[6320] | 1 | """ |
---|
| 2 | ------------------------------------------------------------------------------ |
---|
| 3 | Mango 802.11 Reference Design Experiments Framework - Log File Anonymizer |
---|
| 4 | ------------------------------------------------------------------------------ |
---|
| 5 | License: Copyright 2014-2019, Mango Communications. All rights reserved. |
---|
| 6 | Distributed under the WARP license (http://warpproject.org/license) |
---|
| 7 | ------------------------------------------------------------------------------ |
---|
| 8 | This script uses the wlan_exp Log framework to anonymize a given hdf5 |
---|
| 9 | log file that contains data assocated with an experiment utilizing the |
---|
| 10 | 802.11 reference design and 802.11 Reference Design Experiments Framework. |
---|
| 11 | |
---|
| 12 | Hardware Setup: |
---|
| 13 | - None. Anonymizing log data can be done completely off-line |
---|
| 14 | |
---|
| 15 | Required Script Changes: |
---|
| 16 | - None. Script requires filename of file to be anonymized to be |
---|
| 17 | passed in on the command line. |
---|
| 18 | |
---|
| 19 | Description: |
---|
| 20 | This script parses the log file, removes any personally identifiable |
---|
| 21 | information from the log and write the resulting log data to a new |
---|
| 22 | file. The personally identifiable information that is removed: |
---|
| 23 | - Any MAC address that is not in the following categories: |
---|
| 24 | - Broadcast Address (ff-ff-ff-ff-ff-ff) |
---|
| 25 | - IP v4 Multicast Address (01-00-5E-xx-xx-xx) |
---|
| 26 | - IP v6 Multicast Address (33-33-xx-xx-xx-xx) |
---|
| 27 | - WARP node (40-D8-55-04-2x-xx-xx) |
---|
| 28 | - Any payloads from transmissions / receptions |
---|
| 29 | - Any commands |
---|
| 30 | - Hostnames in the station info |
---|
| 31 | ------------------------------------------------------------------------------ |
---|
| 32 | """ |
---|
| 33 | import sys |
---|
| 34 | import os |
---|
| 35 | import time |
---|
| 36 | import struct |
---|
| 37 | |
---|
| 38 | import wlan_exp.log.util as log_util |
---|
| 39 | import wlan_exp.log.util_hdf as hdf_util |
---|
| 40 | import wlan_exp.log.entry_types as entry_types |
---|
| 41 | |
---|
| 42 | #----------------------------------------------------------------------------- |
---|
| 43 | # Global Variables |
---|
| 44 | #----------------------------------------------------------------------------- |
---|
| 45 | |
---|
| 46 | # Global flag to print performance data |
---|
| 47 | print_time = False |
---|
| 48 | |
---|
| 49 | all_addrs = list() |
---|
| 50 | addr_idx_map = dict() |
---|
| 51 | |
---|
| 52 | |
---|
| 53 | |
---|
| 54 | #----------------------------------------------------------------------------- |
---|
| 55 | # Anonymizer Methods |
---|
| 56 | #----------------------------------------------------------------------------- |
---|
| 57 | def do_replace_addr(addr): |
---|
| 58 | """Determine if the MAC address should be replaced.""" |
---|
| 59 | do_replace = True |
---|
| 60 | |
---|
| 61 | # This list should stay in sync with wlan_exp.util mac_addr_desc_map |
---|
| 62 | |
---|
| 63 | # Don't replace the broadcast address (FF-FF-FF-FF-FF-FF) |
---|
| 64 | if(addr == (0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF)): |
---|
| 65 | do_replace = False |
---|
| 66 | |
---|
| 67 | # Don't replace multicast IP v4 addresses (01-00-5E-00-00-00 to -7F-FF-FF) |
---|
| 68 | # http://technet.microsoft.com/en-us/library/cc957928.aspx |
---|
| 69 | if(addr[0:3] == (0x01, 0x00, 0x5E) and (addr[3] <= 0x7F)): |
---|
| 70 | do_replace = False |
---|
| 71 | |
---|
| 72 | # Don't replace multicast IP v6 addresses (33-33-xx-xx-xx-xx) |
