idnits 2.17.1 draft-ietf-bmwg-mswitch-04.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Looks like you're using RFC 2026 boilerplate. This must be updated to follow RFC 3978/3979, as updated by RFC 4748. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard == It seems as if not all pages are separated by form feeds - found 0 form feeds but 34 pages Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an Abstract section. ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There are 3 instances of too long lines in the document, the longest one being 2 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == Line 90 has weird spacing: '...erfaces are g...' -- The exact meaning of the all-uppercase expression 'MAY NOT' is not defined in RFC 2119. If it is intended as a requirements expression, it should be rewritten using one of the combinations defined in RFC 2119; otherwise it should not be all-uppercase. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame originating from the DUT/SUT (spanning tree, SNMP, RIP, ...) MUST not be counted as a received frame. Frames originating from the DUT/SUT MAY be counted as flooded frames or not counted at all. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame originating from the DUT/SUT MUST not be counted as a received frame. Frames originating from the DUT/SUT MAY be counted as flooded frames or not counted at all. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame originating from the DUT/SUT MUST not be counted as a received frame. Frames originating from the DUT/SUT MAY be counted as flooded frames or not counted at all. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame originating from the DUT/SUT MUST not be counted as a received frame. Frames originating from the DUT/SUT MAY be counted as flooded frames or not counted at all. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame received which does not have the correct destination address MUST not be counted as a received frame and SHOULD be counted as part of a flood count. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame originating from the DUT/SUT MUST not be counted as a received frame. Frames originating from the DUT/SUT MAY be counted as flooded frames or not counted at all. == Using lowercase 'not' together with uppercase 'MUST', 'SHALL', 'SHOULD', or 'RECOMMENDED' is not an accepted usage according to RFC 2119. Please use uppercase 'NOT' together with RFC 2119 keywords (if that is what you mean). Found 'MUST not' in this paragraph: Any frame originating from the DUT/SUT MUST not be counted as a received frame. Frames originating from the DUT/SUT MAY be counted as flooded frames or not counted at all. == The expression 'MAY NOT', while looking like RFC 2119 requirements text, is not defined in RFC 2119, and should not be used. Consider using 'MUST NOT' instead (if that is what you mean). Found 'MAY NOT' in this paragraph: All ports on the traffic generator MUST transmit the exact number of test frames. The exact number of test frames is found by multiplying the Iload of the port by the Trial Duration. All ports MAY NOT transmit the same number of frames if their Iload is not the same. An example would be the Partially meshed many-to-one test. -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (April 2000) is 8767 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Downref: Normative reference to an Informational RFC: RFC 1242 (ref. '1') ** Downref: Normative reference to an Informational RFC: RFC 2285 (ref. '2') ** Downref: Normative reference to an Informational RFC: RFC 2544 (ref. '3') -- Possible downref: Non-RFC (?) normative reference: ref. '4' -- Possible downref: Non-RFC (?) normative reference: ref. '5' Summary: 8 errors (**), 0 flaws (~~), 11 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Network Working Group R. Mandeville 2 Internet-Draft European Network Laboratories 3 Expiration Date: October 2000 J. Perser 4 Netcom Systems 5 April 2000 7 Benchmarking Methodology for LAN Switching Devices 8 10 Status of this Memo 12 This document is an Internet-Draft and is in full conformance with 13 all provisions of Section 10 of RFC2026. 15 Internet-Drafts are working documents of the Internet Engineering 16 Task Force (IETF), its areas, and its working groups. Note that 17 other groups may also distribute working documents as Internet- 18 Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet-Drafts as reference 23 material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt 28 The list of Internet-Draft Shadow Directories can be accessed at 29 http://www.ietf.org/shadow.html. 31 Table of Contents 33 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 34 2. Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 35 3. Test setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 36 4. Frame formats and sizes . . . . . . . . . . . . . . . . . . . . . 3 37 5. Benchmarking Tests . . . . . . . . . . . . . . . . . . . . . . . . 4 38 5.1 Fully meshed throughput, frame loss and forwarding rates . . 4 39 5.2 Partially meshed one-to-many/many-to-one . . . . . . . . . . 7 40 5.3 Partially meshed multiple devices . . . . . . . . . . . . . . 10 41 5.4 Partially meshed unidirectional traffic . . . . . . . . . . . 13 42 5.5 Congestion Control . . . . . . . . . . . . . . . . . . . . . 16 43 5.6 Forward Pressure and Maximum Forwarding Rate . . . . . . . . 19 44 5.7 Address caching capacity . . . . . . . . . . . . . . . . . . 21 45 5.8 Address learning rate . . . . . . . . . . . . . . . . . . . . 24 46 5.9 Errored frames filtering. . . . . . . . . . . . . . . . . . . 26 47 5.10 Broadcast frame Forwarding and Latency . . . . . . . . . . . 28 48 6. Security Considerations . . . . . . . . . . . . . . . . . . . . . 29 49 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 50 8. Authors' Address . . . . . . . . . . . . . . . . . . . . . . . . . 30 51 Appendix A: Formulas . . . . . . . . . . . . . . . . . . . . . . . 31 52 Appendix B: Generating Offered Load . . . . . . . . . . . . . . . 33 54 1. Introduction 56 This document is intended to provide methodology for the benchmarking 57 of local area network (LAN) switching devices. It extends the 58 methodology already defined for benchmarking network interconnecting 59 devices in RFC 2544 [3] to switching devices. 61 This RFC primarily deals with devices which switch frames at the 62 Medium Access Control (MAC) layer. It provides a methodology for 63 benchmarking switching devices, forwarding performance, congestion 64 control, latency, address handling and filtering. In addition to 65 defining the tests, this document also describes specific formats for 66 reporting the results of the tests. 68 A previous document, "Benchmarking Terminology for LAN Switching 69 Devices" [2], defined many of the terms that are used in this 70 document. The terminology document SHOULD be consulted before 71 attempting to make use of this document. 73 2. Requirements 75 The following RFCs SHOULD be consulted before attempting to make use 76 of this document: RFC 1242 [1], RFC 2285 [2], and RFC 2544 [3]. 78 For the sake of clarity and continuity, this RFC adopts the template 79 for benchmarking tests set out in Section 26 of RFC 2544. 81 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 82 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in 83 this document are to be interpreted as described in RFC 2119. 85 3. Test setup 87 This document extends the general test setup described in section 6 88 of RFC 2544 [3] to the benchmarking of LAN switching devices. 89 RFC 2544 [3] primarily describes non-meshed traffic where input and 90 output interfaces are grouped in mutually exclusive sending and 91 receiving pairs. In fully meshed traffic, each interface of a 92 DUT/SUT is set up to both receive and transmit frames to all the 93 other interfaces under test. 95 Prior to each test run, the DUT/SUT MUST learn the MAC addresses used 96 in the test and the address learning SHOULD be verified. Addresses 97 not learned will be forwarded as flooded frames and reduce the amount 98 of correctly forwarded frames. The rate at which address learning 99 frames are offered may have to be adjusted to be as low as 50 frames 100 per second or even less, to guarantee successful learning. The 101 DUT/SUT address aging time SHOULD be configured to be greater than 102 the period of the learning phase of the test plus the trial duration 103 plus any configuration time required by the testing device. 104 Addresses SHOULD NOT age out until the trial duration is completed. 