. long avgPacketsSentPerHost = (sumPacketsSent / senderCountIP);
  687. float avgBandwidthInKBits = (sumBandwidthIn / senderCountIP) * 8;
  688. float avgBandwidthOutInKBits = (sumBandwidthOut / senderCountIP) * 8;
  689. // Create database and write information
  690. statistics_db db(database_path, resourcePath);
  691. db.writeStatisticsFile(packetCount, getCaptureDurationSeconds(),
  692. getFormattedTimestamp(timestamp_firstPacket.seconds(), timestamp_firstPacket.microseconds()),
  693. getFormattedTimestamp(timestamp_lastPacket.seconds(), timestamp_lastPacket.microseconds()),
  694. avgPacketRate, avgPacketSize, avgPacketsSentPerHost, avgBandwidthInKBits,
  695. avgBandwidthOutInKBits, doExtraTests);
  696. db.writeStatisticsIP(ip_statistics);
  697. db.writeStatisticsTTL(ttl_distribution);
  698. db.writeStatisticsIpMac(ip_mac_mapping);
  699. db.writeStatisticsDegree(ip_statistics);
  700. db.writeStatisticsPorts(ip_ports);
  701. db.writeStatisticsProtocols(protocol_distribution);
  702. db.writeStatisticsMSS(mss_distribution);
  703. db.writeStatisticsToS(tos_distribution);
  704. db.writeStatisticsWin(win_distribution);
  705. db.writeStatisticsConv(conv_statistics);
  706. db.writeStatisticsConvExt(conv_statistics_extended);
  707. db.writeStatisticsInterval(interval_statistics, timeIntervals, del, this->default_interval, this->getDoExtraTests());
  708. db.writeDbVersion();
  709. db.writeStatisticsUnrecognizedPDUs(unrecognized_PDUs);
  710. }
  711. else {
  712. // Tinslib failed to recognize the types of the packets in the input PCAP
  713. std::cerr<<"ERROR: Statistics could not be collected from the input PCAP!"<<"\n";
  714. return;
  715. }
  716. }
  717. void statistics::writeIntervalsToDatabase(std::string database_path, std::vector<std::chrono::duration<int, std::micro>> timeIntervals, bool del) {
  718. statistics_db db(database_path, resourcePath);
  719. db.writeStatisticsInterval(interval_statistics, timeIntervals, del, this->default_interval, this->getDoExtraTests());
  720. }