FOA CFOS/DAS study materials
FOA CFOS/DAS study resources

Campus Backbone: Reusing Legacy OM2 Versus Installing Singlemode

FOA CFOS/DAS · Put your knowledge to work in context.

All case studies

Campus Backbone: Reusing Legacy OM2 Versus Installing Singlemode

Ridgeline University needs a 10 Gb/s backbone among four buildings. The fiber route from the core switch in Building A to Building D is 1,850 m and passes splice vaults at Buildings B and C. The existing 2004 plant is 62.5/125 µm OM2 in a 100 mm conduit at 40% fill, with two 90° bends and a 1.2 m elevation change. The IT director wants a 40 Gb/s-capable path within five years and asks whether the OM2 can be reused with mode-conditioning patch cords or must be replaced. Inspection found standing water in two conduit sections; a pull tape is present. The outside-plant budget is $185,000, and cutover must fit an 8-hour outage. Vendors quote 12-strand OM4 at $3.10/m and 12-strand OS2 at $2.40/m; 10GBASE-SR optics are $95 each and 10GBASE-LR optics $210 each.

1. Which medium should be specified for the A–D backbone? Support the choice with a worst-case link-loss calculation at 10 Gb/s and a note on 40 Gb/s migration.

Read the worked answer

Recommendation: OS2 singlemode.

Reach comes first. 10GBASE-SR is specified to 300 m on OM3 and 400 m on OM4; OM2 is worse. The 1,850 m path cannot be rescued with mode-conditioning cords, which exist to let 1000BASE-LX launch into multimode, not to extend 10G reach.

Worst-case loss at 1310 nm:

  • Fiber: 1.85 km × 0.4 dB/km = 0.74 dB
  • Connectors: 4 × 0.3 dB = 1.20 dB (patch panel at each end plus one patch cord each side)
  • Splices: 2 × 0.1 dB = 0.20 dB
  • Total: 2.14 dB

10GBASE-LR has a 6.2 dB budget, leaving 4.06 dB margin, above the usual 3 dB design margin. For 40G, 40GBASE-LR4 reaches 10 km on OS2, while 40GBASE-SR4 is only 150 m on OM4. OS2 cable is also cheaper per metre here, so the only premium is the LR optics, roughly $230 more for the pair, which is trivial against the installation cost.

2. Specify the fiber count, connector type, and splice/termination strategy for the four-building route, and justify each choice.

Read the worked answer

Specify 24-strand OS2, not 12-strand. Twelve strands gives only six duplex channels and leaves no room for a second 40G path, redundancy, or a future building drop. Twenty-four strands gives 12 duplex pairs across four buildings and two splice vaults.

Terminate with LC/UPC at patch panels and fusion-spliced pigtails rather than field-install connectors. Factory pigtails give lower and more repeatable loss, and the 0.3 dB connector allowance is easier to defend. At Buildings B and C, use in-line splice closures and splice through the strands that continue to D; do not terminate and patch through unless the vault is a real access point.

Maintain duplex A-B polarity, label both ends, and test every fiber bi-directionally at 1310 nm and 1550 nm. Keep at least four fibers dark as spares. The splice strategy adds 0.1 dB per splice but removes two connector pairs per intermediate building, a net loss improvement and a reliability gain.

3. The conduit is flooded and the cutover window is 8 hours. List the three highest installation risks and a specific mitigation for each.

Read the worked answer
  1. Water and debris in the conduit. Standing water raises pull friction and can damage non-water-blocked cable. Mitigation: use outdoor water-blocked cable, swab and prove both sections, and pull with a compatible lubricant. If the water is not clean, jet and vacuum the section before pulling.

  2. Pull tension and bend radius at the two 90° bends. The 1.2 m elevation change and bends can exceed the cable's minimum bend radius or tensile limit. Mitigation: install rollers or a back-tension brake, monitor pull tension continuously, and pull from the midpoint toward each end if tension approaches the limit. Leave a visible slack loop at each vault.

  3. The 8-hour outage. Field termination and testing can consume the entire window. Mitigation: pre-terminate and pre-test the cable on reels, pre-stage optics and patch cords, and rehearse the cutover. Have a fallback plan to restore the old link if testing fails, and do not release the old path until bi-directional OTDR and power-meter tests pass.