105 More than one learning trial may be needed for the association of the 106 address to the port to occur. 108 If a DUT/SUT uses a hashing algorithm with address learning, the 109 DUT/SUT may not learn the necessary addresses to perform the tests. 110 The format of the MAC addresses MUST be adjustable so that the 111 address mapping may be re-arranged to ensure that the DUT/SUT learns 112 all the addresses. 114 4. Frame formats and sizes 116 The test frame format is defined in RFC 2544 section 8 [3] and MUST 117 contain a unique signature field located in the UDP DATA area of the 118 Test Frame (see Appendix C [3]). The purpose of the signature field 119 is filter out frames that are not part of the offered load. 121 The signature field MUST be unique enough to identify the frames not 122 originating from the DUT/SUT. The signature field SHOULD be located 123 after byte 56 (collision window [4] ) or at the end of the frame. 124 The length, contents and method of detection is not defined in this 125 memo. 127 The signature field MAY have a unique identifier per port. This 128 would filter out misforwarded frames. It is possible for a DUT/SUT 129 to strip off the MAC layer, send it through its switching matrix, 130 and transmit it out with the correct destination MAC address but the 131 wrong payload. 133 For frame sizes, refer to RFC 2544, section 9 [3]. 135 There are three possible frame formats for layer 2 Ethernet switches: 136 standard MAC Ethernet frames, standard MAC Ethernet frames with 137 vendor-specific tags added to them, and IEEE 802.3ac frames tagged to 138 accommodate 802.1p&Q. The two types of tagged frames may exceed the 139 standard maximum length frame of 1518 bytes, and may not be accepted 140 by the interface controllers of some DUT/SUTs. It is recommended to 141 check the compatibility of the DUT/SUT with tagged frames before 142 testing. 144 Devices switching tagged frames of over 1518 bytes will have a 145 different maximum forwarding rate than untagged frames. 147 5. Benchmarking Tests 149 The following tests offer objectives, procedures, and reporting 150 formats for benchmarking LAN switching devices. 152 5.1 Fully meshed throughput, frame loss and forwarding rates 154 5.1.1 Objective 156 To determine the throughput, frame loss and forwarding rates of 157 DUT/SUTs offered fully meshed traffic as defined in RFC 2285 [2]. 159 5.1.2 Setup Parameters 161 When offering full meshed traffic, the following parameters MUST be 162 defined. Each parameter is configured with the following 163 considerations. 165 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 166 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 167 bytes are included in the frame size specified. 169 Interframe Gap (IFG) - The IFG between frames inside a burst 170 MUST be at the minimum specified by the standard (9.6 us for 171 10Mbps Ethernet, 960 ns for 100Mbps Ethernet, and 96 ns for 172 1 Gbps Ethernet) of the medium being tested. 174 Duplex mode - Half duplex or full duplex. 176 ILoad - Intended Load per port is expressed in a percentage of the 177 medium's maximum theoretical load, regardless of traffic 178 orientation or duplex mode. Certain test configurations will 179 theoretically over-subscribe the DUT/SUT. 181 In half duplex, an ILoad over 50% will over-subscribe the DUT/SUT. 183 Burst Size - The burst size defines the number of frames sent 184 back-to-back at the minimum legal IFG [4] before pausing 185 transmission to receive frames. Burst sizes SHOULD vary between 1 186 and 930 frames. A burst size of 1 will simulate constant load 187 [1]. 189 Addresses per port - Represents the number of addresses which 190 are being tested for each port. Number of addresses SHOULD be a 191 binary exponential (i.e. 1, 2, 4, 8, 16, 32, 64, 128, 256, ...). 192 Recommended value is 1. 194 Trial Duration - The recommended Trial Duration is 30 seconds. 195 Trial duration SHOULD be adjustable between 1 and 300 seconds. 197 5.1.3 Procedure 199 All ports on the tester MUST transmit test frames either in a Frame 200 Based or Time Based mode (Appendix B). All ports SHOULD start 201 transmitting their frames within 1% of the trial duration. For a 202 trial duration of 30 seconds, all ports SHOULD have started 203 transmitting frames within 300 milliseconds of each other. 205 Each port in the test MUST send test frames to all other ports in a 206 round robin type fashion. The sequence of addresses MUST NOT change 207 when congestion control is applied. The following table shows how 208 each port in a test MUST transmit test frames to all other ports in 209 the test. In this example, there are six ports with 1 address per 210 port: 212 Source Port Destination Ports (in order of transmission) 214 Port #1 2 3 4 5 6 2... 215 Port #2 3 4 5 6 1 3... 216 Port #3 4 5 6 1 2 4... 217 Port #4 5 6 1 2 3 5... 218 Port #5 6 1 2 3 4 6... 219 Port #6 1 2 3 4 5 1... 221 As shown in the table, there is an equal distribution of destination 222 addresses for each transmit opportunity. This keeps the test balanced 223 so that one destination port is not overloaded by the test algorithm 224 and all ports are equally and fully loaded throughout the test. Not 225 following this algorithm exactly will produce inconsistent results. 227 For tests using multiple addresses per port, the actual port 228 destinations are the same as described above and the actual 229 source/destination address pairs SHOULD be chosen randomly to 230 exercise the DUT/SUT's ability to perform address lookups. 232 For every address, learning frames MUST be sent to the DUT/SUT to 233 allow the DUT/SUT update its address tables properly. 235 5.1.4 Measurements 237 Each port should receive the same number of test frames that it 238 transmitted. Each receiving port MUST categorize, then count the 239 frames into one of two groups: 241 1.) Received Frames: received frames MUST have the correct 242 destination MAC address and SHOULD match a signature field. 244 2.) Flood count [2]. 246 Any frame originating from the DUT/SUT (spanning tree, SNMP, RIP, 247 ...) MUST not be counted as a received frame. Frames originating 248 from the DUT/SUT MAY be counted as flooded frames or not counted at 249 all. 251 Frame loss rate of the DUT/SUT SHOULD be reported as defined in 252 section 26.3 [3] with the following notes: Frame loss rate SHOULD be 253 measured at the end of the trail duration. The term "rate", for this 254 measurement only, does not imply the units in the fashion of "per 255 second." 257 5.1.4.1 Throughput 259 Throughput measurement is defined in section 26.1 [3]. A search 260 algorithm is employed to find the maximum Oload [2] with a zero Frame 261 loss rate [1]. The algorithm MUST adjust Iload to find the 262 throughput. 264 5.1.4.2 Forwarding Rate 266 Forwarding rate (FR) of the DUT/SUT SHOULD be reported as the number 267 of test frames per second that the device is observed to successfully 268 forward to the correct destination interface in response to a 269 specified Oload. The Oload MUST also be cited. 271 Forwarding rate at maximum offered load (FRMOL) MUST be reported 272 as the number of test frames per second that a device can 273 successfully transmit to the correct destination interface in 274 response to the MOL as defined in section 3.6 [2]. The MOL MUST also 275 be cited. 277 Maximum forwarding rate (MFR) MUST be reported as the highest 278 forwarding rate of a DUT/SUT taken from an iterative set of 279 forwarding rate measurements. The iterative set of forwarding rate 280 measurements are made by adjusting Iload. The Oload applied to the 281 device MUST also be cited. 283 5.1.5 Reporting format 285 The results for these tests SHOULD be reported in the form of a 286 graph. The x coordinate SHOULD be the frame size, the y coordinate 287 SHOULD be the test results. There SHOULD be at least two lines on 288 the graph, one plotting the theoretical and one plotting the test 289 results. 291 To measure the DUT/SUT's ability to switch traffic while performing 292 many different address lookups, the number of addresses per port 293 MAY be increased in a series of tests. 295 5.2 Partially meshed one-to-many/many-to-one 297 5.2.1 Objective 299 To determine the throughput when transmitting from/to multiple ports 300 and to/from one port. As with the fully meshed throughput test, this 301 test is a measure of the capability of the DUT to switch frames 302 without frame loss. Results of this test can be used to determine 303 the ability of the DUT to utilize an Ethernet port when switching 304 traffic from multiple Ethernet ports. 