In short: Dover queues under the EU Entry/Exit System have less to do with the number of booths than with whether every kiosk, tablet and lane sensor stays connected to the systems behind it; the Port of Dover built an 84-kiosk, £40 million facility that still could not be used for car passengers in July 2026, because the French border systems behind the kiosks were not switched on.
Key Takeaways
- The hall was ready, the link was not — Dover's Western Docks EES facility has 84 biometric kiosks and cost the port £40 million, yet it sat unused for car passengers through summer 2026 because the French systems behind it had not been activated.
- Queues move to wherever the network reaches — When fixed kiosks are unavailable, enrolment falls back to officers with handheld tablets in the traffic lanes, and those devices need reliable coverage across open, metal-crowded marshalling yards.
- Ten thousand cars strain the public network — On peak days Dover expects 10,000 or more outbound cars, and thousands of stationary phones in a few kilometres of queue can swamp the same public cells that port staff depend on.
In a nutshell
Why are there queues at Dover?
The queues at Dover come from three things stacked on top of one another: (i) a fixed number of French border control positions at the port, operating under the juxtaposed controls that put French officers on English soil, (ii) a summer peak in which outbound tourist traffic arrives in the few days after the schools break up, and (iii) since October 2025, a new EU border process that takes longer per passenger than the old passport stamp. Any one of these can be managed; together, and with a few kilometres of the A20 as the only buffer, they produce the tailbacks that make the national news most summers.
The May 2026 half-term showed how quickly it happens. Around 8,500 outbound tourist vehicles arrived at the port, a critical incident was declared, and waiting times reached four-and-a-half hours. The port expected about 7,500 cars on a Friday in mid-July, nearly 10,000 the following day, and around 10,500 on each of 24 and 25 July. Those figures come from the Port of Dover's own planning, reported by The Independent in July, and they are the volumes any border process at Dover has to be designed around.
What is the Entry/Exit System, and why does it slow the border?
The Entry/Exit System (EES) is the EU's digital register of non-EU travellers crossing the Schengen external border. It replaces the passport stamp with a biometric record: a facial image for every traveller, including babies, and fingerprints for everyone aged 12 and over. British citizens are in scope. The EU began the rollout on 12 October 2025, phased over six months to April 2026, and a first registration remains valid for three years, so the second crossing within that window is quicker than the first.
At most UK ferry ports the process happens on the far side of the Channel or the North Sea. At three English sites, however, French border checks take place before departure: the Port of Dover, Eurotunnel's Folkestone terminal, and Eurostar at St Pancras. Those three shared £10.5 million of UK Government funding to prepare, and in July 2026 the Department for Transport announced a further £20 million. For a car passenger at Dover, a first EES registration means each occupant presenting a passport, a face and, for most adults, four fingers to a reader, either at a kiosk or to an officer with a handheld device. A process that once took a glance at a passport photo now involves a data capture, a lookup against a central EU system, and a record written back.
Why did a £40 million kiosk hall sit empty?
The Port of Dover's answer to EES was a new processing facility at the Western Docks, separate from the Eastern Docks ferry terminal, with 84 kiosks to record biometric details before cars reach the French border booths. The port paid for it from its own capital, around £40 million, and built it to the EES timeline. In July 2026 it was still unused for car passengers, because the French authorities had not switched the kiosks on and no date had been set. Doug Bannister, the port's chief executive, put it plainly: "it's been specifically designed to handle our peak volumes safely and efficiently, and we can't use it."
The reason given in February 2026, when the BBC reported the delay, is instructive. The French side was having problems communicating with the kiosks that scan faces, fingerprints and passports; the French Interior Ministry said the registration tools were "undergoing technical work that is being finalised". Toby Howe of the Kent and Medway Resilience Forum said the port and Eurotunnel were "ready to go". The hardware was installed and the building finished; the integration between the kiosk estate and the national border systems was the missing piece.
We do not know the detail of that integration, and we would be wary of anyone who claims to from outside. Its shape is familiar to anyone who has run a distributed system, though. A biometric kiosk is an edge device; it is only useful when it can authenticate to, and exchange records with, a central service that someone else owns, over a path that has to be secure, monitored and resilient. When that path is not working, the kiosks cannot register anyone.
What happens to Dover queues when the kiosks are not available?
When the fixed kiosks cannot be used, registration falls back to people: officers walking the lanes with tablets, or passengers called forward to booths, or, at the busiest moments, the "relaxations" the French ministry referred to, under which Schengen states may scale back the full biometric check to keep traffic moving. Alastair Gunn of the Road Haulage Association expected exactly that, saying checks would "inevitably" have to be suspended at busy periods, and that "the infrastructure and resource needed for EES to work at all ports is not fully in place".
