The question everyone asks
Whenever we present an attendance system, the first question is nearly always the same: what happens when the connection drops? It is the right question. In Kinshasa as in the provinces, a building's internet link can go down one morning without warning, the provider can have an outage, and the power can cut out just as all the staff are arriving.
If the time clock needs the internet to record a clock-in, what follows is predictable. Staff wait in front of a device that does not respond, then give up. Someone gets out a notebook and writes names down by hand. At the end of the month, HR rebuilds the hours, with all the mistakes and disputes that brings.
What a connection-dependent system costs you
- Lost clock-ins. Unrecorded arrivals turn into absences on paper, although the staff were there.
- A parallel register. The backup notebook has to be keyed in again, by someone, with errors.
- Disputes. Honest staff see their pay docked; the dishonest have a ready-made excuse: "it was the outage".
- Eroding trust. After a few incidents, nobody believes the system's figures any more, and the paper register creeps back.
Three principles for a system that holds up
Record locally first
A clock-in should be recorded where it happens, on the terminal or on a computer in the building, without waiting for a reply from a remote server. The terminals we install keep clock-ins in local memory throughout an outage.
Resynchronise without anyone stepping in
When the network returns, clock-ins should flow back to the central platform on their own. Nobody should have to walk round with a USB stick, or export a file and email it. And the time that counts must be the moment the person clocked in, not the moment the data arrived.
Flag what is not coming through
The quietest danger is not the one-hour outage. It is the terminal that stopped responding three weeks ago without anyone noticing. You find out when payroll closes and hours are missing, and by then it is too late to rebuild them. A good system watches its devices and raises an alert as soon as one goes silent.
What about electricity?
Working offline does not protect you from a power cut. For that, you need an uninterruptible power supply at each entry point where the mains is unreliable, and generator backup for central equipment. It is a budget line to plan from the outset, not an extra bolted on after the first incident.
Two set-ups, depending on your situation
Everything on the internal network
For a single site, or an organisation with no reliable link at all, the most robust option is not to depend on the internet in the first place. That is the approach behind OpSolver, our desktop software for fingerprint attendance: installed on a computer inside the organisation, it runs on the internal network, reads and stores fingerprints locally, and produces attendance sheets that can be exported to payroll. The data never leaves your building.
A central platform with self-sufficient terminals
As soon as there are several sites, management needs a consolidated view. The platform is then central, but each terminal stays self-sufficient: it records on the spot and sends its clock-ins back when the link allows. An outage delays the data; it does not lose a single arrival.
We apply the same principle to our university and school platforms: work carries on during the outage, and synchronisation resumes by itself when the link comes back.
Questions to put to your supplier
- What happens if the connection is down for two hours? For three days?
- Where are clock-ins stored during an outage, and how many can the terminal hold?
- Is resynchronisation automatic, or does someone have to do something?
- How will I know when a terminal has stopped sending anything?
- What happens in a power cut?
If the answers are vague, the system was designed for a different context from yours.
An outage should delay the data, never make it disappear.
Want to check that your set-up can take the strain, or design one that can? Tell us about your situation.