Cross Connect and Meet Me Room Software: When the Label on the Panel Stops Matching the Record

Why the meet me room record drifts away from the meet me room

A technician is standing in front of frame 14 at 11pm with a work order that says pull the connect on panel B, port 22. The label on port 22 reads a customer name that was acquired two years ago. The spreadsheet says port 22 is free. The cable in port 22 is warm, patched, and carrying somebody's production traffic. He has two choices: pull it and find out, or walk back to the office and spend an hour trying to reconstruct history from email. On a good night he walks back. On a bad night the operator finds out about the error from a customer whose network just dropped, and the incident review turns into a conversation about how the facility does not know its own cabling.

Every meet me room drifts. The mechanism is always the same. Connects get installed under time pressure and the record is updated later, or by someone else, or not at all. Disconnects are the worse half, because there is no customer waiting on a disconnect so it never gets the same attention. A tenant is acquired and the name on 40 labels is now wrong. A frame gets re-terminated during a riser upgrade and the port numbering shifts by one for half a panel. None of these are negligence. They are the accumulated result of a physical estate being maintained faster than a document about it.

The cost of the drift shows up in four places. Outages caused by pulling live circuits, which are the ones that reach the executive team. Unbillable or unbilled connects, because a port in use with no order behind it cannot be invoiced with confidence. Install delays, because an order for a connect between two tenants stalls while somebody physically traces what is actually on the panel. And the slow one nobody counts: your capacity picture is wrong, so you either install panels you did not need or run out of ports in a riser during a customer install.

Problem one: the record is a list of ports, not a path

A cross connect is not one row. It is a path with two ends and usually several intermediate points: customer cabinet, cage panel, riser, meet me room frame A, a jumper, meet me room frame B, riser, carrier cage. Fibre adds strand pairs, media type and connector type at each termination. A spreadsheet stores the two endpoints and hopes the middle is obvious. It is not obvious at 11pm.

NetBox deserves credit here because it does model cables, terminations and path tracing properly, and for many operators it is the correct starting point rather than a compromise. Where it stops is that it is a source of truth, not a workflow system: it does not take orders, verify authority, dispatch a technician, capture evidence or hold the order lifecycle. FNT Command and Sunbird dcTrack do more of the workflow but expect the facility to be described inside their hierarchy and naming conventions, and a carrier hotel that has been operating for twenty years has naming conventions that predate the product and cannot be changed without relabelling the building.

A custom build models the path explicitly, with every termination as a first class object carrying its own label text as it physically appears, not as it should appear. That distinction matters more than it sounds. When the label in the room disagrees with the record, you want the system to hold both and flag the conflict, because a system that silently stores the correct value gives a technician standing in front of the wrong label no help at all.

Problem two: the letter of authority arrives as a PDF and nobody checks it properly

A connect between two tenants requires the party being connected to have authorised it. That authority arrives as a letter of authority or a connecting facility assignment, usually a PDF attached to an email, sometimes on letterhead from a company that has since been renamed. Someone in operations reads it, decides it looks right, and the order proceeds. There is no structured record of which document authorised which port on which date, so when the connected party later asks who approved access to their panel, the answer is a search through an email archive.

The build makes authority a required, structured link on the order: which document, which requesting party, which authorising party, which specific ports it covers and what date range it is valid for. The order cannot move to dispatch without it. Document extraction earns its place here as a narrow job rather than a headline feature: read the inbound PDF, pull out the parties, the site, the requested demarcation and any port references, and present them as a draft order for a human to confirm. That gets your operations coordinator out of retyping and into checking, which is the part that actually requires judgement.

Problem three: the technician has no way to be sure before they cut

Most live circuit incidents are not carelessness. They are a technician with an ambiguous instruction and no cheap way to verify. The fix is procedural and the software has to carry the procedure rather than assume it.

What works in practice: every termination carries a scannable label, so the work order is confirmed by scanning the port rather than reading a number off paper. The mobile view shows the full path and the customer at both ends before the action is allowed. Disconnects require an explicit confirmation state and, for anything a customer has flagged as production critical, a second person or a fibre trace check. Every completed action captures a photo of the port as left. The photo is not bureaucracy, it is the thing that resolves the argument three months later about whether the connect was ever installed.

