docs: networking: convert eql.txt to ReST

+140 -295

Documentation/networking/eql.txt Documentation/networking/eql.rst

··· 1 - EQL Driver: Serial IP Load Balancing HOWTO 1 + .. SPDX-License-Identifier: GPL-2.0 2 + 3 + ========================================== 4 + EQL Driver: Serial IP Load Balancing HOWTO 5 + ========================================== 6 + 2 7 Simon "Guru Aleph-Null" Janes, simon@ncm.com 8 + 3 9 v1.1, February 27, 1995 4 10 5 11 This is the manual for the EQL device driver. EQL is a software device ··· 18 12 which was only created to patch cleanly in the very latest kernel 19 13 source trees. (Yes, it worked fine.) 20 14 21 - 1. Introduction 15 + 1. Introduction 16 + =============== 22 17 23 18 Which is worse? A huge fee for a 56K leased line or two phone lines? 24 19 It's probably the former. If you find yourself craving more bandwidth, ··· 48 41 Hey, we can all dream you know... 49 42 50 43 51 - 2. Kernel Configuration 44 + 2. Kernel Configuration 45 + ======================= 52 46 53 47 Here I describe the general steps of getting a kernel up and working 54 48 with the eql driver. From patching, building, to installing. 55 49 56 50 57 - 2.1. Patching The Kernel 51 + 2.1. Patching The Kernel 52 + ------------------------ 58 53 59 54 If you do not have or cannot get a copy of the kernel with the eql 60 55 driver folded into it, get your copy of the driver from 61 56 ftp://slaughter.ncm.com/pub/Linux/LOAD_BALANCING/eql-1.1.tar.gz. 62 57 Unpack this archive someplace obvious like /usr/local/src/. It will 63 - create the following files: 58 + create the following files:: 64 59 65 - 66 - 67 - ______________________________________________________________________ 68 60 -rw-r--r-- guru/ncm 198 Jan 19 18:53 1995 eql-1.1/NO-WARRANTY 69 61 -rw-r--r-- guru/ncm 30620 Feb 27 21:40 1995 eql-1.1/eql-1.1.patch 70 62 -rwxr-xr-x guru/ncm 16111 Jan 12 22:29 1995 eql-1.1/eql_enslave 71 63 -rw-r--r-- guru/ncm 2195 Jan 10 21:48 1995 eql-1.1/eql_enslave.c 72 - ______________________________________________________________________ 73 64 74 65 Unpack a recent kernel (something after 1.1.92) someplace convenient 75 66 like say /usr/src/linux-1.1.92.eql. Use symbolic links to point 76 67 /usr/src/linux to this development directory. 77 68 78 69 79 - Apply the patch by running the commands: 70 + Apply the patch by running the commands:: 80 71 81 - 82 - ______________________________________________________________________ 83 72 cd /usr/src 84 73 patch </usr/local/src/eql-1.1/eql-1.1.patch 85 - ______________________________________________________________________ 86 74 87 75 88 - 89 - 90 - 91 - 2.2. Building The Kernel 76 + 2.2. Building The Kernel 77 + ------------------------ 92 78 93 79 After patching the kernel, run make config and configure the kernel 94 80 for your hardware. ··· 90 90 After configuration, make and install according to your habit. 91 91 92 92 93 - 3. Network Configuration 93 + 3. Network Configuration 94 + ======================== 94 95 95 96 So far, I have only used the eql device with the DSLIP SLIP connection 96 97 manager by Matt Dillon (-- "The man who sold his soul to code so much ··· 101 100 connection. 