Simply put, we are joining two sites that are 200m apart using 8 core OM4 Multimode
fibre cable. Am I correct in understanding that lc patch panels on either end of the fibre cable are ok and that relying on the SFP modules on lc sfp networked devices on either end to transmit and receive their relevant signals is sufficient?
Or is there any requirement to "boost" the fibre signal at any point with say extra sfp modules within the link?
All current devices on the existing network like switches etc have lc connections.
-
1There's no yes or no answer. It depends on the attenuation of the fiber infrastructure and the type of SFPs. What type of SFPs are you using?– Ron TrunkJul 15, 2019 at 15:37
-
Your cable installer should be able to determine if it fits in your loss budget.– Ron Maupin ♦Jul 15, 2019 at 17:02
-
1Are you actually running Fibre Channel storage protocol or rather just Ethernet?– Zac67 ♦Jul 15, 2019 at 17:28
-
We will have 2 IBM TS3200 LTO Tape drives , one on each site for cross backups. These are linked via a fibre switch with Agilent AFBR-57R5AP SFP modules which are good for 300m with fiber core @ 62.5nm . These will be on a separate fibre circuit to the main ethernet circuit which will use Startech 10GBase-SR SFP+ sfp modules. This is all Multimode shortwave DWDM fibre. Either end of the fibre cable will terminate into a LC patch panel (LC Couplers ) so I am concerned about signal loss at these LC couplers on the patch panels.– Squid_ViciousJul 16, 2019 at 9:18
-
Those FC SFPs are only 4 Gbit/s max, aren't they? Can't really say for sure for FC but it's pretty robust (300m @ 62.5 µm is robust), so I wouldn't worry about the patches too much. Do make sure you don't change the run's diameter though!– Zac67 ♦Jul 17, 2019 at 12:25
Add a comment
|
2 Answers
OM4 is good for 400 m with 10GBASE-SR or for 550 m with 1000BASE-SX, according to the IEEE standards. (These can often be "extended" but there's no guarantee.)
The link power budget isn't usually a problem with short-range multi-mode fiber (MMF), so you should be good with patching in between - provided the fiber and patch cable quality is decent, and all connectors are clean.
[Edit] Since you seem to be using legacy (FDDI) 62.5 µm fiber: don't patch those to or with more modern 50 µm cables. The cable parameters change causes a significant power drop which will likely make the whole run fail. If part of the entire path is 50 µm and another part is 62.5 µm, I'd use another switch with SFPs (or a suitable media converter) in between to adapt.
Mixing FDDI, OM1, and OM2+ is generally not advised. It works with rather short runs but won't on longer ones. Changing the MMF grade can severely degrade the signal.
Mixing OM2 to OM5 isn't generally such a problem but still decreases the reach somewhat.
DWDM might cost you some reach as well. The optical multiplexing is not completely lossless. Still, 200 m shouldn't be a problem.
This will depend on the quality of your fiber, patch panels, patch cords, and specific SFP optics. Several examples from my collection of aging ethernet capable FC optics (1/2/4G):
Ethernet1/15
transceiver is present
type is 1000base-SX
name is AGILENT
part number is HFBR-5710L
nominal bitrate is 1200 MBit/sec
Length_50u : 0x37 550 Meters
Ethernet1/15
transceiver is present
type is Unknown Type-(unknown)
name is FINISAR CORP.
part number is FTRJ8519P1BNL
nominal bitrate is 2100 MBit/sec
Length_50u : 0x1e 300 Meters
Ethernet1/15
transceiver is present
type is Unknown Type-(unknown)
name is Intel Corp.
part number is TXN31115D100000
nominal bitrate is 4300 MBit/sec
Length_50u : 0x0f 150 Meters
(Note: all of these optics, and the cisco n3k they're in, pre-date OM4, they have no data for OM4 lengths)
And a 10G optic:
Ethernet1/1
transceiver is present
type is 10Gbase-SR
name is OEM
part number is 10GB-SFP-SR
nominal bitrate is 10300 MBit/sec
Link length supported for 50/125um OM2 fiber is 82 m
Link length supported for 62.5/125um fiber is 26 m
Link length supported for 50/125um OM3 fiber is 300 m