---|
| 73 | # http://www.cavebear.com/archive/cavebear/Ethernet/multicast.html |
---|
| 74 | if(addr[0:2] == (0x33, 0x33)): |
---|
| 75 | do_replace = False |
---|
| 76 | |
---|
| 77 | # Don't replace Mango addresses (40-D8-55-04-2x-xx) |
---|
| 78 | if(addr[0:4] == (0x40, 0xD8, 0x55, 0x04) and ((addr[4] & 0x20) == 0x20)): |
---|
| 79 | do_replace = False |
---|
| 80 | |
---|
| 81 | return do_replace |
---|
| 82 | |
---|
| 83 | |
---|
| 84 | def addr_to_replace(addr, byte_index, addr_idx_map): |
---|
| 85 | """Build map of all indexes for a particular address so they can all be |
---|
| 86 | replaced by the same value. |
---|
| 87 | """ |
---|
| 88 | global all_addrs |
---|
| 89 | if(do_replace_addr(addr)): |
---|
| 90 | if(addr not in all_addrs): |
---|
| 91 | all_addrs.append(addr) |
---|
| 92 | if addr not in addr_idx_map.keys(): |
---|
| 93 | addr_idx_map[addr] = [byte_index,] |
---|
| 94 | else: |
---|
| 95 | addr_idx_map[addr].append(byte_index) |
---|
| 96 | return |
---|
| 97 | |
---|
| 98 | |
---|
| 99 | def log_anonymize(filename): |
---|
| 100 | """Anonymize the log.""" |
---|
| 101 | global all_addrs |
---|
| 102 | |
---|
| 103 | # Get the log_data from the file |
---|
| 104 | log_bytes = bytearray(hdf_util.hdf5_to_log_data(filename=filename)) |
---|
| 105 | |
---|
| 106 | # Get the raw_log_index from the file |
---|
| 107 | raw_log_index = hdf_util.hdf5_to_log_index(filename=filename) |
---|
| 108 | |
---|
| 109 | # Get the user attributes from the file |
---|
| 110 | log_attr_dict = hdf_util.hdf5_to_attr_dict(filename=filename) |
---|
| 111 | |
---|
| 112 | |
---|
| 113 | # Generate the index of log entry locations sorted by log entry type |
---|
| 114 | # Merge the Rx / Tx subtypes that can be processed together |
---|
| 115 | log_index = log_util.filter_log_index(raw_log_index, |
---|
| 116 | merge={'RX_OFDM': ['RX_OFDM', 'RX_OFDM_LTG'], |
---|
| 117 | 'TX_HIGH': ['TX_HIGH', 'TX_HIGH_LTG'], |
---|
| 118 | 'TX_LOW' : ['TX_LOW', 'TX_LOW_LTG']}) |
---|
| 119 | |
---|
| 120 | # Re-initialize the address-byteindex map per file using the running |
---|
| 121 | # list of known MAC addresses |
---|
| 122 | addr_idx_map = dict() |
---|
| 123 | for addr in all_addrs: |
---|
| 124 | addr_idx_map[addr] = list() |
---|
| 125 | |
---|
| 126 | log_util.print_log_index_summary(log_index, "Log Index Summary (merged):") |
---|
| 127 | |
---|
| 128 | |
---|
| 129 | #--------------------------------------------------------------------- |
---|
| 130 | # Step 1: Build a dictionary of all MAC addresses in the log, then |
---|
| 131 | # map each addresses to a unique anonymous address |
---|
| 132 | # Uses tuple(bytearray slice) since bytearray isn't hashable as-is |
---|
| 133 | # |
---|
| 134 | print("Anonmyizing file step 1 ...") |
---|
| 135 | |
---|
| 136 | start_time = time.time() |
---|
| 137 | |
---|
| 138 | #---------------------------------- |
---|
| 139 | # Rx DSSS entries |
---|
| 140 | # |
---|
| 141 | try: |
---|
| 142 | print(" Anonmyizing {0} RX_DSSS entries".format(len(log_index['RX_DSSS']))) |
---|
| 143 | |
---|
| 144 | pyld_start = struct.calcsize(''.join( |
---|
| 145 | entry_types.entry_rx_dsss.get_field_struct_formats()[:-1]) |
---|
| 146 | ) |
---|
| 147 | |
---|
| 148 | for idx in log_index['RX_DSSS']: |
---|
| 149 | # 6-byte addresses at offsets 4, 10, 16 in the mac_payload |
---|
| 150 | for o in (4, 10, 16): |
---|
| 151 | addr_to_replace(tuple(log_bytes[idx+pyld_start+o:idx+pyld_start+o+6]), idx+pyld_start+o, addr_idx_map) |
---|
| 152 | except KeyError: |
---|
| 153 | pass |
---|
| 154 | |
---|
| 155 | if print_time: |
---|
| 156 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 157 | |
