HANK@TAUNIVM.BITNET (Hank Nussbacher) (02/25/88)
I wish to thank all those who have made invaluable comments on how to improve the scorecard. I have tried to use most of people's comments, but sometimes I decided to not include a hardware card or IP function, simply because I felt there was not enough of a pressing need to include that information (each person has their own subjective feelings). I have also rearranged items, moving some items to different tables. Please review the new Scorecard and even if you sent me a filled in line, please send me an updated Scorecard for the products you know. Please try to use the form below and try to stay away from listing longhand what features the product has or doesn't have. It is only 50% complete and has a long way to go. Please take the time fill in some boxes and send them my way. I and all other people looking into alternative Pc/Ip solutions will be eternally grateful. Thanks, Hank The Pc/Ip Scorecard revision 2: 02/25/88 --------------------- We are the process of trying to decide which Pc/Ip implementation would be the best. After asking around the overwhelming reply I have received has been, "If you ever find a comparison study, I would love to see it too". That is why I have decided to create the Pc/Ip Scorecard. This scorecard will be like a PC Magazine analysis of of hard disks or printers. But I need help in filling in the boxes. So, here is what the scorecard looks like. Please send me your replies and I will integrate all answers and comments and publish the finalized scorecard in the weeks to come. All Pc/Ip implementations support IP, TCP, FTP and Telnet. The question then is to divide other IP protocols into categories of "must have" to "nice to have but not necessary". This first table is called "Must Have". Vendor TFTP POP ICMP SMTP VT- 3270 FTP ARP UDP max cost 100 FTP ($) -----------+----+----+----+----+----+----+----+----+----+----+----+ Beame | Yes| No | Yes| No | | No | | Yes| Yes| | | Bellcore | | | | | | | | | | | | CMC | | | | | | | | | | | | CMU | Yes| No | Yes| No | | No | No | Yes| Yes| | 0| Cornell | | | | | | | | | | | | Excelan | | | Yes| | | | | Yes| Yes| | | FTP | Yes| No | Yes| Yes| | Yes| | Yes| Yes|150k| 400| FUSION | | | | | | | | | | | | IBM | Yes| Yes| Yes| Yes| | Yes| | Yes| Yes| | | KA9Q | No | No | Yes| Yes| | No | | Yes| Yes| | 0| MIT | Yes| | Yes| | | | No | Yes| Yes| | 50| NCSA | No | No | Yes| No | | No | | No | No | 30k| 0| Stanford | No | Yes| Yes| Yes| | No | | Yes| Yes| 50k| 100| SUN | No | Yes| Yes| Yes| | Yes| | Yes| Yes| | 300| UB | | | Yes| | | | | Yes| Yes| | | Wollongong | Yes| No | | | | | | | | 25k| 395| The "max FTP" column is for the fastest FTP to a Pc (*not* from) seen by a user (in Kb/sec). It makes no difference in this table which machine was at the other end (obviously the faster the machine at the other end - the better). The following table lists the most popular Ethernet cards available and whether the Pc/Ip implementation works with the stated card. Vendor 3com Excelan Interlan UB WD 3Com UB NIC 3C501 EXOS205 NI5010 2273A 8003 3C523 PS/2 -----------+-----+--------+--------+-----+-----+-----+------+ Beame | Yes | No | No | No | | | | Bellcore | | | | | | | | CMC | | | | | | | | CMU | Yes | No | Yes | No | | | | Cornell | | | | | | | | Excelan | | | | | | | | FTP | Yes | Yes | Yes | No | Yes | | No | FUSION | | | | | | | | IBM | Yes | No | No | Yes | Yes | Yes | Yes | KA9Q | Yes | No | No | No | | | | MIT | Yes | No | Yes | No | | | | NCSA | Yes | No | No | No | | | | Stanford | Yes | No | No | No | Yes | Yes | | Sun | Yes | No | Yes | Yes | | | | UB | | | | | | | | Wollongong | Yes | | | | | | | This table is called the "Nice to Have" table. The functions listed here are not mandatory but are useful in a Tcp/Ip environment: Vendor name time fing whoi NFS gate srce Net ping SLIP srvr way code BIOS -----------+----+----+----+----+----+----+----+----+----+----+----+----+ Beame | Yes| Yes| Yes| No | | | | | Yes| No | | | Bellcore | | | | | | | | | | | | | CMC | | | | | | | | | | | | | CMU | Yes| Yes| Yes| No | No | No | Yes| | Yes| Yes| | | Cornell | | | | | | | | | | | | | Excelan | | | | | | | | | | | | | FTP | Yes| Yes| Yes| Yes| | | No | | Yes| Yes| | | FUSION | | | | | | | | | | | | | IBM | Yes| Yes| Yes| Yes| No | Yes| | | Yes| No | | | KA9Q | | | | | | | Yes| | No | Yes| | | MIT | No | | Yes| | | | Yes| | Yes| | | | NCSA | Yes| No | No | No | | | Yes| | Yes| No | | | Stanford | No | Yes| Yes| Yes| No | No | No | | Yes| No | | | Sun | Yes| Yes| Yes| Yes| Yes| | | | Yes| Yes| | | UB | | | | | | | | | Yes| | | | Wollongong | | | | | | | | | | | | | FTP Software is OEMed to BICC ISOLAN, Fibronics and Proteon.
