Ref: 99980028 Title: Release Note: CS/100 booting from NCS/AT Ver 13000 Date: 5/26/87 Copyright 3Com Corporation, 1991. All rights reserved. SW/100-TCP-NCS/AT Version 13000 is a software diskette for CS/100s that are booted from an NCS/AT. This release memo contains the instructions for installing the CS/100 software on the NCS/AT and describes the features and bugs fixed in this version of SW/100-TCP-NCS/AT. Installing Software on the NCS/AT Because SW/100-TCP-NCS/AT Version 13000 is stored on a 96 TPI diskette, you must use a blank, 96 TPI, double-sided, high- density diskette if you are making a backup copy of the distribution software. If you want to boot the CS/100 from an NCS/AT, you must first copy the software from the SW/100-TCP-NCS/AT Version 13000 diskette to the NCS/AT disk. The procedure is described below: 1. Follow the instructions in the NCS/AT Installation and Operation Guide to log in to the NCS/AT. A command prompt will appear on the screen after you have logged in. 2. Insert the SW/100-TCP-NCS/AT Version 13000 diskette into the disk drive of the NCS/AT. 3. At the command prompt, type: utility floppyread The Floppyread utility will copy the software from the SW/100-TCP-NCS/AT diskette to the NCS/AT disk. 4. As the file is copied to the disk, its name is displayed on the screen. If floppy read errors or disk write errors occur, remove and replace the diskette; then repeat the procedure starting at step 2. If read or write errors continue to occur, or if other error messages appear, contact an authorized Bridge service representative. 5. When the file has been successfully copied, the command prompt appears. Remove the diskette. Compatibility SW/100-TCP-NCS/AT Version 13000 operates with the following products and software releases: .br;NCS/150 running SW/150-NCS-TCP Version 10000 or later .br;NCS/AT running SW/AT-NCS Version 10000 or later .br;CS/1 running SW/1-TCP Version 10110 or later .br;CS/1-SNA running SW/1-SNA-TCP Version 10000 or later .br;GS/3-IP running SW/3-IP Version 10000 or later .br;CS/200 running SW/200-TCP Version 13000 or later .br;IVECS running SW/IVECS-TCP Version 13000 or later .br;TCPTerm Version 10000 or later Table 1 illustrates the software and firmware compatibility for SW/100-TCP-NCS/AT Version 13000: '' _______________________________________________ | Table 1 Software/Firmware Compatibility | |______________________________________________| | Product| Software | Firmware | | | | | | | | SIO MP | | | | | | CS/100A| SW/100-TCP-| T1ASYN 1B T1MMON 1B| | | NCS/AT | or later or later | |________|_____________|_______________________| New Features SW/100-TCP-NCS/AT Version 13000 has the following features, which were not present in SW/100-TCP Version 12000 or earlier: 1. The NoMacroBreak or MacroBreak field has been added to the InterAction parameter. The NoMacroBreak option prevents users from breaking out of a macro while the macro is in the process of execution. When a NoMacroBreak option is set, the BREAK key is ignored. This feature can be used to prevent a user from breaking out of a macro in Global Network Manager privilege. 2. The BUffersize parameter in the port configuration can be set to a value smaller than the system buffer size. This feature is useful for PC-to-host file transfer applications that require smaller buffer sizes. 3. Whenever a disconnection from a host port is requested, either by a user typing the DisConnect command on the terminal or by the AutoDisconnect timer, the LogoffString will be transmitted to the host port before it closes down the circuit. LogoffString transmission is controlled by the string defined in the LogoffString parameter of the host-end server. The LogoffString can be up to 24 characters long. In order for the LogoffString to be transmitted, the AUtoLogoff parameter should be set to ON. This parameter can be set on a per-port basis. The LogoffString will not be transmitted if the disconnection is caused by the following conditions: a) A network manager uses the Listen command to put the host or terminal port in Listen mode. b) Traffic to the host port is totally flow controlled, and transmission of the LogoffString is not possible within a reasonable amount of time (currently six seconds). c) A user powers off the terminal before disconnecting the session and the UseDTRin parameter is set to AsDTR on the terminal port. The LogoffString and AUtoLogoff parameters have the following syntax: a) To set the LogoffString, enter: