RECOMMENDATIONS REPORT (Revised) for File Transfer Via PC Pursuit (Oct 17, 1988) Telenet Communication Corporation Field Operations HQ Tech Support Reston, Virginia CONTENTS PAGE ---- 1.0 Introduction........................................... 1 2.0 Recommendations........................................ 2 3.0 File Transfer Procedures............................... 5 4.0 Summary................................................ 6 SECTION 1.: Introduction Because of many customer complaints concerning PC Pursuit's inability to allow file transfers, Field Operations was requested to provide recommendations for file transfer via PC Pursuit. In compliance with this request, this document provides the following: * Recommendation on the best file transfer protocols to be used with PC Pursuit. * Recommendation on the hunt-confirm sequence and line parameters which provide optimum performance of various protocols. * The average transfer rates which can be expected using the correct hunt-confirm sequence and optional parameter settings. SECTION 2.: Recommendations This section outlines the most common file transfer protocols used with PC Pursuit. The performance of the protocols in the direct con- nect and PC Pursuit environments are also indicated. The following protocols were tested via the Chicago in-dial to the Washington DC out-dial; the observations are summarized below. XMODEM PC Pursuit XMODEM file transfers performed at an average throughput of 30% when the correct hunt-confirm and terminal type was utilized. XMODEM does not support flow control, therefore it is suggested that the "relaxed" mode be invoked if the user's communications software permits this feature. YMODEM The performance of YMODEM file transfers VIA PC Pursuit was found to have an average throughput of 56% when the correct hunt confirm and terminal type is employed. Although YMODEM does not support flow control, it uses large 1024 byte packets which the network PAD handles quite readily under normal conditions. As a result, YMODEM is rated one of the faster protocols for file transfer via PC Pursuit. WXMODEM WXMODEM file transfers utilizing the correct hunt-confirm and terminal type performed well with an average transfer rate of 52%. This protocol is capable of handling flow control which enables it to perform with better reliability in the PC Pursuit environment. Users should be aware that an early version of PROCOMM is known to have a software problem which can affect the performance of WXMODEM file transfers. KERMIT An optimum average throughput of 47% was obtained by KERMIT file transfers via PC Pursuit. The throughput was optimized by modifying the packet size to 90 and adjusting the (host) window size to 16 (for uploads). KERMIT software which supports the sliding window feature performs with optimum efficiency in the PC Pursuit environment. SEALINK SEALINK file transfers via PC Pursuit performed exceptionally well with an average throughput of 74% with the correct hunt-confirm and terminal type. SEALINK supports flow control and was specifically designed to operate in the networking environment. ZMODEM File transfers utilizing ZMODEM protocol via PC Pursuit yielded an average transfer rate of 61%. ZMODEM performs well in the PC Pursuit environment, however; the local configuration for ZMODEM file trans- fers proved to be cumbersome and difficult for the user. ZMODEM re- quires several parameters to be set locally and on the local pad. These parameter settings can vary depending on the type of machine and the type of communications software. The X.3 PAD parameters which should be employed are 1:0,4:10,5:1,7:8,12:1. In addition, flow con- trol (XON/XOFF) should be enabled at the user PC. Because of the difficulty in configuring ZMODEM for file transfer, it is recommended that only seasoned computer users attempt ZMODEM file transfers via PC Pursuit. The following page summarizes file transfer performance of the protocols tested. The protocols are listed in order (PCP best to worst) in two categories: 1) performance via direct connect, 2) per- formance via PC Pursuit utilizing the recommended hunt confirm and parameter settings shown. It should be noted that the optional parameters need only be employed if a user experiences problems with transferring a particular file. SECTION 3.: File Transfer Procedures This section outlines the step by step procedure for executing file transfers via PC Pursuit. These procedures must be followed exactly to achieve optimum transfer rates. The optional X.3 parameters shown on the previous page indicates the parameters which provide the best transfer rate. STEP 1.: Set PC communications software to 8 bits, no parity, 1 stop bit, full duplex. Depending on the type of protocol to be used, disable or enable local (XON/XOFF) flow control. STEP 2.: Dial local rotary with the communications software set at the desired speed. STEP 3.: Upon answer use the correct hunt confirm sequence: At 300/1200bps use - At 2400bps - <@ D CR> NOTE: "D" MUST BE UPPER CASE. STEP 4.: At prompt "TERMINAL = " enter and return. STEP 5.: At the "@" prompt enter the destination mnemonic, out-dial speed, ID and password. It is important that out-dial speed matches in-dial speed. DO NOT MIX IN-DIAL AND OUT-DIAL SPEEDS. STEP 6.: If pad X.3 parameters are to be changed, do so at this point by entering <@ CR>. Set parameters as prescribed. Return to out-dial port by entering . An example of the proper syntax for modifying X.3 PAD parameters would be "SET 7:8,1:0." To display the current PAD parameter settings, the user should enter "PAR?." These are only two of many