This article is based on a talk at GISEC Global 2026. Operational details about Gulf Air are drawn from public statements and published interviews. The views expressed are the speaker’s own and do not necessarily represent those of Gulf Air Group.
At GISEC Global 2026, Nishanth Kumar Pathi, Director of Cyber Security & Governance at Gulf Air Group, spoke about what the Gulf’s operators of critical infrastructure can learn from seven months of sustained disruption in the region. “Nobody’s plan had a month-three page,” he said.
On 28 February, Bahrain closed its airspace, and Gulf Air could no longer fly from its home airport. Across the region, physical attacks were affecting transport, telecoms and cloud infrastructure, and cyberattacks and GPS interference added to the disruption. These crises overlapped.
For any organisation, evacuating a building changes who can reach critical systems and how. Decisions have to be made with incomplete information while the situation keeps moving. Most business continuity plans are written for disruptions lasting days, not months.
Planning for days, living through months
Bahrain’s airspace stayed closed for 40 days, but the wider disruption went on for months.
Across the industry, disaster recovery budgets, staff rotations, vendor contracts and software licences are often sized for a short outage. Yet teams still have to patch systems, keep audit evidence and meet regulatory deadlines while working from somewhere else.
Each section of a continuity plan needs one more question: if this arrangement is still in place in three months, what gives out?
“Most continuity plans are written for a bad week. We test failover, we test the call tree, and we sign off when the systems come back within hours. But a long crisis tests something different. By month three, the question isn’t whether you can recover. It’s whether the contracts, licences, budgets and people you rely on were ever meant to last that long. Every organisation should go through its plan and ask, line by line: what happens if this is still in place in ninety days?” - Nishanth Kumar Pathi
Recovery sites and the permission to use them
In March, AWS reported that strikes had damaged its facilities in the UAE and near its facilities in Bahrain. In September, AWS stated on its public health dashboard that it was unable to restore access to resources and data hosted exclusively in its Bahrain region, or in one of its three UAE availability zones. Having several facilities in one cloud region did not protect customers against a threat that affected several sites at once.
A recovery location should be chosen against the threat being planned for. A second site a few hundred kilometres away may cover a flood or a local power failure, but it may share the same exposure in a regional conflict. Operators should test whether critical services can run outside the affected region for months, with a second provider where that is practical. Where recovery lives is a business strategy decision, not only a technical one.
There is a decision to make before any switch, too: which data must stay in country, and where may the systems processing it run during an emergency? A technically ready recovery can still wait on approvals, so the permitted location, duration and controls are best agreed in writing with regulators in advance.
“A recovery site is only as good as the threat you chose it for. A second data centre down the road protects you from a flood or a power cut, but in a regional event it may face exactly the same risk as your primary. Resilience means being honest about which scenarios you’re really covering. It also means doing the slower work early: agreeing with regulators where data can go in an emergency, and for how long. The worst time to ask for permission is when your systems are ready and waiting.” - Nishanth Kumar Pathi
When the site closes completely
Many continuity exercises assume a reduced operation can continue at the original site. Bahrain’s airspace closure removed that option. As Gulf Air CEO Martin Gauss has described in published interviews, the airline moved its flights to Dammam in Saudi Arabia with the Saudi government’s permission, set up check-in at a hotel in Bahrain, arranged Saudi visas and shuttle buses across the border, and relocated crews, maintenance, spare parts and catering. He said flights operated from the first day, while the temporary operation took six days to build.
He also described the airline flying about 2,000 tons of food and essentials into Dammam, which were then trucked to Bahrain. Bahrain is an island whose only land link is the causeway to Saudi Arabia, so with the airspace closed, that road carried what the airline flew in.
After Bahrain reopened its airspace in April, Gulf Air brought services back in phases. Mr Gauss has said the airline returned flights to Bahrain one by one over about seven weeks, until the beginning of June.
Keeping a crisis team ready between waves
A prolonged crisis rarely moves in a straight line; it eases and then flares up again. A common pattern is for daily calls to become weekly ones in quieter periods, but stepping a crisis structure down too far makes the next escalation harder to manage.
One rhythm worth considering: twice-daily meetings in the first 72 hours, daily meetings while the response is active, and a short weekly call in calmer periods, with clear triggers, such as sirens, a new airspace notice or a change in a cloud provider’s status, for returning to daily meetings.
Cyber noise
Across the region, security teams faced large volumes of visible attacks during the crisis, alongside quieter intrusion and fraud attempts. Fake refund messages, evacuation notices and requests to change supplier bank details all give criminals ways to exploit confusion. Alert volume alone cannot tell a team which service is at risk. Analysts need to check actual impact while continuing to look for quieter activity in identity and telecom systems.
Unreliable GPS
GPS interference adds another problem. Public reports described spoofing affecting flights and ships across the region. The same satellite signals also provide precise time to some network, payment, and control systems. A continuity review should identify which systems depend on satellite timing, whether they can detect a bad signal, and what trusted alternative they fall back on.
People and decision rights
The hardest part of a prolonged response may be keeping people available. In any organisation, staff may be at home under alerts, caring for children or weighing whether to leave the country.
Mr Gauss has said publicly that his embassy urged departure in early March, and that he chose to stay and run the company. Front-line staff, meanwhile, are dealing with stressed and stranded passengers; Gulf Air accommodated passengers stranded in Bahrain in hotels and rebooked travellers, and looking after them was part of keeping the airline running. No small group can work an emergency rotation indefinitely.
Exercises can test this. Remove the person everyone turns to halfway through a scenario. Keep the recovery site in use for a simulated third month. Combine a cloud-region failure with an order to close the main site. The aim is to find where the team needs more people, clearer authority or a supplier in the room.
“In a long crisis, technology rarely runs out first. People do, and so does the clarity about who can decide. Teams are dealing with alerts at home, families to look after and uncertainty about what comes next, while still expected to keep critical services running. That is why decision rights matter so much. Who can approve emergency access, who can authorise an exception, who speaks to the regulator: these need to be agreed and written down before an incident. The organisations that cope best aren’t the ones with the most tools. They’re the ones that know who decides.” - Nishanth Kumar Pathi
Physical safety and cyber decisions belong with one crisis team. Before an incident, agree who can authorise emergency access, approve temporary, logged exceptions to security controls, approve a move abroad, and speak to the regulator. Left undecided, those calls fall to exhausted people, for months, and then again in July and September.