---|
| 158 | #---------------------------------- |
---|
| 159 | # Rx OFDM entries |
---|
| 160 | # |
---|
| 161 | try: |
---|
| 162 | print(" Anonmyizing {0} RX_OFDM entries".format(len(log_index['RX_OFDM']))) |
---|
| 163 | |
---|
| 164 | pyld_start = struct.calcsize(''.join( |
---|
| 165 | entry_types.entry_rx_ofdm.get_field_struct_formats()[:-1]) |
---|
| 166 | ) |
---|
| 167 | |
---|
| 168 | for idx in log_index['RX_OFDM']: |
---|
| 169 | # 6-byte addresses at offsets 4, 10, 16 in the mac_payload |
---|
| 170 | for o in (4, 10, 16): |
---|
| 171 | addr_to_replace(tuple(log_bytes[idx+pyld_start+o:idx+pyld_start+o+6]), idx+pyld_start+o, addr_idx_map) |
---|
| 172 | except KeyError: |
---|
| 173 | pass |
---|
| 174 | |
---|
| 175 | if print_time: |
---|
| 176 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 177 | |
---|
| 178 | #---------------------------------- |
---|
| 179 | # Tx entries |
---|
| 180 | # |
---|
| 181 | try: |
---|
| 182 | print(" Anonmyizing {0} TX_HIGH entries".format(len(log_index['TX_HIGH']))) |
---|
| 183 | |
---|
| 184 | pyld_start = struct.calcsize(''.join( |
---|
| 185 | entry_types.entry_tx_high.get_field_struct_formats()[:-1]) |
---|
| 186 | ) |
---|
| 187 | |
---|
| 188 | for idx in log_index['TX_HIGH']: |
---|
| 189 | # 6-byte addresses at offsets 4, 10, 16 in the mac_payload |
---|
| 190 | for o in (4, 10, 16): |
---|
| 191 | addr_to_replace(tuple(log_bytes[idx+pyld_start+o:idx+pyld_start+o+6]), idx+pyld_start+o, addr_idx_map) |
---|
| 192 | except KeyError: |
---|
| 193 | pass |
---|
| 194 | |
---|
| 195 | if print_time: |
---|
| 196 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 197 | |
---|
| 198 | #---------------------------------- |
---|
| 199 | # Tx Low entries |
---|
| 200 | # |
---|
| 201 | try: |
---|
| 202 | print(" Anonmyizing {0} TX_LOW entries".format(len(log_index['TX_LOW']))) |
---|
| 203 | |
---|
| 204 | pyld_start = struct.calcsize(''.join( |
---|
| 205 | entry_types.entry_tx_low.get_field_struct_formats()[:-1]) |
---|
| 206 | ) |
---|
| 207 | |
---|
| 208 | for idx in log_index['TX_LOW']: |
---|
| 209 | # 6-byte addresses at offsets 40, 46, 52 |
---|
| 210 | for o in (4, 10, 16): |
---|
| 211 | addr_to_replace(tuple(log_bytes[idx+pyld_start+o:idx+pyld_start+o+6]), idx+pyld_start+o, addr_idx_map) |
---|
| 212 | except KeyError: |
---|
| 213 | pass |
---|
| 214 | |
---|
| 215 | if print_time: |
---|
| 216 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 217 | |
---|
| 218 | |
---|
| 219 | #--------------------------------------------------------------------- |
---|
| 220 | # Step 2: Enumerate actual MAC addresses and their anonymous replacements |
---|
| 221 | # |
---|
| 222 | print("Anonmyizing file step 2 ...") |
---|
| 223 | |
---|
| 224 | print(" Enumerate MAC addresses and their anonymous replacements") |
---|
| 225 | |
---|
| 226 | addr_map = dict() |
---|
| 227 | for ii,addr in enumerate(all_addrs): |
---|
| 228 | # Address should not have a first octet that is odd, as this indicates |
---|
| 229 | # the address is multicast. Hence, use 0xFE as the first octet. |
---|
| 230 | # |
---|
| 231 | # Due to FCS errors, the number of addresses in a log file is |
---|
| 232 | # potentially large. Therefore, the anonymizer supports 2^24 unique |
---|
| 233 | # addresses. |
---|
| 234 | # |
---|
| 235 | anon_addr = (0xFE, 0xFF, 0xFF, (ii//(256**2)), ((ii//256)%256), (ii%256)) |
---|
| 236 | addr_map[addr] = anon_addr |
---|
| 237 | |
---|
| 238 | if print_time: |
---|
| 239 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 240 | |
---|
| 241 | |
---|
| 242 | #--------------------------------------------------------------------- |
