Configuring a blind point-to-point connection with CLI or Dev Manager

[caparker999]

Hello,

I have a somewhat atypical use case that I'm not sure is readily solvable. Networking is definitely not one of my strengths so I apologize if a solution to this problem is already in the docs or in a community thread and I just haven't found it. I'll try to keep this concise.

Objective

Programmatically unplug one non-standard device and plug in a different non-standard device to a controller PC

Background

One of my customers has a custom device that communicates over fiber channel. They also have a custom integration card that makes use of an SFP and PCI slot to pull data from their device. I have a few dozen of the devices but only one of the integration cards. They are internal use only, so my customer is hesitant to invest in additional cards. I have the task of running these devices concurrently and periodically pulling data from them. 

Their system works well with a direct Device <--> PC connection, where the connection via fiber is direct from their device to their integration card. But unfortunately the card only has one port. I am trying to find a way to avoid manually unplugging and replugging cables all day.

A managed switch seems like an obvious solution, but I've been unsuccessful in making that work. I currently have an old MDS 9124 that I'm experimenting with but I intend to purchase a modern switch if I can demonstrate that a managed switch will do what I need it to do. I have a simple trial set up with both the integration card and a single device plugged in to the switch, and I'm trying to simply pass all traffic from one port to the other, and vice versa.

The major confounding factor is that I have no idea how their communications are set up between the device and card, and it seems they're not too sure either (a classic situation with the foundation of a critical process being written by a long gone employee). I speculate they're using custom frames rather any standard, which I'm guessing prevents the switch from recognizing a device is attached.

Trials

I've tried to set up a single port to port interface by creating a new VSAN, port-channel, and setting each of them to be in F mode. I then add the two ports, fc1/13 and fc1/14 to the new channel and vsan. My fc1 ports continue to show physical link problem (see attached interface.txt), though I know the device and integration card work and the cables I'm using are also good.

This is the origin of my question. I think the card and device are communicating with the switch in an unexpected way, making it unable to forward traffic to the specified port. So, is it possible to forward port traffic without any consideration of the traffic structure, e.g. like a bytestream?

Thanks for taking the time to read this, and I apologize if my intentional vagueness made things unclear. Please let me know if any additional information would be helpful.

All the best,

Charley

[caparker999]

Oops, looks like my attachments didn't stay attached. I'll just list the info here:

INTERFACE

vsan2 is up, line protocol is down (port down)

    WWPN is 10:00:00:0d:ec:cf:49:41, FCID not assigned

    MTU 1500 bytes, BW 1000000 Kbit

    0 packets input, 0 bytes, 0 errors, 0 multicast

    0 packets output, 0 bytes, 0 errors, 0 dropped

port-channel 2 is down (No operational members)

    Hardware is Fibre Channel

    Port WWN is 24:02:00:0d:ec:cf:49:40

    Admin port mode is F, trunk mode is on

    snmp link state traps are enabled

    Port vsan is 2

    5 minutes input rate 0 bits/sec, 0 bytes/sec, 0 frames/sec

    5 minutes output rate 0 bits/sec, 0 bytes/sec, 0 frames/sec

      1 frames input, 148 bytes

        0 discards, 0 errors

        0 CRC,  0 unknown class

        0 too long, 0 too short

      1 frames output, 148 bytes

        0 discards, 0 errors

      0 input OLS, 0 LRR, 0 NOS, 0 loop inits

      0 output OLS, 0 LRR, 0 NOS, 0 loop inits

    Member[1] : fc1/13

    Member[2] : fc1/14

fc1/13 is down (Link failure or not-connected)

    Hardware is Fibre Channel, SFP is short wave laser w/o OFC (SN)

    Port WWN is 20:0d:00:0d:ec:cf:49:40

    Admin port mode is F, trunk mode is on

    snmp link state traps are enabled

    Port vsan is 2

    Receive data field Size is 2112

    Beacon is turned off

    Belongs to port-channel 2

    5 minutes input rate 0 bits/sec, 0 bytes/sec, 0 frames/sec

    5 minutes output rate 0 bits/sec, 0 bytes/sec, 0 frames/sec

      1 frames input, 148 bytes

        0 discards, 0 errors

        0 CRC,  0 unknown class

        0 too long, 0 too short

      1 frames output, 148 bytes

        0 discards, 0 errors

      0 input OLS, 0 LRR, 0 NOS, 0 loop inits

      0 output OLS, 0 LRR, 0 NOS, 0 loop inits

fc1/14 is down (Link failure or not-connected)

    Hardware is Fibre Channel, SFP is short wave laser w/o OFC (SN)