102 101 103 102 104 - 3.1. /etc/rc.d/rc.inet1 103 + 3.1. /etc/rc.d/rc.inet1 104 + ----------------------- 105 105 106 106 In rc.inet1, ifconfig the eql device to the IP address you usually use 107 107 for your machine, and the MTU you prefer for your SLIP lines. One 108 108 could argue that MTU should be roughly half the usual size for two 109 109 modems, one-third for three, one-fourth for four, etc... But going 110 110 too far below 296 is probably overkill. Here is an example ifconfig 111 - command that sets up the eql device: 111 + command that sets up the eql device:: 112 112 113 - 114 - 115 - ______________________________________________________________________ 116 113 ifconfig eql 198.67.33.239 mtu 1006 117 - ______________________________________________________________________ 118 - 119 - 120 - 121 - 122 114 123 115 Once the eql device is up and running, add a static default route to 124 116 it in the routing table using the cool new route syntax that makes 125 - life so much easier: 117 + life so much easier:: 126 118 127 - 128 - 129 - ______________________________________________________________________ 130 119 route add default eql 131 - ______________________________________________________________________ 132 120 133 121 134 - 3.2. Enslaving Devices By Hand 122 + 3.2. Enslaving Devices By Hand 123 + ------------------------------ 135 124 136 125 Enslaving devices by hand requires two utility programs: eql_enslave 137 126 and eql_emancipate (-- eql_emancipate hasn't been written because when ··· 131 140 132 141 133 142 The syntax for enslaving a device is "eql_enslave <master-name> 134 - <slave-name> <estimated-bps>". Here are some example enslavings: 143 + <slave-name> <estimated-bps>". Here are some example enslavings:: 135 144 136 - 137 - 138 - ______________________________________________________________________ 139 145 eql_enslave eql sl0 28800 140 146 eql_enslave eql ppp0 14400 141 147 eql_enslave eql sl1 57600 142 - ______________________________________________________________________ 143 - 144 - 145 - 146 - 147 148 148 149 When you want to free a device from its life of slavery, you can 149 150 either down the device with ifconfig (eql will automatically bury the 150 151 dead slave and remove it from its queue) or use eql_emancipate to free 151 152 it. (-- Or just ifconfig it down, and the eql driver will take it out 152 - for you.--) 153 + for you.--):: 153 154 154 - 155 - 156 - ______________________________________________________________________ 157 155 eql_emancipate eql sl0 158 156 eql_emancipate eql ppp0 159 157 eql_emancipate eql sl1 160 - ______________________________________________________________________ 161 158 162 159 163 - 164 - 165 - 166 - 3.3. DSLIP Configuration for the eql Device 160 + 3.3. DSLIP Configuration for the eql Device 161 + ------------------------------------------- 167 162 168 163 The general idea is to bring up and keep up as many SLIP connections 169 164 as you need, automatically. 170 165 171 166 172 - 3.3.1. /etc/slip/runslip.conf 167 + 3.3.1. /etc/slip/runslip.conf 168 + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 173 169 174 - Here is an example runslip.conf: 170 + Here is an example runslip.conf:: 175 171 172 + name sl-line-1 173 + enabled 174 + baud 38400 175 + mtu 576 176 + ducmd -e /etc/slip/dialout/cua2-288.xp -t 9 177 + command eql_enslave eql $interface 28800 178 + address 198.67.33.239 179 + line /dev/cua2 176 180 177 - 178 - 179 - 180 - 181 - 182 - 183 - 184 - 185 - 186 - 187 - 188 - 189 - 190 - ______________________________________________________________________ 191 - name sl-line-1 192 - enabled 193 - baud 38400 194 - mtu 576 195 - ducmd -e /etc/slip/dialout/cua2-288.xp -t 9 196 - command eql_enslave eql $interface 