306 5.2.2 Setup Parameters 308 When offering bursty meshed traffic, the following parameters MUST 309 be defined. Each parameter is configured with the following 310 considerations. 312 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 313 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 314 bytes are included in the frame size specified. 316 Traffic Direction - Traffic can be generated in one direction, the 317 reverse direction, or both directions. 319 Interframe Gap (IFG) - The IFG between frames inside a burst 320 MUST be at the minimum specified by the standard (9.6 us for 321 10Mbps Ethernet, 960 ns for 100Mbps Ethernet, and 96 ns for 322 1 Gbps Ethernet) of the medium being tested. 324 Duplex mode - Half duplex or full duplex. 326 ILoad - Intended Load per port is expressed in a percentage of the 327 medium's maximum theoretical load, regardless of traffic 328 orientation or duplex mode. Certain test configurations will 329 theoretically over-subscribe the DUT/SUT. 331 In half duplex bidirectional traffic, an ILoad over 50% will over- 332 subscribe the DUT/SUT. 334 Burst Size - The burst size defines the number of frames sent 335 back-to-back at the minimum legal IFG [4] before pausing 336 transmission to receive frames. Burst sizes SHOULD vary between 1 337 and 930 frames. A burst size of 1 will simulate constant load 338 [1]. 340 Addresses per port - Represents the number of addresses which 341 are being tested for each port. Number of addresses SHOULD be a 342 binary exponential (i.e. 1, 2, 4, 8, 16, 32, 64, 128, 256, ...). 343 Recommended value is 1. 345 Trial Duration - The recommended Trial Duration is 30 seconds. 346 Trial duration SHOULD be adjustable between 1 and 300 seconds. 348 5.2.3 Procedure 350 All ports on the tester MUST transmit test frames either in a Frame 351 Based or Time Based mode (Appendix B). Depending upon traffic 352 direction, some or all of the ports will be transmitting. All ports 353 SHOULD start transmitting their frames within 1% of the trial 354 duration. For a trial duration of 30 seconds, all ports SHOULD have 355 started transmitting frames within 300 milliseconds of each other. 357 Test frames transmitted from the Many Ports MUST be destined to the 358 One port. Test frames transmitted from the One Port MUST be destined 359 to the Many ports in a round robin type fashion. See section 5.1.3 360 for a description of the round robin fashion. 362 For tests using multiple addresses per port, the actual port 363 destinations are the same as described above and the actual 364 source/destination address pairs SHOULD be chosen randomly to 365 exercise the DUT/SUT's ability to perform address lookups. 367 +----------+ 368 | | 369 | Many | <-------- 370 | | \ 371 +----------+ \ 372 \ 373 +----------+ \ +-------------+ 374 | | ------------> | | 375 | Many | <-----------------------> | One | 376 | | ------------> | | 377 +----------+ / +-------------+ 378 / 379 +----------+ / 380 | | / 381 | Many | <------- 382 | | 383 +----------+ 385 For every address, the testing device MUST send learning frames to 386 allow the DUT/SUT to update its address tables properly. 388 5.2.4 Measurements 390 Each receiving port MUST categorize, then count the frames into one 391 of two groups: 393 1.) Received Frames: received frames MUST have the correct 394 destination MAC address and SHOULD match a signature field. 396 2.) Flood count [2]. 398 Any frame originating from the DUT/SUT MUST not be counted as a 399 received frame. Frames originating from the DUT/SUT MAY be counted 400 as flooded frames or not counted at all. 402 Forwarding rate (FR) of the DUT/SUT SHOULD be reported as the number 403 of test frames per second that the device is observed to successfully 404 transmit to the correct destination interface in response to a 405 specified Oload. The Oload MUST also be cited. 407 Forwarding rate at maximum offered load (FRMOL) MUST be reported 408 as the number of test frames per second that a device can 409 successfully transmit to the correct destination interface in 410 response to the MOL as defined in section 3.6 [2]. The MOL MUST also 411 be cited. 413 Maximum forwarding rate (MFR) MUST be reported as the highest 414 forwarding rate of a DUT/SUT taken from an iterative set of 415 forwarding rate measurements. The iterative set of forwarding rate 416 measurements are made by adjusting Iload. The Oload applied to the 417 device MUST also be cited. 419 5.2.5 Reporting Format 421 The results for these tests SHOULD be reported in the form of a 422 graph. The x coordinate SHOULD be the frame size, the y coordinate 423 SHOULD be the test results. There SHOULD be at least two lines on 424 the graph, one plotting the theoretical and one plotting the test 425 results. 427 To measure the DUT/SUT's ability to switch traffic while performing 428 many different address lookups, the number of addresses per port 429 MAY be increased in a series of tests. 431 5.3 Partially meshed multiple devices 433 5.3.1 Objective 435 To determine the throughput, frame loss and forwarding rates of two 436 switching devices equipped with multiple ports and one high speed 437 backbone uplink (Gigabit Ethernet, ATM, SONET). 439 5.3.2 Setup Parameters 441 When offering bursty partially meshed traffic, the following 442 parameters MUST be defined. Each variable is configured with the 443 following considerations. 445 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 446 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 447 bytes are included in the frame size specified. 449 Interframe Gap (IFG) - The IFG between frames inside a burst 450 MUST be at the minimum specified by the standard (9.6 us for 451 10Mbps Ethernet, 960 ns for 100Mbps Ethernet, and 96 ns for 452 1 Gbps Ethernet) of the medium being tested. 454 Duplex mode - Half duplex or full duplex. 456 ILoad - Intended Load per port is expressed in a percentage of the 457 medium's maximum theoretical load, regardless of traffic 458 orientation or duplex mode. Certain test configurations will 459 theoretically over-subscribe the DUT/SUT. 461 In half duplex, an ILoad over 50% will over-subscribe the DUT/SUT. 463 Burst Size - The burst size defines the number of frames sent 464 back-to-back at the minimum legal IFG [4] before pausing 465 transmission to receive frames. Burst sizes SHOULD vary between 1 466 and 930 frames. A burst size of 1 will simulate constant load 467 [1]. 469 Addresses per port - Represents the number of addresses which 470 are being tested for each port. Number of addresses SHOULD be a 471 binary exponential (i.e. 1, 2, 4, 8, 16, 32, 64, 128, 256, ...). 472 Recommended value is 1. 474 Trial Duration - The recommended Trial Duration is 30 seconds. 475 Trial duration SHOULD be adjustable between 1 and 300 seconds. 477 Local Traffic - A Boolean value of ON or OFF. The frame sequence 478 algorithm MAY be altered to remove local traffic. With local 479 traffic ON, the algorithm is exactly the same as a fully meshed 480 throughput. With local traffic OFF, the port sends frames to all 481 other ports on the other side of the backbone uplink in a round 482 robin type fashion. 484 5.3.3 Procedure 486 All ports on the tester MUST transmit test frames either in a Frame 487 Based or Time Based mode (Appendix B). All ports SHOULD start 488 transmitting their frames within 1% of the trial duration. For a 489 trial duration of 30 seconds, all ports SHOULD have started 490 transmitting frames with 300 milliseconds of each other. 492 Each port in the test MUST send test frames to all other ports in a 493 round robin type fashion as defined in section 5.1.3. Local traffic 494 MAY be removed from the round robin list in order to send the entire 495 load across the backbone uplink. 497 For tests using multiple addresses per port, the actual port 498 destinations are the same as described above and the actual 499 source/destination address pairs SHOULD be chosen randomly to 500 exercise the DUT/SUT's ability to perform address lookups. 502 For every address, the testing device MUST send learning frames to 503 allow the DUT/SUT to update its address tables properly. 505 To measure the DUT/SUT's ability to switch traffic while performing 506 many different address lookups, the number of addresses per port 507 MAY be increased in a series of tests. 509 5.3.4 Measurements 511 Each receiving port MUST categorize, then count the frames into one 512 of two groups: 514 1.) Received frames MUST have the correct destination MAC address 515 and SHOULD match a signature field. 