Exam season shutdowns return across the Middle East and North Africa

Cloudflare’s visibility into global Internet resilience comes from operating in more than 300 cities across over 100 countries, interconnecting with more than 12,000 network providers. That breadth lets us observe how disruptions — from government action to severe weather — affect connectivity around the world. The second quarter of 2023 was especially notable for government-directed shutdowns, many tied to secondary school exam seasons in the Middle East and North Africa.

\Q2 2023 Internet disruption summary

Iraq

Iraq’s Ministry of Communications initially said it had refused the Ministry of Education’s request to impose shutdowns during exams, but that stance reversed within two weeks. Two parallel sets of shutdowns ran during the quarter. The first applied nationwide except for the Kurdistan Region, targeting 9th and 12th grade exams from June 1 through July 15. Those shutdowns occurred between 04:00 and 08:00 local time (01:00–05:00 UTC) and hit a range of providers, including AS203214 (HulumTele), AS59588 (Zain), AS199739 (Earthlink), AS203735 (Net Tech), AS51684 (Asiacell), and AS58322 (Halasat). Traffic on each provider dropped to zero during the shutdown windows, which took place on 13 separate days in June.

The Kurdistan Region saw its own, shorter shutdowns — generally around two hours, between 06:30 and 08:30 local time (03:30–05:30 UTC) — on June 3, 6, 10, 13, 17, and 24. These impacted AS21277 (Newroz Telecom), AS48492 (IQ Online), and AS59625 (KorekTel).

Algeria

Algeria continued a pattern from 2022, using content blocking rather than a total shutdown to prevent exam cheating. Instead of full traffic loss, Cloudflare observed two nominal dips on exam days across mobile providers AS33779 (Ooredoo/Wataniya), AS327931 (Djezzy/Optimum), and AS327712 (Mobilis/Telecom Algeria). The first disruption occurred between 08:00 and 12:00 local time (07:00–11:00 UTC), with a second between 14:00 and 17:00 local time (13:00–16:00 UTC).

Syria and Senegal

After four exam-related shutdowns in 2022, Syria implemented just two this year, on June 25 and 26, between 05:00 and 08:30 local time (02:00–05:30 UTC). Syrian Telecom (AS29256) notified subscribers via Facebook that the Ministry of Education had requested the cut.

Senegal’s shutdowns were protest-related rather than exam-related. After violent demonstrations over the sentencing of opposition leader Ousmane Sonko, the government restricted access to platforms including WhatsApp, Facebook, Twitter, Instagram, TikTok, Telegram, Signal, and YouTube. A June 4 statement from the Ministry of Communication suspended mobile Internet access, with the suspension lifted on June 6. Sudatel Senegal (AS37196) saw disruptions from 15:00 local time on June 3 through 01:00 on June 5, then again from 13:00 on June 5 until 01:00 on June 6. Tigo/Free (AS37649) experienced three shutdowns: 15:30–19:00 on June 3, 13:45 on June 4 to 02:05 on June 5, and 13:05 June 5 to 01:00 June 6. Senegal is UTC+0, so local times match UTC.

Post-arrest shutdowns in Pakistan

Following the May 9 arrest of former Prime Minister Imran Khan, protests led Pakistan’s government to order mobile Internet shutdowns and block several social media platforms. The impact varied by provider: traffic fell completely starting at 22:00 local time (17:00 UTC) on May 9 for Telenor Pakistan (AS24499) and China Mobile Pak (AS59257), at 18:00 local time (13:00 UTC) for Mobilink/Jazz (AS45669), and at 01:00 on May 10 (20:00 UTC May 9) for Ufone/PTML (AS56167). Services were restored at 22:00 local time (17:00 UTC) on May 12.

During the outage, median latency within Pakistan dropped from the 90–100ms range to about 75ms, likely because users shifted to fixed broadband connections, which saw increased traffic volumes while mobile networks were down.

India and Mauritania

In India, shutdowns remain a recurring issue, with digital rights group Access Now reporting at least 84 in 2022. Manipur saw an extended shutdown starting May 3, after ethnic conflict escalated. Mobile data was suspended for five days at a time, with orders extended repeatedly. Traffic from Airtel (AS45609) and BSNL (AS9829) fell significantly around 18:00 local time (12:30 UTC) on May 4. Airtel traffic stayed low and continued dropping through June; BSNL showed slight recovery in early June but remained extremely low. The order was still in effect as of late July.

Mauritania’s shutdown was shorter but severe: mobile Internet was cut at 23:00 local time on May 30 after protests over a death in police custody. Connectivity returned at 23:00 on June 6. Chinguitel (AS37541) and Mattel (AS37508) both showed near-complete traffic loss for that period.