28800 197 - address 198.67.33.239 198 - line /dev/cua2 199 - 200 - name sl-line-2 201 - enabled 202 - baud 38400 203 - mtu 576 204 - ducmd -e /etc/slip/dialout/cua3-288.xp -t 9 205 - command eql_enslave eql $interface 28800 206 - address 198.67.33.239 207 - line /dev/cua3 208 - ______________________________________________________________________ 181 + name sl-line-2 182 + enabled 183 + baud 38400 184 + mtu 576 185 + ducmd -e /etc/slip/dialout/cua3-288.xp -t 9 186 + command eql_enslave eql $interface 28800 187 + address 198.67.33.239 188 + line /dev/cua3 209 189 210 190 211 - 212 - 213 - 214 - 3.4. Using PPP and the eql Device 191 + 3.4. Using PPP and the eql Device 192 + --------------------------------- 215 193 216 194 I have not yet done any load-balancing testing for PPP devices, mainly 217 195 because I don't have a PPP-connection manager like SLIP has with ··· 195 235 year. 196 236 197 237 198 - 4. About the Slave Scheduler Algorithm 238 + 4. About the Slave Scheduler Algorithm 239 + ====================================== 199 240 200 241 The slave scheduler probably could be replaced with a dozen other 201 242 things and push traffic much faster. The formula in the current set ··· 215 254 traffic and the "slower" modem starved. 216 255 217 256 218 - 5. Testers' Reports 257 + 5. Testers' Reports 258 + =================== 219 259 220 260 Some people have experimented with the eql device with newer 221 261 kernels (than 1.1.75). I have since updated the driver to patch ··· 224 262 balancing" driver config option. 225 263 226 264 227 - o icee from LinuxNET patched 1.1.86 without any rejects and was able 265 + - icee from LinuxNET patched 1.1.86 without any rejects and was able 228 266 to boot the kernel and enslave a couple of ISDN PPP links. 229 267 230 - 5.1. Randolph Bentson's Test Report 268 + 5.1. Randolph Bentson's Test Report 269 + ----------------------------------- 231 270 271 + :: 232 272 273 + From bentson@grieg.seaslug.org Wed Feb 8 19:08:09 1995 274 + Date: Tue, 7 Feb 95 22:57 PST 275 + From: Randolph Bentson <bentson@grieg.seaslug.org> 276 + To: guru@ncm.com 277 + Subject: EQL driver tests 233 278 234 279 280 + I have been checking out your eql driver. (Nice work, that!) 281 + Although you may already done this performance testing, here 282 + are some data I've discovered. 235 283 284 + Randolph Bentson 285 + bentson@grieg.seaslug.org 236 286 237 - 238 - 239 - 240 - 241 - 242 - 243 - 244 - 245 - 246 - 247 - 248 - 249 - 250 - 251 - 252 - 253 - 254 - 255 - 256 - 257 - 258 - 259 - 260 - 261 - 262 - 263 - 264 - 265 - 266 - 267 - 268 - 269 - 270 - 271 - 272 - 273 - 274 - 275 - 276 - 277 - 278 - 279 - 280 - 281 - 282 - 283 - 284 - 285 - 286 - 287 - 288 - 289 - 290 - 291 - 292 - 293 - From bentson@grieg.seaslug.org Wed Feb 8 19:08:09 1995 294 - Date: Tue, 7 Feb 95 22:57 PST 295 - From: Randolph Bentson <bentson@grieg.seaslug.org> 296 - To: guru@ncm.com 297 - Subject: EQL driver tests 298 - 299 - 300 - I have been checking out your eql driver. (Nice work, that!) 301 - Although you may already done this performance testing, here 302 - are some data I've discovered. 