Data Center Migration: 1,152 Links in Three Weeks
A colocation provider is migrating 1,152 duplex LC fiber links from an old meet-me room to a new meet-me room in the same building. The new path uses 288-fiber preterminated trunk cables with MPO-24 to LC harnesses; 90 percent of links are preterminated, and 10 percent must be field-terminated because of route changes. The work must be completed in three weeks during a change freeze that permits only four-hour maintenance windows per night. Two test crews are available; each can test about 40 links per shift with an OLTS. The customer requires 100 percent Tier 1 testing and Tier 2 OTDR on any failed link. Labeling must match the customer's naming convention, and any link with more than 1.5 dB loss is rejected. You are asked to approve the migration plan.
1. Determine whether the plan is resourced to finish in three weeks, showing the calculation and identifying the binding constraint.
Read the worked answer
Assume 15 usable weeknights in three weeks. Required rate: 1,152 links / 15 nights = 76.8 links per night. Two crews at 40 links per shift give 80 links per night, a margin of only 3.2 links per night, or about 4 percent. That margin disappears if any night is lost to a change-freeze exception, a failed link requiring OTDR retest, or field termination on the 10 percent of links. The field-terminated links are the binding constraint because they need splicing, cleaning, and retesting, not just patching. The plan is therefore not robust: it can finish only if every night is perfect. Add a third test crew for the first week, or pre-test the preterminated trunks in staging before the maintenance window, and schedule the field-terminated links early so failures have time to be corrected.
2. Recommend a test strategy that satisfies the customer and protects the schedule.
Read the worked answer
Perform 100 percent Tier 1 OLTS testing as required, using the correct reference method and a zeroed meter. Test each duplex link at both wavelengths supported by the transceivers, and record insertion loss and length where the tester provides it. Reject any link above 1.5 dB, but investigate before reworking: a high reading is often a dirty endface or a bad harness, not a bad trunk. Use Tier 2 OTDR on every failed link as required, and also on a 10 percent sample of the MPO trunks before installation. That baseline catches polarity errors, high-loss MPO connections, and macrobends before they are buried in the new pathway. Pre-inspect and clean every endface in staging, and keep known-good reference cords and spare harnesses at the test station.
3. Build a risk list for the migration and state a contingency for each of the top three risks.
Read the worked answer
Top risks: labeling mismatch, polarity errors, and change-freeze overrun. For labeling, pre-print labels to the customer's convention and have the customer approve a sample before the first window; keep a second labeler on standby. A mismatch can invalidate the whole test record even when the optics are good. For polarity, verify MPO harness type and orientation in staging, and test a sample trunk end to end before installation. Keep Type A, B, and C harnesses available because a single wrong harness can strand 24 links. For change-freeze overrun, pre-stage all trunks and patch cords, cap nightly work at 90 percent of the calculated capacity, and hold a third crew on call for the final week. Maintain a written rollback plan so any incomplete window can be restored to the old MMR before the freeze ends.