. This document answers the most frequently asked questions (FAQs) related to LAN-2-LAN Software.
A. This is because propagation delays on typical circuits are 1 millisecond per 100 miles. When you use the NetWare ping-pong file transfer protocol, each transaction suffers the round trip propagation delay. This delay can be further exaggerated because circuit miles can be more than twice as long as straight-line miles.
For example, a line from Oregon to New Jersey was routed through Seattle, Los Angeles, Chicago, and Ohio—a circuit distance of 5,402 miles. A round-trip propagation delay on this circuit is 108 milliseconds.
This means that, for such a line, if you use the NetWare ping-pong file transfer protocol, regardless of the line speed and data compression effectiveness, and ignore any other delays, fewer than 10 transactions a second can occur. If each transaction results in a payload of 512 bytes, the maximum possible transfer rate for a file on this line is approximately 5,000 bytes/second.
As an example, it took 40 seconds at 56 Kbps with data compression to execute PCONSOLE when the system was set up locally with no propagation delay. Because of the propagation delay, this time expanded to 170 seconds when connected 5,400 miles across country.
Excessive delays can be alleviated if you use the large-packet feature provided by LAN²LAN. If you use Ethernet, the Packet Size used by your NICs is probably 1024. If you configure Packet Size to be 1024 on your LAN²LANs (see your user guide), you can expect to substantially lower your file-transfer time. The exact amount depends upon your line speed, use of compression, and circuit mile length of your line. If you are a Thomas Conrad ARCNET or Token Ring user, you possibly can increase the packet size even further.
A different solution to the problem will be provided by the Novell Packet Burst NLM when it becomes available.
A. If you need to investigate excessive delays, these suggestions are helpful:
Calculate your expected time per file transaction (in milliseconds) which is given by:Request transmission time + Response transmission time + Two way Propagation delay + other delays
Request transmission time is given by:File request size (34 bytes)/(line speed in bps/8) * 1000
Response transmission time is given by:File response size (552 bytes)/(line speed in bps/8) * 1000
Two-way propagation delay is given by:(Circuit miles one way * 2)/100
Other delays = 5 milliseconds
A. PC²LAN hangs after the USERID is entered during a load of the network version of Word Perfect v5.1 for Windows, because WP 5.1 grabs the COM1 interrupt from PC²LAN. A workaround for this problem is to move PC²LAN to the COM2 port or use the /x option on the WP 5.1 command line to set WP 5.1 back to its default values for the setup feature.
If you experience problems when you load similar applications in a PC²LAN, try to use another COM port. If your application uses the COM port that you have selected for PC²LAN, check to see if there is a way to disable or change this selection.
A. You receive a message similar to this (messages can vary with earlier NetWare releases):
Primary interrupt controller detected a lost hardware interrupt, or
Secondary interrupt controller detected a lost hardware interrupt
You can disable these messages with this SET command:SET Display Lost Interrupt Alerts = OFF
The problem might not occur with all revisions of the motherboard and processor module. Novell has informed AST, and the problem is under investigation.
A similar problem was experienced when you use an ALR VIESA 486/33. The workaround was to substitute an IDE.DSK driver for the standard ISA.DSK disk driver.
A. LAN²LAN makes minimum demands of the modem/DSU to which it is attached and provides all the outgoing signals required by normal modems and DSUs.
LAN²LAN activates output signal DTR (TR for RS422) continuously from initiation time.
LAN²LAN requires the presence of the incoming modem signal DSR (DM for RS422) before it activates the line. When DSR is present, LAN²LAN activates output signal RTS (RS for RS422), begins to send flags to the line, and starts the line protocol.
LAN²LAN also requires modem-supplied transmit and receive clocks to clock data in and out. All other incoming modem signals are ignored.
LAN²LAN dynamically senses the line speed by means of periodic test transmissions. LAN²LAN adjusts its protocol timer in accordance to the measured line speed.
Lines between LAN²LANs can be implemented with a primary leased line and a backup dialup line. A typical example might be a leased line that operates at 56 Kbps, backed up with a 9600 bps dial line. The switch between primary and secondary can be activated by external equipment. In this situation, if you assume a switch-over period of a few seconds, LAN²LAN detects a short line outage and continues operation on the new line with the line protocol that protects against data loss or corruption. LAN²LAN adjusts its protocol timer for the different line speed at the next periodic speed sample.
Q. What is the problem and associated solution for when you use Cisco Systems' WNIC revision "D" and "E" adapters in AST Symmetric Multi Processor (SMP) 486/50 machines?
When you use NetWare version 2.x or version 3.11, the machine ABENDs with a non-maskable interrupt (NMI). An error message similar to this appears on the console screen:System Halted Tue. Dec. 1, 1992 8:40:00 pm ABEND: Non-maskable Interrupt (Error Code XXXXXXXX)
The problem occurs only with the AST SMP 486/50.
On early releases of the AST SMP 486/50, I/O bus refresh is disabled. I/O bus refresh must be enabled in order for the WNIC to properly use its 512 Kbytes of memory. Later versions of the AST SMP 486/50 are shipped with I/O bus refresh enabled.
Run the EISA configuration program in order to enable the I/O bus refresh. In the configuration program, in the System AST Manhattan SMP System Board menu, I/O bus refresh can be enabled/disabled. I/O bus refresh must be set to enabled.
A. PC 386 Motherboards that use the SYMPHONY CHIP SET with a chip identified as 461 and marked v073 do not operate properly when you do16-bit transfers on odd byte boundaries.
This deficiency results in the incorrect operation of LAN²LAN when installed in a PC that uses the chipset and causes the WD diagnostic to fail during the High Speed Ports test.
If you experience a problem when you run LAN²LAN in a 386 clone, run WD. If you get an error code in the 4141-4147 range, inspect your PC to see if it contains the bad chip set.
The Cisco Support Community is a forum for you to ask and answer questions, share suggestions, and collaborate with your peers.
Refer to Cisco Technical Tips Conventions for information on conventions used in this document.