South Africans recently experienced sluggish internet speeds or breaks in connectivity as a result of multiple undersea cable breaks. While this was frustrating when people were trying to access their company networks, it became even more challenging for anyone using Microsoft’s cloud services.
By Richard Firth, CEO of MIP Holdings
Microsoft services were some of the hardest hit by the cable problems, causing many companies to lose access to e-mail systems and other cloud-based services such as Teams for several hours. Many company websites that rely on cloud platforms were also inaccessible for most of Thursday.
Initial reports indicated that the cable breaks were the result of a fishing trawler dragging its anchor across the cables in shallow water, but further investigation found that the break was caused at a depth of three kilometres due to seismic activity causing undersea rock falls. Four cables – Wacs, Ace, MainOne and Sat-3 – were affected, with repairs estimated to take around five weeks.
Considering that a total of nine cables carry all of South Africa’s international bandwidth, four of them breaking is more than a minor inconvenience. The fact that it took four days for companies to gain an understanding of the cause and an ETA on repairs, which will take over a month, is a massive challenge for cloud providers like Microsoft and users alike.
A new set of challenges
This is not the first time that South Africa has experienced loss of connectivity as a result of cable breaks. Wacs and Sat-3 went down in 2023 as a result of a ship’s anchor, and again internet users had to wait for days to find out what was going on, only to be given a six- to eight-week timeline for repairs.
South Africa experienced similar disruption in 2020 when the Wacs cable went down for weeks, following a simultaneous dual cable break of Wacs and Sat-3 a few months earlier. Seacom’s cable is also currently experiencing an outage in the Red Sea, with repairs being delayed as a result of the geopolitical tensions in the area.
A great deal of international traffic has been rerouted through cables such as Google’s recently completed Equiano line, but this is not a permanent solution – nor is it one that will stop the same thing from happening again. Sat-3 is one of the oldest undersea cables, but it still sees heavy traffic, and 39% of all of South Africa’s international traffic goes through Wacs.
Many companies have moved their business applications to the cloud in order to ensure redundancy and stability, but this latest outage highlights the fact that they have effectively replaced one set of challenges with another. This was highlighted as far back as 2021, when the fire that crippled OVH’s data centre in Strasbourg, France, caused more than $120-million in damages, affected more than 65 000 customers, and pulled down around 3,6-million websites worldwide.
Eyes wide open
In a report issued at the beginning of the year, the International Data Corporation (IDC) indicated that spend on cloud services in South Africa is expected to increase this year. The analysts found that only 9% of respondents were spending on traditional on-premises IT infrastructure, while 60% indicated a hybrid cloud approach.
A recent McKinsey survey of technology leaders at more than 50 major African businesses found that these companies have, on average, about 45% of their workloads in public cloud today. Although based on a relatively small sample size, that’s on par or ahead of the rates of adoption in North America and China.
While there are some compelling benefits of moving to the cloud, the recent connectivity challenges are just the latest headaches felt by CIOs, who are also grappling with issues such as major vendors going all-cloud, cost spikes driving repatriation from the cloud, and other unexpected business risks.
Since cloud services are Dollar-based, cloud can become very expensive for local companies, causing a number of organisations to start moving applications back on-premise. Despite the fact that the three main global cloud providers – AWS, Microsoft and Google Cloud – have got local cloud data centres, which are supposed to alleviate concerns like latency and data residency requirements, an outage like the one we saw last week will add to the repatriation drive.
Unfortunately, many organisations – particularly those without dedicated CIOs – lack the technical knowledge to minimise business risks around cloud. For example, some companies have not considered business continuity in the event of a loss of cloud services, ransomware attacks on the cloud service, or believe they don’t need backups if they are in the cloud. Security remains a concern, with most breaches being caused by user error and malicious insiders.
This latest outage just highlights the fact that companies need to start evaluating the high stakes risks they are taking by using global cloud infrastructure. As more businesses go further on their cloud journeys, they are finding that their investments are not as sound as they thought. Even hybrid cloud is not suitable for every organisation.
Going forward, companies should look at finding the right tool for what they want to do, evaluating all of the potential risks and cost implications, and make an educated decision that doesn’t rely only on the hype around the cloud.