In the fallback, connectivity becomes a daily operational question. A tablet enrolling a family of four in the third lane of a marshalling area needs a reliable data connection to the same central systems as the kiosk, from a position surrounded by steel car bodies, coaches, lighting columns and canopies, often within a few hundred metres of the sea. Wi-Fi access points mounted on canopies can cover a booth plaza well; they struggle across an open yard where the line of sight changes every time a coach pulls in, and they hand over poorly as an officer walks the length of a queue. Public mobile coverage is the other option, and on a peak day it is the least reliable one, for a reason covered in the next section.
The same applies to everything else a port uses to manage the queue: ANPR cameras counting vehicles into the stacking lanes, variable message signs, lane-control barriers, CCTV for the control room, and the radios or push-to-talk handsets that marshals use to call lanes forward. Each one is a small, unglamorous data link, and each one becomes important on exactly the days when the port is busiest.
Why does mobile signal disappear in a long queue?
A public mobile cell is sized for the people normally within it. The cells covering a stretch of the A20, a dual carriageway approach or a ferry port's marshalling area are built for a few hundred people passing through at any one time. A summer Saturday at Dover puts several thousand stationary vehicles, and three or four phones in each, into the same few square kilometres for hours. Those phones stream video to keep children occupied, refresh travel updates and make calls about missed sailings. The cells were never planned for that load, and throughput per user falls sharply once the cell is saturated.
Port and border staff using SIM-based tablets on the same public network share that congestion. Their devices compete for the same radio resources as the passengers in the queue they are trying to clear, which means the moment of greatest demand on the border process is also the moment when its fallback connectivity is weakest. Priority schemes on the public networks exist for the emergency services, but a ferry operator's lane marshal or a contractor's kiosk technician is unlikely to qualify.
What would a resilient border-processing network at a port look like?
We would start from the assumption that every device involved in moving traffic through the border should have a connection that does not depend on how many holidaymakers are in the queue. In practice that means a private network covering the processing area, the stacking lanes and the approaches the port controls, with three characteristics.
Firstly, dedicated capacity. A private 5G network on spectrum the port controls, typically a shared-access licence from Ofcom in the 3.8 to 4.2 GHz band, carries only the port's devices, so a tablet enrolling a passenger is not competing with thousands of phones watching video. Capacity can be planned for the peak rather than hoped for.
Secondly, coverage designed for an open yard. Cellular radio handles mobility across a large outdoor area better than Wi-Fi, with a small number of masts covering lanes that would need dozens of access points, and devices handing over cleanly as officers walk the queue. Kiosks inside a hall can stay on wired connections; the network's job is everything outside the walls.
Thirdly, separation and monitoring. Border systems belong to a foreign state's police service, and port systems belong to the port; a private network can carry separate, isolated slices for each, with the port able to see in real time whether every kiosk, tablet and camera is connected, rather than discovering a failure when the queue starts to grow.
There are trade-offs. A private network is a capital project in its own right, it needs operating, and it does nothing to solve the integration between kiosks and national systems, which is the problem that held up Dover's hall. It also cannot add border officers or booths. What it can do is remove connectivity from the list of things that fail on the busiest day of the year, and give the port measured evidence when it discusses capacity with the French and UK governments.
Which UK ports does this apply to, and what should they do now?
EES checks happen before departure only at the three juxtaposed sites, so Dover, Eurotunnel at Folkestone and St Pancras carry the queueing risk on English roads. Routes from Portsmouth, Newhaven, Poole, Plymouth, Harwich and Hull put the EES check on the continental side, and Irish routes from Holyhead, Fishguard, Pembroke Dock, Liverpool and Cairnryan are outside it altogether, since Ireland is not in the Schengen area. Ports on those routes still deal with peak-day congestion in marshalling yards, along with the same pattern of public-network saturation, and any future change to UK or Irish border processes will raise the same questions there.
The EU's travel authorisation scheme, ETIAS, is due to follow EES and will add another data check for many travellers. For a port planning the next phase, we would suggest three practical steps before the 2027 Easter peak: (i) survey public mobile performance across the marshalling and stacking areas on a peak day, so the decision is based on measured throughput rather than a coverage map; (ii) list every device that touches traffic flow, from kiosks and tablets to ANPR and signage, and record what each one depends on for connectivity; and (iii) price a private network for the processing and stacking areas against the cost of a day like the May 2026 half-term, when 8,500 vehicles waited up to four-and-a-half hours.