---|
| 243 | # Step 3: Replace all MAC addresses in the log |
---|
| 244 | # |
---|
| 245 | print("Anonmyizing file step 3 ...") |
---|
| 246 | |
---|
| 247 | print(" Replace all MAC addresses in the log") |
---|
| 248 | |
---|
| 249 | for old_addr in addr_idx_map.keys(): |
---|
| 250 | new_addr = bytearray(addr_map[old_addr]) |
---|
| 251 | for byte_idx in addr_idx_map[old_addr]: |
---|
| 252 | log_bytes[byte_idx:byte_idx+6] = new_addr |
---|
| 253 | |
---|
| 254 | if print_time: |
---|
| 255 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 256 | |
---|
| 257 | |
---|
| 258 | #--------------------------------------------------------------------- |
---|
| 259 | # Step 4: Other annonymization steps |
---|
| 260 | # |
---|
| 261 | print("Anonmyizing file step 4 ...") |
---|
| 262 | |
---|
| 263 | print(" Remove all payloads") |
---|
| 264 | |
---|
| 265 | # Overwrite all payloads with zeros |
---|
| 266 | try: |
---|
| 267 | for key in log_index.keys(): |
---|
| 268 | log_util.overwrite_payloads(log_bytes, log_index[key]) |
---|
| 269 | except: |
---|
| 270 | pass |
---|
| 271 | |
---|
| 272 | if print_time: |
---|
| 273 | print(" Time = {0:.3f}s".format(time.time() - start_time)) |
---|
| 274 | |
---|
| 275 | |
---|
| 276 | #--------------------------------------------------------------------- |
---|
| 277 | # Write output files |
---|
| 278 | # |
---|
| 279 | |
---|
| 280 | # Write the modified log to a new HDF5 file |
---|
| 281 | (fn_fldr, fn_file) = os.path.split(filename) |
---|
| 282 | |
---|
| 283 | # Find the last '.' in the file name and classify everything after that as the <ext> |
---|
| 284 | ext_i = fn_file.rfind('.') |
---|
| 285 | if (ext_i != -1): |
---|
| 286 | # Remember the original file extension |
---|
| 287 | fn_ext = fn_file[ext_i:] |
---|
| 288 | fn_base = fn_file[0:ext_i] |
---|
| 289 | else: |
---|
| 290 | fn_ext = '' |
---|
| 291 | fn_base = fn_file |
---|
| 292 | |
---|
| 293 | newfilename = os.path.join(fn_fldr, fn_base + "_anon" + fn_ext) |
---|
| 294 | |
---|
| 295 | print("Writing new file {0} ...".format(newfilename)) |
---|
| 296 | |
---|
| 297 | # Copy any user attributes to the new anonymized file |
---|
| 298 | hdf_util.log_data_to_hdf5(log_bytes, newfilename, attr_dict=log_attr_dict) |
---|
| 299 | |
---|
| 300 | return |
---|
| 301 | |
---|
| 302 | |
---|
| 303 | #----------------------------------------------------------------------------- |
---|
| 304 | # Main |
---|
| 305 | #----------------------------------------------------------------------------- |
---|
| 306 | |
---|
| 307 | if __name__ == '__main__': |
---|
| 308 | if(len(sys.argv) < 2): |
---|
| 309 | print("ERROR: must provide at least one log file input") |
---|
| 310 | sys.exit() |
---|
| 311 | else: |
---|
| 312 | for filename in sys.argv[1:]: |
---|
| 313 | # Ensure the log file actually exists; Print an error and continue to the next file. |
---|
| 314 | if(not os.path.isfile(filename)): |
---|
| 315 | print("\nERROR: File {0} not found".format(filename)) |
---|
| 316 | else: |
---|
| 317 | print("\nAnonymizing file '{0}' ({1:5.1f} MB)\n".format(filename, (os.path.getsize(filename)/1E6))) |
---|
| 318 | log_anonymize(filename) |
---|
| 319 | |
---|
| 320 | print("\nMAC Address Mapping:") |
---|
| 321 | for ii,addr in enumerate(all_addrs): |
---|
| 322 | anon_addr = (0xFE, 0xFF, 0xFF, (ii//(256**2)), ((ii//256)%256), (ii%256)) |
---|
| 323 | print("%2d: %02x:%02x:%02x:%02x:%02x:%02x -> %02x:%02x:%02x:%02x:%02x:%02x" % |
---|
| 324 | (ii, addr[0], addr[1], addr[2], addr[3], addr[4], addr[5], |
---|
| 325 | anon_addr[0], anon_addr[1], anon_addr[2], anon_addr[3], anon_addr[4], anon_addr[5])) |
---|