517 2.) Flood count [2]. 519 Any frame originating from the DUT/SUT MUST not be counted as a 520 received frame. Frames originating from the DUT/SUT MAY be counted 521 as flooded frames or not counted at all. 523 Frame loss rate of the DUT/SUT SHOULD be reported as defined in 524 section 26.3 [3] with the following notes: Frame loss rate SHOULD be 525 measured at the end of the trial duration. The term "rate", for this 526 measurement only, does not imply the units in the fashion of "per 527 second." 529 5.3.4.1 Throughput 531 Throughput measurement is defined in section 26.1 [3]. A search 532 algorithm is employed to find the maximum Oload [2] with a zero Frame 533 loss rate [1]. The algorithm MUST adjust Iload to find the 534 throughput. 536 5.3.4.2 Forwarding rate 538 Forwarding rate (FR) of the DUT/SUT SHOULD be reported as the number 539 of test frames per second that the device is observed to successfully 540 forward to the correct destination interface in response to a 541 specified Oload. The Oload MUST also be cited. 543 Forwarding rate at maximum offered load (FRMOL) MUST be reported 544 as the number of test frames per second that a device can 545 successfully transmit to the correct destination interface in 546 response to the MOL as defined in section 3.6 [2]. The MOL MUST also 547 be cited. 549 Maximum forwarding rate (MFR) MUST be reported as the highest 550 forwarding rate of a DUT/SUT taken from an iterative set of 551 forwarding rate measurements. The iterative set of forwarding rate 552 measurements are made by adjusting Iload. The Oload applied to the 553 device MUST also be cited. 555 5.3.5 Reporting format 557 The results for these tests SHOULD be reported in the form of a 558 graph. The x coordinate SHOULD be the frame size, the y coordinate 559 SHOULD be the test results. There SHOULD be at least two lines on 560 the graph, one plotting the theoretical and one plotting the test 561 results. 563 To measure the DUT/SUT's ability to switch traffic while performing 564 many different address lookups, the number of addresses per port 565 MAY be increased in a series of tests. 567 5.4 Partially meshed unidirectional traffic 569 5.4.1 Objective 571 To determine the throughput of the DUT/SUT when presented multiple 572 streams of unidirectional traffic with half of the ports on the 573 DUT/SUT are transmitting frames destined to the other half of the 574 ports. 576 5.4.2 Setup Parameters 578 The following parameters MUST be defined. Each variable is 579 configured with the following considerations. 581 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 582 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 583 bytes are included in the frame size specified. 585 Interframe Gap (IFG) - The IFG between frames inside a burst 586 MUST be at the minimum specified by the standard (9.6 us for 587 10Mbps Ethernet, 960 ns for 100Mbps Ethernet, and 96 ns for 588 1 Gbps Ethernet) of the medium being tested. 590 Duplex mode - Half duplex or full duplex. 592 ILoad - Intended Load per port is expressed in a percentage of the 593 medium's maximum theoretical load, regardless of traffic 594 orientation or duplex mode. Certain test configurations will 595 theoretically over-subscribe the DUT/SUT. 597 ILoad will not over-subscribe the DUT/SUT in this test. 599 Burst Size - The burst size defines the number of frames sent 600 back-to-back at the minimum legal IFG [4] before pausing 601 transmission to receive frames. Burst sizes SHOULD vary between 1 602 and 930 frames. A burst size of 1 will simulate constant load 603 [1]. 605 Addresses per port - Represents the number of addresses which 606 are being tested for each port. Number of addresses SHOULD be a 607 binary exponential (i.e. 1, 2, 4, 8, 16, 32, 64, 128, 256, ...). 608 Recommended value is 1. 610 Trial Duration - The recommended Trial Duration is 30 seconds. 611 Trial duration SHOULD be adjustable between 1 and 300 seconds. 613 5.4.3 Procedure 615 Ports do not send and receive test frames simultaneously. As a 616 consequence, there should be no collisions unless the DUT is 617 misforwarding frames, generating flooded or Spanning-Tree frames or 618 is enabling some flow control mechanism. Ports used for this test 619 are either transmitting or receiving, but not both. Those ports which 620 are transmitting send test frames destined to addresses corresponding 621 to each of the ports receiving. This creates a unidirectional mesh 622 of traffic. 624 All ports on the tester MUST transmit test frames either in a Frame 625 Based or Time Based mode (Appendix B). All ports SHOULD start 626 transmitting their frames within 1% of the trial duration. For a 627 trial duration of 30 seconds, all ports SHOULD have started 628 transmitting frames with 300 milliseconds of each other. 630 Each transmitting port in the test MUST send frames to all receiving 631 ports in a round robin type fashion. The sequence of addresses MUST 632 NOT change when congestion control is applied. The following table 633 shows how each port in a test MUST transmit test frames to all other 634 ports in the test. In this 8 port example, port 1 through 4 are 635 transmitting and ports 5 through 8 are receiving; each with 1 address 636 per port: 638 Source Port, then Destination Ports (in order of transmission) 640 Port #1 5 6 7 8 5 6... 641 Port #2 6 7 8 5 6 7... 642 Port #3 7 8 5 6 7 8... 643 Port #4 8 5 6 7 8 5... 645 As shown in the table, there is an equal distribution of destination 646 addresses for each transmit opportunity. This keeps the test balanced 647 so that one destination port is not overloaded by the test algorithm 648 and all receiving ports are equally and fully loaded throughout the 649 test. Not following this algorithm exactly will product inconsistent 650 results. 652 For tests using multiple addresses per port, the actual port 653 destinations are the same as described above and the actual 654 source/destination address pairs SHOULD be chosen randomly to 655 exercise the DUT/SUT's ability to perform address lookups. 657 For every address, the testing device MUST send learning frames to 658 allow the DUT/SUT to load its address tables properly. The address 659 table's aging time SHOULD be set sufficiently longer than the 660 learning time and trial duration time combined. If the address table 661 ages out during the test, the results will show a lower performing 662 DUT/SUT. 664 To measure the DUT/SUT's ability to switch traffic while performing 665 many different address lookups, the number of addresses per port 666 MAY be increased in a series of tests. 668 5.4.4 Measurements 670 Each receiving port MUST categorize, then count the frames into one 671 of two groups: 673 1.) Received Frames: received frames MUST have the correct 674 destination MAC address and SHOULD match a signature field. 676 2.) Flood count [2]. 678 Any frame originating from the DUT/SUT MUST not be counted as a 679 received frame. Frames originating from the DUT/SUT MAY be counted 680 as flooded frames or not counted at all. 682 Frame loss rate of the DUT/SUT SHOULD be reported as defined in 683 section 26.3 [3] with the following notes: Frame loss rate SHOULD be 684 measured at the end of the trial duration. The term "rate", for this 685 measurement only, does not imply the units in the fashion of "per 686 second." 688 5.4.4.1 Throughput 690 Throughput measurement is defined in section 26.1 [3]. A search 691 algorithm is employed to find the maximum Oload [2] with a zero Frame 692 loss rate [1]. The algorithm MUST adjust Iload to find the 693 throughput. 695 5.4.4.2 Forwarding rate 697 Forwarding rate (FR) of the DUT/SUT SHOULD be reported as the number 698 of test frames per second that the device is observed to successfully 699 forward to the correct destination interface in response to a 700 specified Oload. The Oload MUST also be cited. 702 Forwarding rate at maximum offered load (FRMOL) MUST be reported 703 as the number of test frames per second that a device can 704 successfully transmit to the correct destination interface in 705 response to the MOL as defined in section 3.6 [2]. The MOL MUST also 706 be cited. 708 Maximum forwarding rate (MFR) MUST be reported as the highest 709 forwarding rate of a DUT/SUT taken from an iterative set of 710 forwarding rate measurements. The iterative set of forwarding rate 711 measurements are made by adjusting Iload. The Oload applied to the 712 device MUST also be cited. 