Weather and Infrastructure Failures

Several weather-related and physical infrastructure incidents across the globe caused measurable Internet degradation throughout Q2. From Pacific typhoons to data center power failures, these events demonstrate how dependent modern connectivity is on physical systems.

Typhoon Mawar Hits Guam

"Super Typhoon" Mawar made landfall on the US territory of Guam on May 24, causing extensive physical damage to trees, buildings, power lines, and communications infrastructure. The storm produced an immediate and significant drop in Internet traffic at the country level. Restoration began promptly, with Guam Power Authority, Docomo Pacific, and GTA Teleguam providing regular updates.

For GTA Teleguam (AS9246), recovery was achieved relatively quickly, with traffic returning to pre-storm levels around June 17. In a June 20 post, the provider noted that a majority of its wireless network cell sites were operational. Docomo Pacific (AS3605) faced a longer road to recovery, with traffic volumes remaining well below normal through the end of June.

Fiber Cuts in Bolivia and The Gambia

A fiber optic cable cut in the town of Pongo, Bolivia, announced by telecommunications company COTAS on June 19, disrupted connectivity for three networks — AS25620 (COTAS), AS27839 (Comteco), and AS52495 (Cotel) — for a roughly five-hour window between 13:00 and 18:00 local time (17:00–22:00 UTC).

In The Gambia, state provider Gamtel disclosed a localized fiber cut on June 7 via Twitter, followed by additional cuts on June 8. These incidents led to a traffic reduction of up to 80% for AS25250 (Gamtel) between 14:00 local time on June 7 and 00:00 on June 9. Given The Gambia’s UTC+0 offset, the local and UTC times align.

Submarine Cable Issue in the Philippines

A June 5 advisory from PLDT described a loss of capacity on a partner's submarine cable, which degraded Internet speeds for its customers. The traffic impact was a decrease in throughput beginning around 14:00 local time. Simultaneously, latency ticked upward and bandwidth declined, confirming user complaints of slower browsing. Reduced impact began after the cable partner provisioned alternate capacity, a process completed by late June 5 UTC, when normal service resumed.

Power Outages and Connectivity

Curaçao-wide Power Failure

A malfunction in a main power cable connected to the Parera substation prompted a series of alerts from Aqualectra on June 8. The utility said the outage impacted "all neighborhoods," and traffic across Curaçao fell substantially. The effect was visible across multiple providers, including AS11081 (UTS), AS52233 (Columbus Communications), and AS27660 (Curaçao Telecom). Power was restored to all areas just after midnight local time on June 9.

Lisbon Data Center Incident

A power outage at an Equinix data center in Prior Velho, Portugal, on June 6 affected utilities, banking services, and court systems, according to published reports. The blackout also took a toll on ICT, with Internet service provider MEO suffering traffic declines for both AS3243 (MEO-RESIDENCIAL) and AS15525 (MEO-EMPRESAS). Users likely felt this as reduced bandwidth and higher latency, both visible in network quality metrics.

Botswana and Barbados

A countrywide Botswana power outage on May 19 lasted from 10:45 to 12:15 local time (08:45–10:15 UTC), representing about a 90-minute Internet disruption. The Botswana Power Corporation publicly informed residents via tweet but did not cite a root cause for the outage.

In Barbados, The Barbados Light & Power Company acknowledged a widespread power failure via an "Outage Notice" tweeted April 4. The national traffic graph dipped notably prior to that post. Although power was restored by 18:00 local time, traffic required several additional hours to make a complete recovery.

Technical Faults and Network Hiccups

Namibia and the Solomon Islands

Telecom Namibia cited undefined "technical challenges" on fixed and mobile data services on June 15 and 16, which produced intermittent connectivity and lower traffic for both the provider and the country. The situation lasted roughly seven hours.

Likewise, in the Solomon Islands, a Facebook post from Our Telekom on April 26 said mobile data was down for unspecified technical reasons. Traffic for AS45891 (Our Telekom/SBT) experienced a decline between 19:30 UTC April 26 and 06:00 UTC April 27. Country-wide traffic likewise dipped during that interval.

SpaceX Starlink's global user base can feel the repercussions of comparatively small technical issues. On April 7, traffic from AS14593 (SpaceX-Starlink) dropped sharply just before midnight UTC. A post from Elon Musk stated the issue was "Caused by expired ground station cert," tagging an expired digital certificate as the culprit for disrupted satellite-to-ground communications. Service normalized after just two hours.

Backbone and OSI Layer Faults

In Madagascar, a Telma-posted Twitter alert on April 7 cited a "problem with the backbone," leading to a loss of up to two-thirds of Internet traffic between 06:15 and 11:00 UTC for AS37054 (Telma Madagascar) and the country as a whole.