SETDefault LogoffString = "string" The "string" can contain any control characters; for example, "^M" for control-M, "^J" for control-J. b) To enable/disable AUtoLogoff on a particular port, enter: SETDefault (!portid) AUtoLogoff = [ON | OFF] Note that the AUtoLogoff parameter is only available on a host port. c) To display the AUtoLogoff parameter for a particular port, enter: SHow (!portid) AUtoLogoff d) To display the LogoffString, enter: SHow LogoffString 4. Permanent Virtual Circuits (PVCs) support is now available. Permanent Virtual Circuits allow automatic circuit reestablishment upon disconnection. Once a permanent virtual circuit is defined, it will be reestablished whenever the server is rebooted. After the PermanentVC parameter is set, the first connection attempt will begin in approximately 30 seconds. If the connection attempt fails or if the permanent virtual circuit is disconnected for any reason, the system will attempt to reestablish the circuit every 60 seconds. PVCs can only be defined on a host port. The destination end of the connection can be either a terminal or host port. The PVC may be defined on both ends of the circuit. In addition, PVCs may accept names. PermanentVC parameters have the following syntax: - SETDefault (!portid) PermanentVC = "
" Defines a permanent virtual circuit between a local host port and any port on the network.
can be either a physical address or a name. The name can be up to 12 characters long. - SETDefault (!portid) PermanentVC = "" Removes any previous PVC definition, but does not disconnect any existing connection. The existing circuits can be disconnected using the DisConnect or Listen command. - SHow (!portid) DefaultParameter PermanentVC Displays the current permanent virtual circuits. 5. The SHow AllSessions command has an option allowing it to display specified physical addresses or logical names of destination devices when a connection is established. For example, to display physical addresses of destination devices, use the following command: SHow AllSessions p To display logical names of destination devices, use the following command: SHow AllSessions 6. A sysgen parameter "Number of macro buffers" has been added. This parameter is useful when the macro service is used extensively. The default value for this parameter is 4 and the maximum value is 14. This parameter value determines the number of users that can execute macros simultaneously without seeing the "Can't - no memory" message. For example, if this parameter is set to 14, then 14 users can execute macros simultaneously. Note that the greater the number of macro buffers, the smaller the number of sessions in the systems. 7. The ARP-based routing has been improved. When a packet is transmitted to a destination with a different network number in its address, an initial route is set up using the default gateway. If the default gateway is zero when such an initial transmission is made, SW/100-TCP-NCS/AT Version 13000 will attempt to resolve the address on the local segment via an ARP request. 8. User Telnet has been enhanced. During setup negotiation, User Telnet now agrees to DO ECHO if instructed by the host. This feature is useful in offloading processing from some hosts whose TELNET implementations might not echo keyboard input (e.g., a TELNET that is aware of entry/exit from a screen- based editor). 9. If the BReakAction parameter is set to InBand or OutofBand or both, a TELNET Interrupt Process command will be sent. If the BReakAction parameter is set to FlushVC, a TELNET DO- TIMING-MARK will be sent and all incoming data will be flushed until the mark is echoed. FlushVC is useful in preventing long displays from a dead host process. The FlushVC, however, depends on the TELNET timing mark being returned. Some systems may not return this mark reliably. In particular, Bridge has experienced this problem with a VAX running 4.2 BSD UNIX. In a VAX running 4.2 BSD UNIX, we found that a second timing mark sent by a second BREAK was usually returned. 10. The Connect and REMOTE commands reject invalid Internet addresses. The SW/100-TCP-NCS/AT Version 13000 checks the legality of Internet address assignments (i.e., any of the InternetServer addresses and the InternetPort addresses) as follows: - Any address may be assigned 0.0.0.0, indicating no assignment (all addresses start out with this value). - No address may be assigned a broadcast or loopback address (host portion all 0s or all 1s). - The DefaultGateway may not be the local address. - If the server is booted from the Network Control Server, neither the PrimaryNameServer nor SecondaryNameServer may be the local address. 