user commands available. Many of the user commands are clearly defined in the Telenet document "How to Use Telenet's Asynchronous Dial Service." STEP 7.: Upon connecting to the destination pad, insure communication with the out-dial modem by entering . The destination modem will respond with "OK". STEP 8.: Enter and the local number you wish to dial. STEP 9.: Queue host file transfer and start file transfer. Please note that these are the basic steps needed to achieve success- ful file transfers. Since communications software may vary from package to package, additional steps may be needed to initiate the start of the file transfer at the user software level. SECTION 4.: Summary Extensive testing has resulted in identifying the expected performance of six file transfer protocols when used with PC Pursuit. These protocols have been determined to perform satisfactorily with PC Pursuit when the correct hunt-confirm, terminal type and parameters are employed. It is the recommendation of Field Operations that customers be in- formed of the correct logon procedures and the protocols which provide the most reliable file transfers. Customers should also be reminded that PCP users can expect a small degree of network delay which is considered a common characteristic of packet switched networks. In addition, users should also be informed that poor quality voice grade telephone lines can adversely affect file transfer sessions. Field Operations is one of many Telenet groups dedicated to providing customers with complete support for PC Pursuit. Field Operations will offer assistance with file transfer problems providing the customer is willing to release a copy of the problem software as well as provide the pertinent information necessary to resolve the problem. Because problems can vary in nature, the information required to resolve problems can differ from one problem to the next. But at the very least, the following information will be required: * Type of communication software * Type of PC, make, model * Type of modem * Call origin * Call destination * Speed in * Speed out * Type of session * Time of failure * Date of Failure * Point of failure in session * Out-dial number * Network Address (if possible) * Copy of PC Software (in some cases) Additional information may be required depending on the nature of the problem. FILE TRANSFER PERFORMANCE STATISTICS PERFORMANCE STATISTICS DIRECT CONNECT General Communication Parameters = 8 bits 1 stop bit N no parity Terminal Type = D1 | PERFORMANCE STATISTICS DIRECT CONNECT | | | | | | XFR | | PROTOCOL SPEED SECONDS CPS BPS RATE | |======================================================| | | | SEALINK UP 1200 ***517 87.15 871.49 73% | | SEALINK UP 2400 ***270 166.87 1668.74 70% | | SEALINK DN 1200 ***428 105.27 1052.71 88% | | SEALINK DN 2400 ***230 195.90 1958.96 82% | | | | ZMODEM UP 1200 ***408 110.43 1104.31 92% | | ZMODEM UP 2400 ***201 224.16 2241.59 93% | | ZMODEM DN 1200 ***391 115.23 1152.33 96% | | ZMODEM DN 2400 ***195 231.06 2310.56 96% | | | | YMODEM UP 1200 **259 102.80 1027.95 86% | | YMODEM UP 2400 **126 211.30 2113.02 88% | | YMODEM DN 1200 **244 109.11 1091.15 91% | | YMODEM DN 2400 **130 204.80 2048.00 85% | | | | WXMODEM UP 1200 **277 96.12 961.16 80% | | WXMODEM UP 2400 **159 167.45 1674.47 70% | | WXMODEM DN 1200 **405 65.74 657.38 55% | | WXMODEM DN 2400 **131 203.24 2032.37 85% | | | | KERMIT UP 1200 **356 74.79 747.87 62% | | KERMIT UP 2400 **211 126.18 1261.80 53% | | KERMIT DN 1200 **322 82.68 826.83 69% | | KERMIT DN 2400 **178 149.57 1495.73 62% | | | | XMODEM UP 1200 **259 102.80 1027.95 86% | | XMODEM UP 2400 **140 190.17 1901.71 79% | | XMODEM DN 1200 **268 99.34 993.43 83% | | XMODEM DN 2400 **146 182.36 1823.56 76% | ====================================================== **File size = 26624 ***File size = 45056 FILE TRANSFER PERFORMANCE STATISTICS PERFORMANCE STATISTICS VIA PCP General Communication Parameters = 8 bits 1 stop bit N no parity Terminal Type = D1 PERFORMANCE STATISTICS VIA PCP XFR CONFIRM PROTOCOL SP SECONDS CPS BPS RATE SEQUENCE CONTROL SEALINK UP 12 ***561 80.31 803.14 67% CR D CR XON/XOFF SEALINK UP 24 ***300 150.19 1501.87 63% @ D CR XON/XOFF SEALINK DN 12 ***447 100.80 1007.96 84% CR D CR XON/XOFF SEALINK DN 24 ***231 195.05 1950.48 81% @ D CR XON/XOFF *7:8,1:0 End-to-End flow control on ZMODEM UP 12 ***622 72.44 724.37 60% CR D CR XON/XOFF ZMODEM UP 24 ***363 124.12 1241.21 52% @ D CR XON/XOFF ZMODEM DN 12 ***573 78.63 786.32 66% CR D CR XON/XOFF ZMODEM DN 24 ***288 156.44 1564.44 65% @ D CR XON/XOFF *1:0,4:10,5:1,7:8,12:1 End-to-End flow control on YMODEM UP 12 **443 60.10 600.99 50% CR D CR NONE YMODEM UP 24 **200 133.12 1331.20 55% @ D CR NONE YMODEM DN 12 **335 79.47 794.75 66% CR D CR NONE YMODEM DN 24 **211 126.18 1261.80 53% @ D CR NONE *7:8,1:0 WXMODEM UP 12 **480 55.47 554.67 46% CR D CR XON/XOFF WXMODEM UP 24 **258 103.19 1031.94 43% @ D CR XON/XOFF WXMODEM DN 12 **363 73.34 733.44 61% CR D CR XON/XOFF WXMODEM DN 24 **193 137.95 1379.48 57% @ D CR XON/XOFF *7:8,1:0 End-to-End flow control on KERMIT UP 12 **597 44.60 445.96 37% CR D CR XON/XOFF KERMIT UP 24 **298 89.34 893.42 37% @ D CR XON/XOFF KERMIT DN 12 **389 68.44 684.42 57% CR D CR XON/XOFF KERMIT DN 24 **196 135.84 1358.37 57% @ D CR XON/XOFF *7:8,1:0,5:1,12:1 XMODEM UP 12 **734 36.27 362.72 30% CR D CR NONE XMODEM UP 24 **424 62.79 627.92 26% @ D CR NONE XMODEM DN 12 **638 41.73 417.30 35% CR D CR NONE XMODEM DN 24 **373 71.38 713.78 30% @ D CR NONE *7:8,1:0 ============================ * Optional PAD parameters which optimize performance **File size = 26624 ***File size = 45560