714 5.4.5 Reporting format 716 The results for these tests SHOULD be reported in the form of a 717 graph. The x coordinate SHOULD be the frame size, the y coordinate 718 SHOULD be the test results. There SHOULD be at least two lines on 719 the graph, one plotting the theoretical and one plotting the test 720 results. 722 To measure the DUT/SUT's ability to switch traffic while performing 723 many different address lookups, the number of addresses per port 724 MAY be increased in a series of tests. 726 5.5 Congestion Control 728 5.5.1 Objective 730 To determine how a DUT handles congestion. Does the device implement 731 congestion control and does congestion on one port affect an 732 uncongested port. This procedure determines if Head of Line Blocking 733 and/or Backpressure are present. 735 5.5.2 Setup Parameters 737 The following parameters MUST be defined. Each variable is 738 configured with the following considerations. 740 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 741 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 742 bytes are included in the frame size specified. 744 Interframe Gap (IFG) - The IFG between frames inside a burst 745 MUST be at the minimum specified by the standard (9.6 us for 746 10Mbps Ethernet, 960 ns for 100Mbps Ethernet, and 96 ns for 747 1 Gbps Ethernet) of the medium being tested. 749 Duplex mode - Half duplex or full duplex. 751 Addresses per port - Represents the number of addresses which 752 are being tested for each port. Number of addresses SHOULD be a 753 binary exponential (i.e. 1, 2, 4, 8, 16, 32, 64, 128, 256, ...). 754 Recommended value is 1. 756 Trial Duration - The recommended Trial Duration is 30 seconds. 757 Trial duration SHOULD be adjustable between 1 and 300 seconds. 759 5.5.3 Procedure 761 This test MUST consist of a multiple of four ports with the same MOL. 762 Four ports are REQUIRED and MAY be expanded to fully utilize the 763 DUT/SUT in increments of four. Each group of four will contain a 764 test block with two of the ports as source transmitters and two of 765 the ports as receivers. The diagram below depicts the flow of traffic 766 between the switch ports: 768 +----------+ 50 % MOL +-------------+ 769 | | ------------------------> | | 770 | | 50 % MOL | uncongested | 771 | | --------- | | 772 +----------+ \ +-------------+ 773 \ 774 \ 775 \ 776 +----------+ \ +-------------+ 777 | | ---------> | | 778 | | 100 % MOL | congested | 779 | | ------------------------> | | 780 +----------+ +-------------+ 782 Both source transmitters MUST transmit the exact number of test 783 frames. The first source MUST transmit test frames at the MOL with 784 the destination address of the two receive ports in an alternating 785 order. The first test frame to the uncongested receive port, second 786 test frame to the congested receive port, then repeat. The second 787 source transmitter MUST transmit test frames at the MOL only to the 788 congested receive port. 790 Both receive ports SHOULD distinguish between test frames originating 791 from the source ports and frames originating from the DUT/SUT. Only 792 test frames from the source ports SHOULD be counted. 794 The uncongested receive port should be receiving at a rate of half 795 the MOL. The number of test frames received on the uncongested port 796 SHOULD be 50% of the test frames transmitted by the first source 797 transmitter. The congested receive port should be receiving at the 798 MOL. The number of test frames received on the congested port should 799 be between 100% and 150% of the test frames transmitted by one source 800 transmitter. 802 Test frames destined to uncongested ports in a switch device should 803 not be dropped due to other ports being congested, even if the source 804 is sending to both the congested and uncongested ports. 806 5.5.4 Measurements 808 Any frame received which does not have the correct destination 809 address MUST not be counted as a received frame and SHOULD be counted 810 as part of a flood count. 812 Any frame originating from the DUT/SUT MUST not be counted as a 813 received frame. Frames originating from the DUT/SUT MAY be counted 814 as flooded frames or not counted at all. 816 Frame loss rate of the DUT/SUT's congested and uncongested ports MUST 817 be reported as defined in section 26.3 [3] with the following notes: 818 Frame loss rate SHOULD be measured at the end of the trial duration. 819 The term "rate", for this measurement only, does not imply the units 820 in the fashion of "per second." 822 Offered Load to the DUT/SUT MUST be reported as the number of test 823 frames per second that the DUT/SUT observed to accept. This may be 824 different that the MOL. 826 Forwarding rate (FR) of the DUT/SUT's congested and uncongested ports 827 MUST be reported as the number of test frames per second that the 828 device is observed to successfully transmit to the correct 829 destination interface in response to a specified offered load. The 830 offered load MUST also be cited. 832 5.5.5 Reporting format 834 This test MUST report the frame lost rate at the uncongested port, 835 the forwarding rate (at 50% offered load) at the uncongested port, 836 and the frame lost rate at the congested port. This test MAY report 837 the frame counts transmitted and frame counts received by the 838 DUT/SUT. 840 5.5.5.1 HOLB 842 If there is frame loss at the uncongested port, "Head of Line" 843 blocking is present. The DUT cannot forward the amount of traffic to 844 the congested port and as a result it is also losing frames destined 845 to the uncongested port. 847 5.5.5.2 Back Pressure 849 If there is no frame loss on the congested port, then backpressure 850 is present. It should be noted that this test expects the overall 851 load to the congested port to be greater than 100%. Therefore if the 852 load is greater than 100% and no frame loss is detected, then the DUT 853 must be implementing a flow control mechanism. The type of flow 854 control mechanism used is beyond the scope of this memo. 856 It should be noted that some DUTs may not be able to handle the 100% 857 load presented at the input port. In this case, there may be frame 858 loss reported at the uncongested port which is due to the load at the 859 input port rather than the congested port's load. 861 If the uncongested frame loss is reported as zero, but the maximum 862 forwarding rate is less than 7440 (for 10Mbps Ethernet), then this 863 may be an indication of congestion control being enforced by the DUT. 864 In this case, the congestion control is affecting the throughput of 865 the uncongested port. 867 If no congestion control is detected, the expected percentage frame 868 loss for the congested port is 33% at 150% overload. It is receiving 869 100% load from 1 port, and 50% from another, and can only get 100% 870 possible throughput, therefore having a frame loss rate of 33% 871 (150%-50%/150%). 873 5.6 Forward Pressure and Maximum Forwarding Rate 875 5.6.1 Objective 877 The Forward Pressure test overloads a DUT/SUT port and measures the 878 output for forward pressure [2]. If the DUT/SUT transmits frames 879 with an interframe gap less than 96 bits (section 4.2.3.2.2 [4]), 880 then forward pressure is detected. 882 The objective of the Maximum Forwarding Rate test is to measure the 883 peak value of the Forwarding Rate when the Offered Load is varied 884 between the throughput [1] and the Maximum Offered Load [2]. 886 5.6.2 Setup Parameters 888 The following parameters MUST be defined. Each variable is 889 configured with the following considerations. 891 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 892 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 893 bytes are included in the frame size specified. 895 Duplex mode - Half duplex or full duplex. 897 Trial Duration - The recommended Trial Duration is 30 seconds. 898 Trial duration SHOULD be adjustable between 1 and 300 seconds. 900 Step Size - The minimum incremental resolution that the Iload will 901 be incremented in frames per second. The smaller the step size, 902 the more accurate the measurement and the more iterations 903 required. As the Iload approaches the MOL, the minimum step size 904 will increase because of gap resolution on the testing device. 906 5.6.3 Procedure 908 5.6.3.1 Maximum forwarding rate 910 If the Throughput [1] and the MOL [2] are the same, then MFR [2] 911 is equal to the MOL [2]. 913 This test MUST at a minimum be performed in a two-port configuration 914 as described below. Learning frames MUST be sent to allow the 915 DUT/SUT to update its address tables properly. 