11. A new port configuration parameter "NetAScii" has been added. Its values are UseLF and UseNUL. This parameter allows the user to choose the character sequence transmitted by the server when a is entered. If the "NetAScii" parameter is set to UseLF, a will be transmitted. The "NetAScii" parameter is used only by a terminal (active) port; it has no effect on the host (passive) side of a connection. The "NetAScii" parameter has been implemented to provide for the variety of TELNET implementations that interpret the TELNET data stream differently. For hosts that correctly view and as encoded Network Virtual Terminal functions (e.g., 4.3 BSD UNIX), Bridge recommends that the "NetAScii" parameter be set to UseNUL. When the "NetAScii" parameter is set to UseNUL, it allows as a . For hosts that neglect to strip a following a , Bridge recommends that "NetAScii" be set to UseLF. As a consequence of this new parameter addition, the EOM parameter has been deleted. Bugs Fixed The following problems, which were present in SW/100-TCP Version 12000 or earlier, have been corrected in SW/100-TCP-NCS/AT Version 13000: 1. If a question mark is entered at the end of a SETDefault InternetPort (setd ip = ?) command, the syntax for this command will be displayed. 2. When creating a macro to establish a session with ECM and RESume after transmitting a string, the macro buffer is no longer lost if the connection fails due to a busy or disabled destination port. 3. The SHow STATisticS command now displays correct information. Known Problems The following is a list of current problems in SW/100-TCP-NCS/AT Version 13000: 1. The message "Remote is Busy" appears when the remote system is out of memory. 2. If the host crashes and the user continues to enter characters on the terminal, the terminal port will hang. After a few minutes, the terminal port will be reactivated automatically. However, this terminal port can also be reactivated by being put in Listen mode from another terminal. 3. If the remote-end of an inactive session is disconnected, the SHow SESsions and SHow AllSessions commands still show that the local-end is connected. The disconnection message only appears if a user resumes that session. That problem does not occur on an active session. 4. If a wrong password is entered in a SET PRIVilege command, this information is not forwarded to the NCS/AT Audit Trail. 5. If the same Internet address is defined for different physical ports, there is no error checking message to inform users. As a result, when a first connection is established to this Internet address, the lowest physical port number will be connected; the second connection request to the same Internet address will cause the message "remote is busy" to appear. 6. On a busy network with high cable utilization rate, the CS/100 may experience difficulties booting from an NCS/AT. If this occurs, please contact an authorized Bridge service representative for assistance. User Interface SW/100-TCP-NCS/AT Version 13000 includes the following user interface commands: - SETDefault (!) InternetPort = Assigns Internet addresses to individual physical or rotary ports. - SETDefault (!) InternetPort = 0.0.0.0 Removes the Internet address of an individual physical or rotary port. - SETDefault DefaultGateway = Allows all remote net packets to be sent to the specified gateway. - SETDefault PrimaryNameServer = Assigns an Internet address for the primary Name Server. The primary Name Server will be queried first for name resolution. The primary Name Server's Internet address can be defined to be on a remote network. - SETDefault SecondaryNameServer = Assigns an Internet address for the secondary Name Server. The secondary Name Server will be queried if the primary Name Server does not yield a name to address resolution. - SHow InternetPorts Replaces the SHow InternetMap command. It displays the Internet address of the Communications Server itself and all other Internet addresses associated with the Communications Server (physical and rotary ports). - SHow InternetServers Displays Internet addresses of the Name Servers and Default Gateway. - SHow VirtualPorts Displays the connection state (passive or active) of the physical port to which these virtual ports are connected. - SHow ARptable