The other half is the audit. Nobody audits a meet me room because it is a two week walk with a clipboard and the results are stale before they are typed up. Give the audit a mobile tool that walks the technician panel by panel, scan by scan, records found versus expected inline, and lets the walk happen in fragments over weeks without losing state. Facilities that adopt this typically find their first pass exceptions cluster in exactly two places: ports marked free that carry live cable, and connects with no order behind them at all. Both are recoverable value.

Problem four: riser and panel capacity gets discovered during a customer install

Ports are only half the capacity question. The other half is strands in the riser between the customer cage and the meet me room, and the available positions on the frame itself. When the record is a list of connects rather than a model of the physical plant, exhaustion is discovered by a technician who cannot complete an install, which means the customer finds out at the same time you do.

A modelled build knows how many positions each panel has, how many strands each riser carries, how many are lit, how many are reserved by pending orders, and what the lead time is to install more. Pending orders matter as much as installed connects, because an order accepted today consumes capacity tomorrow. Once that exists, you can put a simple constraint check into order acceptance and stop selling paths you cannot deliver without a construction job.

What this costs and how long it takes

A first release covering the path and termination model, connect order intake with structured authority, and the mobile technician workflow with scan confirmation and evidence capture runs $45,000 to $110,000 and ships in 10 to 14 weeks. A full platform adding the audit tool, riser and panel capacity, customer self service ordering, disconnect lifecycle with reclaim, and a clean handoff to billing runs $120,000 to $300,000 over 5 to 9 months.

What drives cost in this specific category: the state of your labelling. If the room does not follow a consistent scheme, and many older carrier hotels do not, then somebody has to decide whether to relabel or to model the mess as it stands, and modelling the mess is usually the right answer but it is more engineering. Multiple buildings with different conventions multiply that. Whether you already run NetBox or a DCIM matters, because importing an existing cable model is far cheaper than reconstructing one. Customer facing ordering adds authentication, entitlement rules and a support surface. And migration is real work: getting the current spreadsheet into a modelled path structure means resolving every ambiguity in it, which is unglamorous and takes weeks.

Build versus buy, and when buying is right

Buy, or rather adopt, if you have a few hundred connects in a single room with consistent labelling and one operations coordinator who handles every order. NetBox will hold your cable model properly, costs nothing in licence, and paired with a disciplined order process in whatever ticket system you already run, it will serve you for years. dcTrack or FNT Command make sense if you want the DCIM and the connection record in one product and your facility fits their hierarchy comfortably.

Build when the connect volume is high enough that order intake is a full time job, when you operate more than one building with incompatible conventions, when you have had a live circuit incident and the review pointed at record quality, or when you cannot answer with confidence how many connects are installed but not billed. The strongest signal is organisational rather than technical: if there is one technician whose knowledge of the room is load bearing, and everyone knows who that is, you are one resignation away from a very expensive audit.

How to choose a developer for meet me room software

Ask them to model a fibre cross connect on a whiteboard, from customer cabinet to carrier cage. If they draw two endpoints and a line, they are thinking about a database table. If they draw terminations, strand pairs, intermediate frames and ask which panel positions are spliced versus patched, they have done this before.

Ask specifically how the system behaves when the physical label disagrees with the record, because the honest answer is that it will, regularly, and the design has to hold both values. Ask how a disconnect is protected. Ask whether the mobile workflow functions in a room with no signal, since meet me rooms are frequently the worst coverage in the building and an offline first design is not optional.

Settle ownership before kickoff: repository, cloud accounts and the freedom to bring in another firm. At Digital Heroes the client owns the code from the first commit. A useful first step before you scope anything, take one frame, walk it with a clipboard, and compare what you find to what the spreadsheet says. The exception rate on a single frame is a fair estimate of the exception rate across the room, and it will tell you quickly whether this is a process problem or a systems problem.