From August to early October, the subsea cables around Singapore suffered one fault after another. If you host in Singapore, or reach servers from this part of the world, these events show up directly as higher latency and more packet loss — and repairs are measured in weeks.

Three Events, in Order

  • August 8 to 9: shunt faults on the Indigo West and Indigo Central cables — the earlier of the two incidents on the Perth–Singapore route.
  • Early September: as Light Reading reported on September 7, the Australia–Singapore Cable (ASC, 4,600km, operated by Vocus) broke in Indonesian waters between Anyer and Singapore, knocking out wavelength and Ethernet services. It was the second major incident on the Perth–Singapore route within four weeks. Traffic was rerouted over Australian east-coast cables via Japan or the United States, and Melbourne-to-Singapore detours reached 400ms. Vocus said it was planning the repair with a marine repair partner, including a technical review, confirming spare materials and coordinating permits, without giving a completion date.
  • Around October 2: the Philippines saw a nationwide drop in international bandwidth, affecting links toward Singapore and Hong Kong. Henry Aguda, Secretary of the Department of Information and Communications Technology, said the cause was still being determined — "either a fiber cut through natural causes or equipment failure". Globe, PLDT and Converge all switched to backup routes, with PLDT saying it had rerouted most of the affected traffic, and local media reported that repairs were expected to take several weeks.

Why a "Singapore Node" Feels Different Depending on Where You Are

Latency depends on the path, not on the city where the node sits. During a cable fault, the same Singapore server can be unchanged from one country and take a detour halfway around the world from another. When you see "it's slower to Singapore", first work out which leg is detouring.

  • Measure from the user's side: ping and mtr show which hop adds latency or loss — see how to test VPS latency and packet loss — and how to troubleshoot packet loss covers narrowing it down.
  • Check both directions: a healthy outbound path says nothing about the return path, and during cable faults the two often detour differently.
  • Don't conclude from a single test: carriers keep adjusting routes during an incident, and the same link can change several times in a day.

Practical Advice When Choosing Nodes

If your users are spread across Australia, the Philippines and several Southeast Asian countries, a single node becomes the weak point in events like these. Place entry points according to where users are, or at least keep a standby region you can switch to. The Singapore VPS buying guide explains what a Singapore node suits and where it falls short, and current plans are on the Singapore node page. Looking further ahead, as we noted in our piece on APAC's record data-centre pipeline, Singapore has very little capacity under construction — so the reliability of cables and cross-border paths will only matter more for business in this region.