    Port WWN is 20:0e:00:0d:ec:cf:49:40

    Admin port mode is F, trunk mode is on

    snmp link state traps are enabled

    Port vsan is 2

    Receive data field Size is 2112

    Beacon is turned off

    Belongs to port-channel 2

    5 minutes input rate 0 bits/sec, 0 bytes/sec, 0 frames/sec

    5 minutes output rate 0 bits/sec, 0 bytes/sec, 0 frames/sec

      1 frames input, 180 bytes

        0 discards, 0 errors

        0 CRC,  0 unknown class

        0 too long, 0 too short

      1 frames output, 180 bytes

        0 discards, 0 errors

      0 input OLS, 0 LRR, 0 NOS, 0 loop inits

      0 output OLS, 0 LRR, 0 NOS, 0 loop inits

HARDWARE

Software
BIOS: version 1.0.17
loader: version N/A
kickstart: version 4.1(3a)
system: version 4.1(3a)
BIOS compile time: 05/28/09
kickstart image file is: bootflash:/m9100-s2ek9-kickstart-mz.4.1.3a.bin
kickstart compile time: 2/12/2009 19:00:00 [03/06/2009 05:34:54]
system image file is: bootflash:/m9100-s2ek9-mz.4.1.3a.bin
system compile time: 2/12/2009 19:00:00 [03/06/2009 07:48:25]

Hardware
cisco MDS 9124 (1 Slot) Chassis ("1/2/4 Gbps FC/Supervisor-2")
Motorola, e500 with 516136 kB of memory.
Processor Board ID JAF1321AQED

Device name: MDS-9100-ReliabilityLab
bootflash: 254464 kB
Kernel uptime is 0 day(s), 3 hour(s), 24 minute(s), 13 second(s)

Last reset at 322446 usecs after Wed Jul 27 06:53:17 2022

Reason: Reset Requested by CLI command reload
System version: 4.1(3a)
Service:
--------------------------------
Switch hardware ID information
--------------------------------

Switch is booted up
Switch type is : MDS 9124 (1 Slot) Chassis
Model number is DS-C9124-K9
H/W version is 5.1
Part Number is 73-10583-06
Part Revision is A0
Manufacture Date is Year 13 Week 19
Serial number is FOX1319G3QF
CLEI code is COMTU10ARB

--------------------------------
Chassis has 1 Module slots
--------------------------------

Module1 ok
Module type is : 1/2/4 Gbps FC/Supervisor-2
0 submodules are present
Model number is DS-C9124-K9-SUP
H/W version is 5.1
Part Number is 73-10583-06
Part Revision is A0
Manufacture Date is Year 13 Week 21
Serial number is JAF1321AQED
CLEI code is COMTU10ARB

---------------------------------------
Chassis has 2 PowerSupply Slots
---------------------------------------

PS1 fail/shutdown
Power supply type is: 299.88W DC
Model number is DS-C24-300AC
H/W version is 160.0
Part Number is 341-0250-01
Part Revision is A0
Manufacture Date is Year 14 Week 32
Serial number is PAC1432315W
CLEI code is

PS2 ok
Power supply type is: 299.88W DC
Model number is DS-C24-300AC
H/W version is 160.0
Part Number is 341-0250-01
Part Revision is A0
Manufacture Date is Year 13 Week 17
Serial number is PAC131730R3
CLEI code is

----------------------------------
Chassis has 3 Fan slots
----------------------------------

Fan1 ok

Fan2 ok

Fan3 ok

[caparker999]

Hi everyone,

It's finally dawned on me that the term I'm looking for is multiplexing. I'm trying to set up a simple multiplexor that can selectively channel one input to the designated output and ignore all other inputs. This seems like a basic use case but I'm having trouble finding resources on setting up the system. Most articles and posts are about the much more useful WDM compared to just selecting FC inputs. I'm not sure if WDM could solve my use case but it seems like overkill.

I'm still having trouble getting my customer's device to be recognized by the switch. Connections from their integration card and to their device are listed as LinkFailure by the switch. My switch has some old SFPs that I initially thought to be the issue, but when I tried removing the SFP+ used by the integration card and device and replacing them with the same SFPs my switch has there was still a LinkFailure. I tried using the switch SFPs for a direct connection from device to integration card and that worked fine, so I don't think the SFPs are the problem. 

I'm looking into more simple devices that are dedicated to multiplexing, though everything I'm finding is more focused on maximizing bandwidth of a single fiber rather than selecting an input to route to the output. If anyone knows of a device that does what I'm looking for I would greatly appreciate the guidance. I'll update this thread with whatever I find for posterity's sake.

All the best,

Charley

[caparker999 · Accepted solution]

Hello,

I did end up finding a class of devices that suit my needs. While searching it was hard to filter out the regular network switches from the fiber switches that ended up working for me. My search keywords that got the best results were "all optical fiber switch" and "fiber-fiber switch". After some digging I eventually found a few companies that provide this niche product. 

I'm not trying to advertise a specific product so I'll just link a couple of the devices I was considering. I only bought one of these products and I was able to get it to work. I have no reason to doubt that all of these products would work for my situation, as they are all more or less built on the same switching method, I believe. 

Dual 1xN Broad Wavelength Optical Switch Box – Fiber-Fiber™ - Agiltron Inc.
MEMS Special Configuration Optical Switches | DiCon Fiberoptics
1xN/ 2xN/ 4xN Optical Switch Series (senko.com)

Hopefully this will help to save someone some pain in the future. 

All the best,
Charley