303 - 304 - Randolph Bentson 305 - bentson@grieg.seaslug.org 306 - 307 - --------------------------------------------------------- 287 + ------------------------------------------------------------------ 308 288 309 289 310 290 A pseudo-device driver, EQL, written by Simon Janes, can be used ··· 267 363 Once a link was established, I timed a binary ftp transfer of 268 364 289284 bytes of data. If there were no overhead (packet headers, 269 365 inter-character and inter-packet delays, etc.) the transfers 270 - would take the following times: 366 + would take the following times:: 271 367 272 368 bits/sec seconds 273 369 345600 8.3 ··· 292 388 that the connection establishment seemed fragile for the higher 293 389 speeds. Once established, the connection seemed robust enough.) 294 390 295 - #lines speed mtu seconds theory actual %of 296 - kbit/sec duration speed speed max 297 - 3 115200 900 _ 345600 298 - 3 115200 400 18.1 345600 159825 46 299 - 2 115200 900 _ 230400 300 - 2 115200 600 18.1 230400 159825 69 301 - 2 115200 400 19.3 230400 149888 65 302 - 4 57600 900 _ 234600 303 - 4 57600 600 _ 234600 304 - 4 57600 400 _ 234600 305 - 3 57600 600 20.9 172800 138413 80 306 - 3 57600 900 21.2 172800 136455 78 307 - 3 115200 600 21.7 345600 133311 38 308 - 3 57600 400 22.5 172800 128571 74 309 - 4 38400 900 25.2 153600 114795 74 310 - 4 38400 600 26.4 153600 109577 71 311 - 4 38400 400 27.3 153600 105965 68 312 - 2 57600 900 29.1 115200 99410.3 86 313 - 1 115200 900 30.7 115200 94229.3 81 314 - 2 57600 600 30.2 115200 95789.4 83 315 - 3 38400 900 30.3 115200 95473.3 82 316 - 3 38400 600 31.2 115200 92719.2 80 317 - 1 115200 600 31.3 115200 92423 80 318 - 2 57600 400 32.3 115200 89561.6 77 319 - 1 115200 400 32.8 115200 88196.3 76 320 - 3 38400 400 33.5 115200 86353.4 74 321 - 2 38400 900 43.7 76800 66197.7 86 322 - 2 38400 600 44 76800 65746.4 85 323 - 2 38400 400 47.2 76800 61289 79 324 - 4 19200 900 50.8 76800 56945.7 74 325 - 4 19200 400 53.2 76800 54376.7 70 326 - 4 19200 600 53.7 76800 53870.4 70 327 - 1 57600 900 54.6 57600 52982.4 91 328 - 1 57600 600 56.2 57600 51474 89 329 - 3 19200 900 60.5 57600 47815.5 83 330 - 1 57600 400 60.2 57600 48053.8 83 331 - 3 19200 600 62 57600 46658.7 81 332 - 3 19200 400 64.7 57600 44711.6 77 333 - 1 38400 900 79.4 38400 36433.8 94 334 - 1 38400 600 82.4 38400 35107.3 91 335 - 2 19200 900 84.4 38400 34275.4 89 336 - 1 38400 400 86.8 38400 33327.6 86 337 - 2 19200 600 87.6 38400 33023.3 85 338 - 2 19200 400 91.2 38400 31719.7 82 339 - 4 9600 900 94.7 38400 30547.4 79 340 - 4 9600 400 106 38400 27290.9 71 341 - 4 9600 600 110 38400 26298.5 68 342 - 3 9600 900 118 28800 24515.6 85 343 - 3 9600 600 120 28800 24107 83 344 - 3 9600 400 131 28800 22082.7 76 345 - 1 19200 900 155 19200 18663.5 97 346 - 1 19200 600 161 19200 17968 93 347 - 1 19200 400 170 19200 17016.7 88 348 - 2 9600 600 176 19200 16436.6 85 349 - 2 9600 900 180 19200 16071.3 83 350 - 2 9600 400 181 19200 15982.5 83 351 - 1 9600 900 305 9600 9484.72 98 352 - 1 9600 600 314 9600 9212.87 95 353 - 1 9600 400 332 9600 8713.37 90 391 + ====== ======== === ======== ======= ======= === 392 + #lines speed mtu seconds theory actual %of 393 + kbit/sec duration speed speed max 394 + ====== ======== === ======== ======= ======= === 395 + 3 115200 900 _ 345600 396 + 3 115200 400 18.1 345600 159825 46 397 + 2 115200 900 _ 230400 398 + 2 115200 600 18.1 230400 159825 69 399 + 2 115200 400 19.3 230400 149888 65 400 + 4 57600 900 _ 234600 401 + 4 57600 600 _ 234600 402 + 4 57600 400 _ 234600 403 + 3 57600 600 20.9 172800 138413 80 404 + 3 57600 900 21.2 172800 136455 78 405 + 3 115200 600 21.7 345600 133311 38 406 + 3 57600 400 22.5 172800 128571 74 407 + 4 38400 900 25.2 153600 114795 74 408 + 4 38400 600 26.4 153600 109577 71 409 + 4 38400 400 27.3 153600 105965 68 410 + 2 