Prominent Internet provider back online after week-long outage

Internet service provider, fibre operator, and web hosting company Cybersmart says it has resolved the problem that caused its network to go offline this week.

“Our engineering team successfully completed the critical hardware replacement overnight,” it said in a status update on Friday morning.

“We can confirm that all remaining problems that customers were still experiencing have been resolved, including high latency, packet loss and intermittent connectivity.”

Cybersmart customers have been affected by a national outage of the company’s networking services since around 12:40 on Monday, 12 May 2025.

Fibre customers, web hosting clients, and Cybersmart’s own websites were all offline.

One user said when Cybersmart’s website came back, its most recent network status notice initially said it would take four hours to repair the issue.

However, the estimated time to repair (ETR) soon increased to six hours.

By around 17:00, some customers reported partial connectivity. One customer told MyBroadband they could access Facebook, YouTube, and Zoom, but speed tests would not run.

Another customer said that throughput was generally intermittent and slow.

Customers continued to report intermittent connectivity throughout the week.

By 22:30 on Monday, Cybersmart had adjusted the ETR to 12 hours, suggesting that it hoped to fix the problem by midnight.

However, on Tuesday morning, the ETR had been updated again and was now undetermined.

Data from Cloudflare Radar suggested that something had gone wrong on Cybersmart’s core network.

It showed that traffic from Cybersmart’s Autonomous System (AS36874) dropped suddenly at 12:40 on Monday.

Cloudflare’s data also shows that the number of IP addresses Cybersmart announced also suddenly dropped on Monday night before recovering.

Cybersmart’s WhatsApp support line later told a customer that it had experienced an issue with one of its core switches.

It later provided an official public update stating that early indications pointed to a routing issue on some legacy hardware as the cause of the problem.

Core network problem resolved

On Friday, Cybersmart said that all critical work has been completed and connectivity services on its core network was fully restored.

“Our team will now begin the work of completing a full network diagnosis and further investigations into mitigating an unprecedented event like this in the future,” it said.

“A full communication, which will include a detailed Reason For Outage, will be sent out to all customers following this.”

Cybersmart thanked its customer for their patience and its teams for their commitment to resolving the outage.

“Please note that during this period, we have also discovered customers with physical faults on their fibre connections,” Cybersmart said.

“This proved difficult to diagnose due to the vast amount of support queries our teams have been handling, either on our Lightspeed network or the many open access networks we provide services over.”

Cybersmart said if customers were still experiencing downtime, this may be related to a physical break requiring a technician to investigate.

MyBroadband contacted Cybersmart for comment, but it did not respond by publication.

However, Daily Maverick reported that Cybersmart CTO Laurie Failkov reached out to them to explain what happened.

Fialkov blamed the outage on old Cisco 6500 routers and a sudden increase in global routing data.

According to the report, China Telecom, Hurricane Electric, and Saudi Telecom added 150,000 additional routes onto the Internet, and Cybersmart’s old routers couldn’t cope.