917 Test frames are transmitted to the first port (port 1) of the DUT/SUT 918 at the Iload. The FR [2] on the second port (port 2) of the DUT/SUT 919 is measured. The Iload is incremented for each Step Size to find the 920 MFR. The algorithm for the test is as follows: 922 CONSTANT 923 MOL = ... frames/sec; {Maximum Offered Load} 924 VARIABLE 925 MFR := 0 frames/sec; {Maximum Forwarding Rate} 926 ILOAD := starting throughput in frames/sec; {offered load} 927 STEP := ... frames/sec; {Step Size} 928 BEGIN 929 ILOAD := ILOAD - STEP; 930 DO 931 BEGIN 932 ILOAD := ILOAD + STEP 933 IF (ILOAD > MOL) THEN 934 BEGIN 935 ILOAD := MOL 936 END 937 AddressLearning; {Port 2 broadcasts with its source address} 938 Transmit(ILOAD); {Port 1 sends frames to Port 2 at Offered load} 939 IF (Port 2 Forwarding Rate > MFR) THEN 940 BEGIN 941 MFR := Port 2 Forwarding Rate; {A higher value than before} 942 END 943 END 944 WHILE (ILOAD < MOL); {ILOAD has reached the MOL value} 945 DONE 947 5.6.3.2 Minimum Interframe Gap 949 The Minimum Interframe gap test SHOULD, at a minimum, be performed in 950 a two-port configuration as described below. Learning frames MUST be 951 sent to allow the DUT/SUT to update its address tables properly. 953 Test frames SHOULD be transmitted to the first port (port 1) of the 954 DUT/SUT with an interframe gap of 88 bits. This will apply forward 955 pressure to the DUT/SUT and overload it at a rate of one byte per 956 frame. The test frames MUST be constructed with a source address of 957 port 1 and a destination address of port 2. 959 The FR on the second port (port 2) of the DUT/SUT is measured. The 960 measured Forwarding Rate should not exceed the medium's maximum 961 theoretical utilization (MOL). 963 5.6.4 Measurements 965 Port 2 MUST categorize, then count the frames into one of two groups: 967 1.) Received Frames: received frames MUST have the correct 968 destination MAC address and SHOULD match a signature field. 970 2.) Flood count [2]. 972 Any frame originating from the DUT/SUT MUST not be counted as a 973 received frame. Frames originating from the DUT/SUT MAY be counted 974 as flooded frames or not counted at all. 976 5.6.5 Reporting format 978 MFR MUST be reported as the highest forwarding rate of a DUT/SUT 979 taken from an iterative set of forwarding rate measurements. The 980 Iload applied to the device MUST also be cited. 982 Forwarding rate (FR) of the DUT/SUT SHOULD be reported as the number 983 of frames per second that the device is observed to successfully 984 transmit to the correct destination interface in response to a 985 specified Oload. The Iload MUST be cited and the Oload MAY be 986 recorded. 988 If the FR exceeds the MOL during the Minimum Interframe gap test, 989 this MUST be highlighted with the expression "Forward Pressure 990 detected". 992 5.7 Address Caching Capacity 994 5.7.1 Objective 996 To determine the address caching capacity of a LAN switching device 997 as defined in RFC 2285, section 3.8.1 [2]. 999 5.7.2 Setup Parameters 1001 The following parameters MUST be defined. Each variable is 1002 configured with the following considerations. 1004 Age Time - The maximum time that a DUT/SUT will keep a learned 1005 address in its forwarding table. 1007 Addresses Learning Rate - The rate at which new addresses are 1008 offered to the DUT/SUT to be learned. The rate at which address 1009 learning frames are offered may have to be adjusted to be as low as 1010 50 frames per second or even less, to guarantee successful 1011 learning. 1013 Initial Addresses - The initial number of addresses to start the 1014 test with. The number MUST be between 1 and the maximum number 1015 supported by the implementation. 1017 5.7.3 Procedure 1019 The aging time of the DUT/SUT MUST be known. The aging time MUST be 1020 longer than the time necessary to produce frames at the specified 1021 rate. If a low frame rate is used for the test, then it may be 1022 possible that sending a large amount of frames may actually take 1023 longer than the aging time. 1025 This test MUST at a minimum be performed in a three-port 1026 configuration described below. The test MAY be expanded to fully 1027 utilized the DUT/SUT in increments of two or three ports. An 1028 increment of two would include an additional Learning port and Test 1029 port. An increment of three would include an additional Learning 1030 port, Test port, and Monitoring port. 1032 The Learning port (Lport) transmits learning frames to the DUT/SUT 1033 with varying source addresses and a fixed destination address 1034 corresponding to the address of the device connected to the Test port 1035 (Tport) of the DUT/SUT. By receiving frames with varying 1036 source addresses, the DUT/SUT should learn these new addresses. The 1037 source addresses MAY be in sequential order. 1039 The Test port (Tport) of the DUT/SUT acts as the receiving port for 1040 the learning frames. Test frames will be transmitted back to the 1041 addresses learned on the Learning port. The algorithm for this is 1042 explained below. 1044 The Monitoring port (Mport) on the DUT/SUT acts as a monitoring port 1045 to listen for flooded or mis-forwarded frames. If the test spans 1046 multiple broadcast domains (VLANs), each broadcast domain REQUIRES a 1047 Monitoring port. 1049 It is highly recommended that SNMP, Spanning Tree, and any other 1050 frames originating from the DUT/SUT be disabled when running this 1051 test. If such protocols cannot be turned off, the flood count MUST 1052 be modified only to count test frame originating from Lport and MUST 1053 NOT count frames originating from the DUT/SUT. 1055 The algorithm for the test is as follows: 1057 CONSTANT 1058 AGE = ...; {value greater that DUT aging time} 1059 MAX = ...; {maximum address support by implementation} 1060 VARIABLE 1061 LOW := 0; {Highest passed valve} 1062 HIGH := MAX; {Lowest failed value} 1063 N := ...; {user specified initial starting point} 1064 BEGIN 1065 DO 1066 BEGIN 1067 PAUSE(AGE); {Age out any learned addresses} 1068 AddressLearning(TPort); {broadcast a frame with its source 1069 Address and broadcast destination} 1070 AddressLearning(LPort); {N frames with varying source addresses 1071 to Test Port} 1072 Transmit(TPort); {N frames with varying destination addresses 1073 corresponding to Learning Port} 1074 IF (MPort receive frame != 0) OR 1075 (LPort receive frames < TPort transmit) THEN 1076 BEGIN {Address Table of DUT/SUT was full} 1077 HIGH := N; 1078 END 1079 ELSE 1080 BEGIN {Address Table of DUT/SUT was NOT full} 1081 LOW := N; 1082 END 1083 N := LOW + (HIGH - LOW)/2; 1084 END WHILE (HIGH - LOW < 2); 1085 END {Value of N equals number of addresses supported by DUT/SUT} 1087 Using a binary search approach, the test targets the exact number of 1088 addresses supported per port with consistent test iterations. Due 1089 to the aging time of DUT/SUT address tables, each iteration may take 1090 some time during the waiting period for the addresses to clear. If 1091 possible, configure the DUT/SUT for a low value for the aging time. 1093 Once the high and low values of N meet, then the threshold of address 1094 handling has been found. 1096 5.7.4 Measurements 1098 Whether the offered addresses per port was successful forwarded 1099 without flooding. 1101 5.7.5 Reporting format 1103 After the test is run, results for each iteration SHOULD be displayed 1104 in a table to include: 1106 The number of addresses used for each test iteration (varied). 1108 The intended load used for each test iteration (fixed). 1110 Number of test frames that were offered to Tport of the DUT/SUT. 1111 This SHOULD match the number of addresses used for the test 1112 iteration. Test frames are the frames sent with varying 1113 destination addresses to confirm that the DUT/SUT has learned 1114 all of the addresses for each test iteration. 1116 The flood count on Tport during the test portion of each test. 1117 If the number is non-zero, this is an indication of the DUT/SUT 1118 flooding a frame in which the destination address is not in the 1119 address table. 1121 The number of frames correctly forwarded to test Lport during 1122 the test portion of the test. Received frames MUST have the 1123 correct destination MAC address and SHOULD match a signature 1124 field. For a passing test iteration, this number should be equal 1125 to the number of frames transmitted by Tport. 