Moving a fax fleet to IP: a checklist that keeps every number

Short answer: a fax migration that loses nothing comes down to five steps: inventory every line and everywhere its number is published, pick one replacement per line (the existing machine behind a T.38 ATA, a network multifunction printer, or an online service), pilot with two or three real counterparts, cut over with the number ported, then watch a short dashboard for a month. The copper switch-off sets the deadline; this checklist lets you keep the order.

The copper switch-off removes the “wait and see” option

Across Europe and North America, carriers are retiring analogue lines exchange by exchange; only the schedule varies. The script repeats everywhere: the copper pair that carried your faxes is pulled in favour of fibre, and the machine plugged into it falls silent. The last weeks of a condemned line are rarely clean — crackling, half-transmitted pages, sessions that fail at the last page and look like a device fault when it is the line dying.

The full background is in our copper switch-off and fax overview; the operational conclusion fits in one sentence: decide the replacement while the line still works — a pilot remains possible, counterparts can be warned calmly, the inventory is taken without pressure. The day after the cut, everything happens inside an incident.

Three options, chosen line by line

The existing fax machine behind an ATA or the router

The current fleet can stay: an analogue telephone adapter (ATA) — or the carrier’s router — converts the IP line back into a signal the machine understands. The non-negotiable condition is T.38 support, at both the carrier and the adapter end, for sending and receiving alike: that is what carries a fax as relayed data instead of compressed audio. The mechanics are detailed in our explanation of fax over IP and the T.38 protocol. Order of magnitude: €40–80 per adapter, half a day of configuration per site.

The network multifunction printer

A recent MFP often sends and receives over the network with a T.38/SIP card or option. It fits where printing is already pooled: no dedicated line any more, scanning replaces paper filing. Check one thing before buying the option — whether sending is traceable, with an exportable journal; otherwise every transmission will have to be justified by hand.

The online service

An organisation that mostly sends PDFs no longer needs a machine: documents leave from the desktop, inbound faxes arrive as PDFs in a shared mailbox, and the organisation’s fax number is kept by porting — correspondents see nothing change. It is the fastest option to roll out on small sites and the easiest to monitor centrally. To arbitrate on real cost, run your inventoried volumes through the fax sending price calculator rather than through sales brochures.

Inventory before any decision

No option is chosen in a meeting; the inventory decides. For each line, record:

1. The line itself: number, site, device, owner, monthly volume sent and received. Thirty pages and a thousand pages a month do not lead to the same choice.
2. The published number: letterhead, quotes and invoices, website, email signatures, terms and conditions, contract annexes. A number that changes must be chased everywhere it is printed; a number kept by porting spares you that — the strongest argument for not creating a new one.
3. Dependencies: a line shared with other equipment, an inbound flow feeding a critical process (orders, patient files, litigation), occasional users nobody has listed.
4. The carrier’s date: the scheduled withdrawal on each site — it sets the order of the migration.

Example: of twelve inventoried lines, three numbers are heavily printed on contractual documents and one line feeds inbound orders — the three go to an online service with porting, the last to the department’s MFP, the rest behind an ATA.

The pilot: two or three real counterparts

A test between two machines you just bought proves nothing about your network. A useful pilot exchanges with real correspondents — the lab that sends its reports, the customer returning signed purchase orders, the insurer requesting documents:

• one week of live exchanges, on each option in the running if the fleet is mixed;
• a systematic comparison of the transmission reports: success rate, complete pages received, legibility on the recipient’s side;
• one written verdict per option: pass, adjust, or drop.

Two criteria settle it: the failure rate — beyond a few per cent, the option is not mature on your network — and the quality of the pages as your counterparts receive them, because they are the ones judging it.

Cut-over and the watching window

Cut over site by site, starting with the quietest one. Three rules hold the project together:

1. keep the previous line in service until the carrier confirms the number porting;
2. warn active correspondents a week ahead, stating that the number does not change where that is the case;
3. keep a fallback ready: a spare machine, a mobile service, the neighbouring site’s line.

Then comes the watching window — two to four weeks: every failure is logged with its option, its probable cause and its remedy; a repeat incident on the same option sends you back to the pilot, never to habit.

The tracking dashboard

The migration is only over once it is measured. Five indicators are enough:

A monthly half-hour review, one shared sheet, one named owner: that dashboard is what turns the migration into a closed project instead of background noise. At the first review with no incident and no line left, file the checklist with the project folder: that is the real end of the job.