Displays the mapping between Internet addresses and Ethernet addresses of devices that communicate with the CS/100. Note that the ARP table always contains the unit's Internet address and the broadcast address. - ZeroArp Cleans up the contents of the ARP table. Normally, the software automatically removes an ARP entry if it does not communicate with the device within 15 minutes. However, if the device has moved to a different location, this command can update the information in the ARP table faster. Documentation The current manuals for the CS/100 running SW/100-TCP-NCS/AT are listed below: + Series/100 Installation Guide (10/85) .br;(Part number: 09-0023-02) + Connection Service User's Guide (8/86) .br;(Part number: 09-0063-01) + Connection Service User's Quick Reference Guide (12/85) .br;(Part number: 09-0062-00) + Network Management Guide (1/87) .br;(Part number: 09-0067-01) + Configuration Guide (1/87) .br;(Part number: 09-0093-00) + Getting Started Guide (1/87) .br;(Part number: 09-0043-01) If you have any questions, please contact an authorized Bridge service representative for assistance. Trademarks UNIX is a registered trademark of AT&T. VAX is a registered trademark of Digital Equipment Corporation. APPENDIX SUBNETS Overview If a private network wishes to establish communications to a public TCP/IP network (e.g., ARPANET), the public network first assigns a unique network number to the private network. In this case, all stations on this private network will have this unique network number. Gateways internal to this network may be used to further subdivide the network into segments by incorporating subnetworking as described below. Definition of Subnets and the Subnet Mask An Internet address consists of 32 bits divided into four 8-bit subfields. Normally, these subfields are divided between the and fields. Some TCP/IP networks, however, have inserted an additional field between the and fields called . The field is used to indicate the particular physical segment. The field is formed by taking the leading bits from the field. The field is assigned in the system program under the "Server subnet mask" option in the "IP Parameters" menu. However, if your network does not use subnets, do not assign a subnet mask. When entering the subnet mask, enter the decimal instead of binary representation. The subnet mask is then defined as follows: It is a 32-bit number divided into four equal subfields. Each bit that coincides with the network field (the 8, 16, and 24 most significant bits for a class A, B, and C Internet address, respectively) is always set to 1. The remaining bits in the host field that are set to 1 define the subnet field to be incorporated. The subnetwork address is then obtained by performing the logical AND operation of the Internet address with the selected subnet mask field. For the subnet addressing to be meaningful, all servers on this private network must be sysgenned with identical subnet masks. Example 1: The subnets of a class A Internet address might be defined as follows: 126.(11110000) (00000000) (00000000) In this example, the number of subnets defined on this network is 16 and the subnet mask is (11111111) (11110000) (00000000) (00000000), which is the binary equivalent of 255.240.000.000 decimal. The possible subnetwork addresses for this network would then be: '' 126.0.0.0 126.128.0.0 126.16.0.0 126.144.0.0 126.32.0.0 126.160.0.0 126.48.0.0 126.176.0.0 126.64.0.0 126.192.0.0 126.80.0.0 126.208.0.0 126.96.0.0 126.224.0.0 126.112.0.0 126.240.0.0 Example 2: The subnets of a class B Internet address might be defined as follows: 129.000.(11100000) (00000000) In this example, the number of subnets is 8 and the subnet mask is (11111111) (11111111) (11100000) (00000000), which is the binary equivalent of 255.255.224.000 decimal. The possible subnetwork addresses for this network would then be: '' 129.0.0.0 129.0.128.0 129.0.32.0 129.0.160.0 129.0.64.0 129.0.192.0 129.0.96.0 129.0.224.0 Example 3: The subnets of a class C Internet address might be defined as follows: 193.129.130.(11000000) In this example, the number of subnets is 4 and the subnet mask is (11111111) (11111111) (11111111) (11000000), which is the binary equivalent of 255.255.255.192 decimal. The possible subnetwork addresses for this network would then be: '' .br;193.129.130.0 193.129.130.128 .br;193.129.130.64 193.129.130.192 Note that in the three examples above, the numbers in parentheses are in binary format.