1127 The flood count on Lport during the test portion of each test. 1128 If the number is non-zero, this is an indication of the DUT/SUT 1129 flooding a frame in which the destination address is not in the 1130 address table. 1132 The flood count on Mport. If the value is not zero, then this 1133 indicates that for that test iteration, the DUT/SUT could not 1134 determine the proper destination port for that many frames. In 1135 other words, the DUT/SUT flooded the frame to all ports since its 1136 address table was full. 1138 5.8 Address Learning Rate 1140 5.8.1 Objective 1142 To determine the rate of address learning of a LAN switching device. 1144 5.8.2 Setup Parameters 1146 The following parameters MUST be defined. Each variable is 1147 configured with the following considerations. 1149 Age Time - The maximum time that a DUT/SUT will keep a learned 1150 address in its forwarding table. 1152 Initial Addresses Learning Rate - The starting rate at which new 1153 addresses are offered to the DUT/SUT to be learned. 1155 Number of Addresses - The number of addresses that the DUT/SUT must 1156 learn. The number MUST be between 1 and the maximum number 1157 supported by the implementation. It is recommended no to exceed 1158 the address caching capacity found in section 5.9 1160 5.8.3 Procedure 1162 The aging time of the DUT/SUT MUST be known. The aging time MUST be 1163 longer than the time necessary to produce frames at the specified 1164 rate. If a low frame rate is used for the test, then it may be 1165 possible that sending a large amount of frames may actually take 1166 longer than the aging time. 1168 This test MUST at a minimum be performed in a three-port 1169 configuration in section 5.9.3. The test MAY be expanded to fully 1170 utilized the DUT/SUT in increments of two or three ports. An 1171 increment of two would include an additional Learning port and Test 1172 port. An increment of three would include an additional Learning 1173 port, Test port, and Monitoring port. 1175 An algorithm similar to the one used to determine address caching 1176 capacity can be used to determine the address learning rate. This 1177 test iterates the rate at which address learning frames are offered 1178 by the test device connected to the DUT/SUT. It is recommended to 1179 set the number of addresses offered to the DUT/SUT in this test to 1180 the maximum caching capacity. 1182 The address learning rate might be determined for different numbers 1183 of addresses but in each test run, the number MUST remain constant 1184 and SHOULD be equal to or less than the maximum address caching 1185 capacity. 1187 5.8.4 Measurements 1189 Whether the offered addresses per port were successful forwarded 1190 without flooding at the offered learning rate. 1192 5.8.5 Reporting format 1194 After the test is run, results for each iteration SHOULD be displayed 1195 in a table: 1197 The number of addresses used for each test iteration (fixed). 1199 The intended load used for each test iteration (varied). 1201 Number of test frames that were transmitted by Tport. This SHOULD 1202 match the number of addresses used for the test iteration. Test 1203 frames are the frames sent with varying destination addresses to 1204 confirm that the DUT/SUT has learned all of the addresses for each 1205 test iteration. 1207 The flood count on Tport during the test portion of each test. 1208 If the number is non-zero, this is an indication of the DUT/SUT 1209 flooding a frame in which the destination address is not in the 1210 address table. 1212 The number of frames correctly forwarded to test Lport during 1213 the test portion of the test. Received frames MUST have the 1214 correct destination MAC address and SHOULD match a signature 1215 field. For a passing test iteration, this number should be equal 1216 to the number of frames transmitted by Tport. 1218 The flood count on Lport during the test portion of each test. 1219 If the number is non-zero, this is an indication of the DUT/SUT 1220 flooding a frame in which the destination address is not in the 1221 address table. 1223 The flood count on Mport. If the value is not zero, then this 1224 indicates that for that test iteration, the DUT/SUT could not 1225 determine the proper destination port for that many frames. In 1226 other words, the DUT/SUT flooded the frame to all ports since its 1227 address table was full. 1229 5.9 Errored frames filtering 1231 5.9.1 Objective 1233 The objective of the Errored frames filtering test is to determine 1234 the behavior of the DUT under error or abnormal frame conditions. 1235 The results of the test indicate if the DUT/SUT filters the errors, 1236 or simply propagates the errored frames along to the destination. 1238 5.9.2 Setup Parameters 1240 The following parameters MUST be defined. Each variable is 1241 configured with the following considerations. 1243 ILoad - Intended Load per port is expressed in a percentage of the 1244 medium's maximum theorical load possible. The actual transmitted 1245 frame per second is dependent upon half duplex or full duplex 1246 operation. The test SHOULD be run multiple times with a different 1247 load per port in each case. 1249 Trial Duration - The recommended Trial Duration is 30 seconds. 1250 Trial duration SHOULD be adjustable between 1 and 300 seconds. 1252 5.9.3 Procedure 1254 Each of the illegal frames for Ethernet MUST be checked: 1256 Oversize - The DUT/SUT MAY filter frames larger than 1518 bytes 1257 from being propagated through the DUT/SUT section 4.2.4.2.1 [4]. 1258 Oversized frames transmitted to the DUT/SUT should not be forwarded. 1259 DUT/SUT supporting tagged Frames MAY forward frames up to and 1260 including 1522 bytes long (section 4.2.4.2.1 [5]). 1262 Undersize - The DUT/SUT MUST filter frames less than 64 bytes from 1263 being propagated through the DUT/SUT (section 4.2.4.2.2 [4]). 1264 Undersized frames (or collision fragments) received by the DUT/SUT 1265 must not be forwarded. 1267 CRC Errors - The DUT/SUT MUST filter frames that fail the Frame Check 1268 Sequence Validation (section 4.2.4.1.2 [4]) from being propagated 1269 through the DUT/SUT. Frames with an invalid CRC transmitted to the 1270 DUT/SUT should not be forwarded. 1272 Dribble Bit Errors - The DUT/SUT MUST correct and forward frames 1273 containing dribbling bits. Frames transmitted to the DUT/SUT that do 1274 not end in an octet boundary but contain a valid frame check sequence 1275 MUST be accepted by the DUT/SUT (section 4.2.4.2.1 [4]) and forwarded 1276 to the correct receive port with the frame ending in an octet 1277 boundary (section 3.4 [4]). 1279 Alignment Errors - The DUT/SUT MUST filter frames that fail the Frame 1280 Check Sequence Validation AND do not end in an octet boundary. This 1281 is a combination of a CRC error and a Dribble Bit error. When both 1282 errors are occurring in the same frame, the DUT/SUT MUST determine 1283 the CRC error takes precedence and filters the frame (section 1284 4.2.4.1.2 [4]) from being propagated. 1286 5.9.5 Reporting format 1288 For each of the error conditions in section 5.6.3, a "pass" or "fail" 1289 MUST be reported. Actual frame counts MAY be reported for diagnostic 1290 purposes. 1292 5.10 Broadcast frame Forwarding and Latency 1294 5.10.1 Objective 1296 The objective of the Broadcast Frame Forwarding and Latency Test is 1297 to determine the throughput and latency of the DUT when forwarding 1298 broadcast traffic. The ability to forward broadcast frames will 1299 depend upon a specific function built into the device for that 1300 purpose. It is therefore necessary to determine the ability of 1301 DUT/SUT to handle broadcast frames, since there may be many different 1302 ways of implementing such a function. 1304 5.10.2 Setup Parameters 1306 The following parameters MUST be defined. Each variable is 1307 configured with the following considerations. 1309 Frame Size - Recommended frame sizes are 64, 128, 256, 512, 1024, 1310 1280 and 1518 bytes, per RFC 2544 section 9 [3]. The four CRC 1311 bytes are included in the frame size specified. 1313 Duplex mode - Half duplex or full duplex. 1315 ILoad - Intended Load per port is expressed in a percentage of the 1316 medium's maximum theoretical load, regardless of traffic 1317 orientation or duplex mode. Certain test configurations will 1318 theoretically over-subscribe the DUT/SUT. 1320 ILoad will not over-subscribe the DUT/SUT in this test. 1322 Trial Duration - The recommended Trial Duration is 30 seconds. 1323 Trial duration SHOULD be adjustable between 1 and 300 seconds. 