jbvb@VAX.FTP.COM (James Van Bokkelen) (02/26/88)
FTP Software's PC/TCP has both H19 and VT100 emulators. We are working on a VT220 (60 days or so). PC/TCP also has an FTP client, and a server (two separate programs). The 150K bytes transfer rate is between our client and our server, on two 8Mhz ATs, from RAM disk to NUL. PC/TCP supports the U-B 2273 (PC-NIC). As Joe Cimmino pointed out, the U-B NIU is part #2261, and that is the only card U-B themselves supported, as of the last I knew (their TCP is resident on the board). When version 2.02 of PC/TCP ships, an RFC-conforming NETBIOS will be available ($80 extra) as an add-on. This is presently in beta test, and is reported to interoperate with Excelan, U-B, Syntax and Bridge. (probably about 30 days). jbvb
jbvb@VAX.FTP.COM (James Van Bokkelen) (03/02/88)
Reassembly doesn't help you much unless you have an Ethernet board with multiple packet buffers and relatively fast logic. We added reassembly support (only two fragments, resulting packet must be smaller than the network MTU) for someone who had a Bridge GS-7? that fragmented everything bigger than 500 bytes of IP length. The result works, and it is in the code we ship, but it was just about useless with the 3C500s he had. The problem is that gateway developers fragment the packet immediately before sending it, and they pride themselves on how fast they make their Ethernet interfaces go. You can't get a single-buffered interface ready in time for the 2nd fragment, period. Pairs of fragments are very good ways of determining just how good your multiple-buffered interface is. I don't have high-precision instrumentation to measure intervals, but I don't believe that any of the PC interfaces available today can grab two packets as fast as some other vendors send them. I have also heard criticism (possibly inaccurate) of some Unix vendors for sending packets closer together than the Ethernet spec allows. I think generating fragments from a PC is a bad idea, and should be avoided. Unless you can arrange to send a re-transmission with the same IP ID field, sending fragments doubles the dropped-packet rate once for each fragmentation. We deal with this issue by requesting a smaller TCP MSS when a connection is routed off the local subnet. Our TCP window is also settable. Philip Prindeville's version of the CMU code would seem to have to generate fragments, because he supports TFTP on ARCnet, with its 508-byte MTU, but I haven't looked at it or used it. Speaking of subnets, I think you should have a "supports subnets" column. Our code does, and I know that Excelan, at least, doesn't. MIT-derived code ought to support them, vendors with older on-board TCPs are less likely to. I also think that 'subnets' settable on 8-bit boundaries only is not "subnet support". IP frag Subnet FTP | reass | yes |..... James B. VanBokkelen FTP Software Inc.
BLASCO@ICNUCEVM.BITNET ("A. Blasco Bonito") (03/02/88)
Hank, I think the scorecard is missing an important element: the ability or inability of the different packages to reassemble IP fragments. This feature is relevant in general if one wants to estabilish connections traversing different networks and gateways which could fragment an IP message. My knowledge is that KA9Q can reassemble, MIT can't, the latest version of CMU (or a patch to CMU made by somebody else, I don't remember) can, Stanford can not. Best regards ---------- ---------- A. Blasco Bonito Internet: Blasco@cnuce-vm.arpa CNUCE - Istituto del CNR Earn/Bitnet: Blasco@icnucevm Reparto Reti e Sistemi Distribuiti Tel: +39 (50) 593246 Via S. Maria, 36 Telex: 500371 CNUCE I 56100 PISA Italy Fax: +39 (50) 576571 ---------- ----------
Philip Prindeville@UDEL.EDU, philipp@larry.mcrcim.mcgill.EDU (03/06/88)
Actually, reassembly and fragmentation is very handy when you are going over a slow, 'small' link, such as a serial line used to connect two LAN islands. In such a case, the slower speed of the line will space out the packets, possibly giving you time to recover for the subsequent fragments. And you might not have any way of knowing that the 'small' network separates you and your destination, especially if it is not directly connected, and therefore can't negotiate the MSS. -Philip N.B.: 'small' means less than MTU of directly connected network.
davidc@TERMINUS.UMD.EDU ("David R. Conrad") (03/19/88)
Sorry this took so long, first some corrections/clarifications using John Romkey's suggestion of yes meaning both client and server, client meaning just client, server meaning just server and no meaning neither. Table 1: Vendor TFTP POP ICMP SMTP VT- 3270 FTP ARP UDP max cost 100 FTP ($) -----------+----+----+----+-------+----+----+-------+----+----+----+----+ IBM | Yes| Yes| Yes| Client| No | Yes| Client| Yes| Yes|114*| ** | * Max FTP rate is 114 KBytes/Sec on PS/2M60 from Vaxstation2000 Ultrix 2.0 to NUL in PC-DOS 3.3 ** Anywhere from about ~$150 to ~$300 depending on quantity. Table 2: Vendor 3com Excelan Interlan UB WD 3Com UB NIC 3C501 EXOS205 NI5010 2273A 8003 3C523 PS/2 -----------+-----+--------+--------+-----+-----+-----+------+ IBM | Yes | No | No | Yes | No | No | Yes | Table 3: Vendor name time fing whoi NFS gate srce Net ping SLIP srvr way code BIOS -----------+------+------+------+------+----+----+----+----+----+----+ IBM |Client|Client|Client|Client| No | Yes| ***| No | Yes| No | *** If you have a VM license yes. Some possible additions: rexec, lpr, network monitoring (IBM's has these) The rest of the r's, programmer's toolkit (socket library). -drc