57600 900 29.1 115200 99410.3 86 411 + 1 115200 900 30.7 115200 94229.3 81 412 + 2 57600 600 30.2 115200 95789.4 83 413 + 3 38400 900 30.3 115200 95473.3 82 414 + 3 38400 600 31.2 115200 92719.2 80 415 + 1 115200 600 31.3 115200 92423 80 416 + 2 57600 400 32.3 115200 89561.6 77 417 + 1 115200 400 32.8 115200 88196.3 76 418 + 3 38400 400 33.5 115200 86353.4 74 419 + 2 38400 900 43.7 76800 66197.7 86 420 + 2 38400 600 44 76800 65746.4 85 421 + 2 38400 400 47.2 76800 61289 79 422 + 4 19200 900 50.8 76800 56945.7 74 423 + 4 19200 400 53.2 76800 54376.7 70 424 + 4 19200 600 53.7 76800 53870.4 70 425 + 1 57600 900 54.6 57600 52982.4 91 426 + 1 57600 600 56.2 57600 51474 89 427 + 3 19200 900 60.5 57600 47815.5 83 428 + 1 57600 400 60.2 57600 48053.8 83 429 + 3 19200 600 62 57600 46658.7 81 430 + 3 19200 400 64.7 57600 44711.6 77 431 + 1 38400 900 79.4 38400 36433.8 94 432 + 1 38400 600 82.4 38400 35107.3 91 433 + 2 19200 900 84.4 38400 34275.4 89 434 + 1 38400 400 86.8 38400 33327.6 86 435 + 2 19200 600 87.6 38400 33023.3 85 436 + 2 19200 400 91.2 38400 31719.7 82 437 + 4 9600 900 94.7 38400 30547.4 79 438 + 4 9600 400 106 38400 27290.9 71 439 + 4 9600 600 110 38400 26298.5 68 440 + 3 9600 900 118 28800 24515.6 85 441 + 3 9600 600 120 28800 24107 83 442 + 3 9600 400 131 28800 22082.7 76 443 + 1 19200 900 155 19200 18663.5 97 444 + 1 19200 600 161 19200 17968 93 445 + 1 19200 400 170 19200 17016.7 88 446 + 2 9600 600 176 19200 16436.6 85 447 + 2 9600 900 180 19200 16071.3 83 448 + 2 9600 400 181 19200 15982.5 83 449 + 1 9600 900 305 9600 9484.72 98 450 + 1 9600 600 314 9600 9212.87 95 451 + 1 9600 400 332 9600 8713.37 90 452 + ====== ======== === ======== ======= ======= === 354 453 454 + 5.2. Anthony Healy's Report 455 + --------------------------- 355 456 457 + :: 356 458 459 + Date: Mon, 13 Feb 1995 16:17:29 +1100 (EST) 460 + From: Antony Healey <ahealey@st.nepean.uws.edu.au> 461 + To: Simon Janes <guru@ncm.com> 462 + Subject: Re: Load Balancing 357 463 358 - 359 - 5.2. Anthony Healy's Report 360 - 361 - 362 - 363 - 364 - 365 - 366 - 367 - Date: Mon, 13 Feb 1995 16:17:29 +1100 (EST) 368 - From: Antony Healey <ahealey@st.nepean.uws.edu.au> 369 - To: Simon Janes <guru@ncm.com> 370 - Subject: Re: Load Balancing 371 - 372 - Hi Simon, 464 + Hi Simon, 373 465 I've installed your patch and it works great. I have trialed 374 466 it over twin SL/IP lines, just over null modems, but I was 375 467 able to data at over 48Kb/s [ISDN link -Simon]. I managed a 376 468 transfer of up to 7.5 Kbyte/s on one go, but averaged around 377 469 6.4 Kbyte/s, which I think is pretty cool. :) 378 - 379 - 380 - 381 - 382 - 383 - 384 - 385 - 386 - 387 - 388 - 389 - 390 - 391 - 392 - 393 - 394 - 395 - 396 - 397 - 398 - 399 - 400 - 401 - 402 - 403 - 404 - 405 - 406 - 407 - 408 - 409 - 410 - 411 - 412 - 413 - 414 - 415 - 416 - 417 - 418 - 419 - 420 - 421 - 422 - 423 - 424 - 425 - 426 - 427 - 428 - 429 - 430 - 431 - 432 -

+1

Documentation/networking/index.rst

··· 54 54 defza 55 55 dns_resolver 56 56 driver 57 + eql 57 58 58 59 .. only:: subproject and html 59 60

+1 -1

drivers/net/Kconfig

··· 126 126 Linux driver or with a Livingston Portmaster 2e. 127 127 128 128 Say Y if you want this and read 129 - <file:Documentation/networking/eql.txt>. You may also want to read 129 + <file:Documentation/networking/eql.rst>. You may also want to read 130 130 section 6.2 of the NET-3-HOWTO, available from 131 131 <http://www.tldp.org/docs.html#howto>. 132 132