1325 5.10.3 Procedure 1327 For this test, there are two parts to be run. 1329 Broadcast Frame Throughput - This portion of the test uses a single 1330 source test port to transmit test frames with a broadcast address 1331 using the frame specified in RFC 2544 [3]. Selected receive ports 1332 then measure the forwarding rate and Frame loss rate. 1334 Broadcast Frame Latency - This test uses the same setup as the 1335 Broadcast Frame throughput, but instead of a large stream of test 1336 frames being sent, only one test frame is sent and the latency to 1337 each of the receive ports are measured in seconds. 1339 5.10.4 Measurements 1341 Frame loss rate of the DUT/SUT SHOULD be reported as defined in 1342 section 26.3 [3] with the following notes: Frame loss rate SHOULD be 1343 measured at the end of the trial duration. The term "rate", for this 1344 measurement only, does not imply the units in the fashion of "per 1345 second." 1347 Forwarding rate (FR) of the DUT/SUT SHOULD be reported as the number 1348 of test frames per second that the device is observed to successfully 1349 forward to the correct destination interface in response to a 1350 specified Oload. The Oload MUST also be cited. 1352 5.10.5 Reporting format 1354 The results for these tests SHOULD be reported in the form of a 1355 graph. The x coordinate SHOULD be the frame size, the y coordinate 1356 SHOULD be the test results. There SHOULD be at least two lines on 1357 the graph, one plotting the theoretical and one plotting the test 1358 results. 1360 To measure the DUT/SUT's ability to switch traffic while performing 1361 many different address lookups, the number of addresses per port 1362 MAY be increased in a series of tests. 1364 6. Security Considerations 1366 As this document is solely for the purpose of providing metric 1367 methodology and describes neither a protocol nor a protocol's 1368 implementation, there are no security considerations associated with 1369 this document. 1371 7. References 1373 [1] Bradner, S., Editor, "Benchmarking Terminology for Network 1374 Interconnection Devices", RFC 1242, July 1991. 1376 [2] Mandeville, R., Editor, "Benchmarking Terminology for LAN 1377 Switching Devices", RFC 2285, February 1998. 1379 [3] Bradner, S., Editor, "Benchmarking Methodology for Network 1380 Interconnect Devices", RFC 2544, March 2000. 1382 [4] ANSI/IEEE, "CSMA/CD Access Method and Physical Layer 1383 Specifications," ISO/IEC 8802-3, ISBN 0-7381-0330-6, 1998. 1385 [5] IEEE Draft, "Frame Extensions for Virtual Bridged Local Area 1386 Networks (VLAN) Tagging on 802.3 Networks", 802.3ac/D3.1, 1387 July 1998. 1389 8. Authors' Address 1391 Robert Mandeville 1392 European Network Laboratories (ENL) 1393 2, rue Helene Boucher 1394 87286 Guyancourt Cedex 1395 France 1397 Phone: + 33 1 39 44 12 05 1398 EMail: bob@enl.net 1400 Jerry Perser 1401 Netcom Systems 1402 26750 Agoura Road 1403 Calabasas, CA 91302 1404 USA 1406 Phone: + 1 818 676 2300 1407 Email: jerry_perser@netcomsystems.com 1409 Appendix A: Formulas 1411 A.1 Calculating the InterBurst Gap 1413 IBG is defined in RFC 2285 [2] as the interval between two bursts. 1414 To achieve a desired load, the following Input Parameter need to be 1415 defined: 1417 LENGTH - Frame size in bytes including the CRC. 1419 LOAD - The intended load in percent. Range is 0 to 100. 1421 BURST - The number of frames in the burst (integer value). 1423 SPEED - media's speed in bits/sec 1424 Ethernet is 10,000,000 bits/sec 1425 Fast Ethernet is 100,000,000 bits/sec 1426 Gigabit Ethernet is 1,000,000,000 bits/sec 1428 IFG - A constant 96 bits for the minimum interframe gap. 1430 The IBG (in seconds) can be calculated: 1432 [(100/LOAD - 1) * BURST * (IFG + 64 + 8*LENGTH)] + IFG 1433 IBG = ----------------------------------------------------------- 1434 SPEED 1436 A.2 Calculating the Number of Bursts for the Trial Duration 1438 The number of bursts for the trial duration is rounded up to the 1439 nearest integer number. The follow Input Parameter need to be 1440 defined: 1442 LENGTH - Frame size in bytes including the CRC. 1444 BURST - The number of frames in the burst (integer value). 1446 SPEED - media's speed in bits/sec 1447 Ethernet is 10,000,000 bits/sec 1448 Fast Ethernet is 100,000,000 bits/sec 1449 Gigabit Ethernet is 1,000,000,000 bits/sec 1451 IFG - A constant 96 bits for the minimum interframe gap. 1453 IBG - Found in the above formula 1455 DURATION - Trial duration in seconds. 1457 An intermediate number of the Burst duration needs to be calculated 1458 first: 1460 IFG*(BURST-1) + BURST*(64 + 8*LENGTH) 1461 TXTIME = ----------------------------------------- 1462 SPEED 1464 Number of Burst for the Trial Duration (rounded up): 1466 DURATION 1467 #OFBURSTS = -------------- 1468 (TXTIME + IBG) 1470 Example: 1472 LENGTH = 64 bytes per frame 1473 LOAD = 100 % offered load 1474 BURST = 24 frames per burst 1475 SPEED = 10 Mbits/sec (Ethernet) 1476 DURATION = 10 seconds test 1478 IBG = 1612.8 uS 1479 TXTIME = 1603.2 uS 1480 #OFBURSTS = 3110 1482 Appendix B: Generating Offered Load 1484 In testing, the traffic generator is configured with the Iload 1485 (Intended Load) and measures the Oload (Offered Load). If the 1486 DUT/SUT applies congestion control, then the Iload and the Oload are 1487 not the same value. The question arises, how to generate the Oload? 1488 This appendix will describe two different methods. 1490 The unit of measurement for Oload is bits per second. The two 1491 methods described here will hold one unit constant and let the 1492 DUT/SUT vary the other unit. The traffic generator SHOULD specify 1493 which method it uses. 1495 B.1 Frame Based Load 1497 Frame Based Load holds the number of bits constant. The Trial 1498 Duration will vary based upon congestion control. Advantage is 1499 implementation is a simple state machine (or loop). The 1500 disadvantage is that Oload needs to be measured independently. 1502 All ports on the traffic generator MUST transmit the exact number of 1503 test frames. The exact number of test frames is found by multiplying 1504 the Iload of the port by the Trial Duration. All ports MAY NOT 1505 transmit the same number of frames if their Iload is not the same. 1506 An example would be the Partially meshed many-to-one test. 1508 All ports SHOULD start transmitting their frames within 1% of the 1509 trial duration. For a trial duration of 30 seconds, all ports SHOULD 1510 have started transmitting frames within 300 milliseconds of each 1511 other. 1513 The reported Oload SHOULD be the average during the Trial Duration. 1514 If the traffic generator continues to transmit after the Trial 1515 Duration due to congestion control, Oload MAY be averaged over the 1516 entire transmit time. Oload for the DUT/SUT MUST be the aggregate of 1517 all the Oloads per port. Oload per port MAY be reported. 1519 B.2 Time Based Load 1521 Time based load holds the Trial Duration constant, while allowing the 1522 number of octets transmitted to vary. Advantages are an accurate 1523 Trial Duration and integrated Oload measurement. Disadvantage is 1524 that the starting and stopping of the traffic generator MUST be more 1525 accurate. 1527 All ports on the traffic generator are configured to transmit the 1528 Iload for a finite amount of time. Each port MUST count the number 1529 of octets successfully transmitted. 1531 The start and stop is initiated at a layer defined by the test 1532 parameters. The layer can be the MAC layer, IP layer, or some other 1533 point in the protocol stack. The traffic generator MUST complete its 1534 layer specific transmit process when the stop time is reached (i.e. 1535 no fragments, finish the frame). 1537 All ports MUST start transmitting their frames within 1% of the trial 1538 duration. For a trial duration of 30 seconds, all ports SHOULD have 1539 started transmitting frames within 300 milliseconds of each other. 1541 All ports SHOULD stop transmitting frames after the specified trail 1542 duration within 0.01% of the trial duration. Each port's stop time 1543 MUST be reference to its start time. This trial duration error 1544 controls the accuracy of the Oload measurement and SHOULD be reported 1545 with the Oload measurement. 1547 Each port is allowed an offset error of 0.1% and a trial duration 1548 error of 0.01%. 1550 Oload is found by taking the number of octets successfully 1551 transmitted and dividing by the trial duration. Oload for the 1552 DUT/SUT MUST be the aggregate of all the Oloads per port. Oload per 1553 port MAY be reported for diagnostic purposes.