/**
* * Initializes the scheme
*
* @param password user given password
* @return
*/
public static boolean initScheme(String password, long[] features) {
Position[] positions = new Position[Constants.M];
Arrays.fill(
positions,
Position
.BOTH); /* Initially both Alpha and Beta positions are filled with correct entries */
BigInteger hpwd = Generator.getHPWD(Constants.Q); /* generate a random hardened password */
initFunctions(password);
HistoryData histData = new HistoryData(features);
histData.persist(new File(Constants.HISTORY_FILE_PATH), hpwd);
generateScheme(hpwd, positions, password);
LoginHandler.getPreferences().put(Constants.PREF_INITIALIZED, "true");
return true;
}
/**
* @param password
* @param features : array of current login features
* @return true if successfully authenticated
*/
private static boolean _authenticate(String password, long[] features) {
initFunctions(password);
BigInteger hpwd = extractHardenedPwd(features, password);
File historyFile = new File(Constants.HISTORY_FILE_PATH);
HistoryData historyData = HistoryData.loadHistory(historyFile, hpwd);
if (historyData == null) {
/* returns null if decryption of history file fails*/
return false;
}
historyData.addEntry(features);
Position[] positions = computeDistFeatures(historyData);
historyData.persist(historyFile, hpwd);
generateScheme(hpwd, positions, password);
return true;
}
/**
* * computes the distinguishing features from History File Data
*
* @param histData
* @return Position[] containing which features are distinguishing
*/
private static Position[] computeDistFeatures(HistoryData histData) {
Position[] positions = new Position[Constants.M];
if (histData == null || histData.numEntries() < Constants.H) {
Arrays.fill(positions, Position.BOTH); /* If num of entries in history file < 10 */
} else {
double mean, sd;
for (int i = 0; i < Constants.M; i++) {
mean = histData.getMean(i);
sd = histData.getStandardDeviation(i);
if ((Constants.THRESHOLD_FEATURE_VALUES[i] - (Constants.K * sd)) > mean) {
positions[i] = Position.ALPHA;
} else if ((Constants.THRESHOLD_FEATURE_VALUES[i] + (Constants.K * sd)) < mean) {
positions[i] = Position.BETA;
} else {
positions[i] = Position.BOTH;
}
}
}
return positions;
}