Virgin Media Service Slip-Ups

Virgin Media in the UK saw its broadband customers hit with repeat outages on April 4. The first wave arrived around midnight UTC and lasted through 08:00 UTC. A second round began about 14 hours later, with jittery traffic consolidation until approximately 20:30 UTC.

The company used Twitter for rolling updates. The first acknowledgement came hours into the incident; the second acknowledgment was faster. Root-cause analysis continued, with engineering activity seeming to involve re-seating fibre cards and modifying time-to-live values, per posts from a self-identified customer. Cloudflare published a precise technical review of this particular incident in an April 4 blog entry.

Attack Activity: From Crimea to Russian Satellites

On May 26, an outage struck Miranda Media (AS201776), the Crimean-based, Russian-controlled network operator. Traffic began declining around 16:30 local time (13:30 UTC) and hit zero by 18:15 local time (15:15 UTC), cutting connectivity across the Crimean Peninsula and parts of occupied Ukraine. Service did not recover until roughly 06:00 local time on May 27 (03:00 UTC). Published reports attribute the incident to a cyberattack by Ukrainian hacktivists—another offline front in the ongoing conflict that has repeatedly reshaped routing and uptime in the region since February 2022.

Russia itself also saw network-level fallout from an attack. Dozor Teleport, a satellite provider serving the Ministry of Defense, the Northern Fleet, Gazprom, the FSB, Rosatom, and the Bilibino nuclear power plant, went down for about 16 hours on June 29. The outage ran from 01:30 to 17:30 UTC, and at least two groups claimed responsibility for the cyberattack that reportedly caused it.

When Conflict Takes Down the Wire

On April 23-24, users in Chad faced repeated connectivity failures spanning multiple providers. Outages were visible from 04:30 to 06:00 and 15:00 to 20:00 local time on April 23, then again from 04:00 to 08:30 local time on April 24. The affected networks included AS327802 (Millicom Chad), AS327756 (Airtel Chad), AS328594 (Sudat Chad), and AS327975 (ILNET-TELECOM). The root cause was reportedly damaged fiber infrastructure linking Chad to Cameroon and Sudan—where fighting between the Sudanese Armed Forces (SAF) and Rapid Support Forces (RSF) was simultaneously wreaking havoc on Internet service.

Sudan’s own providers bore the brunt of that conflict. Fuel shortages and power cuts destabilized connectivity from mid-April onward:

  • AS15706 (Sudatel): A complete outage from 03:00 April 23 to 17:00 April 24 local time (01:00 April 23 to 15:00 April 24 UTC), followed by another from 03:00 April 25 to 01:00 April 28 local time (01:00 April 25 to 23:00 April 27 UTC).
  • AS36972 (MTN): Disruption from 03:00 to 12:00 local time on April 16 (01:00-10:00 UTC), another from 20:00 April 27 to 02:00 April 29 (18:00 April 27 to 00:00 April 29 UTC), and then a prolonged near-total outage beginning May 5.
  • AS33788 (Kanar Telecommunication): After a traffic dip on April 19, a near-complete outage from 12:00 April 21 to 01:00 April 29 local time (10:00 April 21 to 23:00 April 28 UTC), followed by a multi-week outage starting around 00:00 local time on May 11 (22:00 May 10 UTC).

Planned Maintenance, Unplanned Downtime

Sometimes the cause is not an attack but a repair. On April 6, maintenance on the West Africa Cable System (WACS) submarine cable disrupted traffic across Togo, the Republic of Congo (Brazzaville), and Burkina Faso. Operator GVA-Canalbox (AS36924) warned of “very strong disruptions on Internet connections with the risk of a total outage.” The observed impact confirmed it:

  • In the Republic of Congo (Brazzaville), a significant drop in traffic occurred between 15:15 and 22:15 UTC.
  • In Burkina Faso, the disruption was briefer and less severe, from 09:15 to 18:00 UTC.
  • In Togo, traffic was interrupted from 11:00 to 23:15 UTC.

The Sticker Price of an Offline Nation

Whether the disruption stems from a cut cable, severe weather, or armed conflict, the economic cost is inevitable. Operators responding to accidental damage typically work to limit the blast radius and restore service quickly. Government-mandated shutdowns, however, inflict deliberate economic self-harm. The Internet Society’s Net Loss Calculator aims to quantify that damage, projecting the cost of an outage in terms of gross domestic product (GDP), foreign direct investment (FDI), and unemployment—useful data for advocates and policymakers weighing the true price of a shutdown.

Ongoing tracking and analysis of such disruptions are available via Cloudflare Radar.