When the Cloud Goes Dark, Are There Solutions? How Ready Is Cyprus?

Header Image

Commissioner Marios Pieri and security officer Andreas Nicolaou explain how organisations can keep working.

The recent Microsoft 365 outage from 12 to 14 November, which hit Outlook, Exchange Online, Teams and Copilot, affected organisations in more than 70 countries. According to the Commissioner of Electronic Communications and Postal Regulation, Marios Pieri, the incident showed that dependence on a single provider is an operational risk, however large and well known the provider may be.

For Andreas Nicolaou, Information Security Officer at Cowin, the outage brought back a simple question for businesses and government services: can they keep operating when a key digital platform is unavailable? As he notes, email, communication, access to documents and collaboration between departments often rely on the same ecosystem of services. An outage can therefore affect many activities at once, without any cyberattack having taken place.

The question directly concerns Cyprus. According to Eurostat, about one in two Cypriot businesses with ten or more employees uses paid cloud services, mainly for email, office software and file storage.

How exposed each organisation is

Mr Nicolaou argues that there is no single answer. Every business and public service has different infrastructure, dependencies and needs, and not everyone faces the same risk or has the same ability to deal with its consequences. That is why, he explains, the starting point is a risk assessment. It should answer practical questions: which services are critical, which systems and suppliers they depend on, how likely an outage is, and whether current measures are sufficient.

He proposes the ISO/IEC 27005:2022 standard as a basis. He notes that the assessment should consider not only malicious acts but also technical failures, human error and third-party service outages. Then comes risk treatment, with measures proportionate to the severity of the risk, named owners and timelines. According to him, this process is in line with ENISA guidance.

In a similar vein, Commissioner Pieris recommends that businesses map the functions that depend on the cloud, including "hidden" dependencies such as user authentication or third-party applications that send emails. He proposes assessing the impact of outages lasting 4, 8, 24 and 48 hours, and setting for each function the maximum acceptable downtime and the maximum acceptable data loss.

Alternative channels and backups

Among the practical measures the commissioner describes are communication channels that do not depend on the same provider, such as SMS, mobile phones or a second messaging platform, as well as an up-to-date contact list kept locally and in print. He recommends emergency accounts protected by physical security keys, and planning so that key functions can continue offline, with documented manual procedures.

For data, he points to the 3-2-1-1 rule: three copies, on two different media, one off-site and one immutable or offline. He stresses that the provider guarantees the availability of the service, not the recovery of the customer's data. That is why email, OneDrive, SharePoint and Teams need backups made with a third-party tool. For critical functions such as email, he proposes a continuity service or an alternative provider.

Plans that work when everything stops

According to Mr Nicolaou, the Business Continuity Plan plays a central role, setting out which functions are maintained and by what alternative means. It is linked to the Incident Response Plan and the Disaster Recovery Plan, with clear roles, activation criteria and communication channels, based on the ISO 22301 and ISO/IEC 27031:2025 standards.

He highlights an often overlooked risk: a second application is not a real alternative if it depends on the same login mechanism or the same infrastructure. The emergency plan must also remain accessible when the main platform is down. "An organisation cannot repair Microsoft's infrastructure," he notes. "It can, however, have prepared the way it will work until it is restored."

Testing as a measure of readiness

"Having a plan does not prove that it works," Mr Nicolaou stresses. He proposes exercises with realistic scenarios, such as Microsoft 365 being unavailable for 24 or 48 hours. Tabletop exercises would test decisions and communication, and technical tests would check whether the alternative means actually work. The assessment needs measurements: how quickly the plan was activated, which services were maintained, whether recovery targets were met, and whether the data was checked after restoration.

Mr Pieris recommends business continuity exercises at least once a year, a review after every incident, and a response plan that is also available in print. He also proposes independent monitoring by a different provider, with SMS alerts, as the provider's status page is often updated late.

Who is responsible

According to Mr Nicolaou, responsibility lies primarily with management, which provides resources, tools, training and testing. Employees need to know the procedures for their role and which channel to use. The pressure to keep working does not justify moving company data to personal accounts or unapproved applications. No plan, he adds, eliminates all risks. Management decides which risks are reduced, which are avoided and what residual risk is accepted.

The real measure of readiness, he concludes, is concrete: when a key digital service goes down, the organisation should know which tasks continue, who takes on what, and how quickly it can recover.