EAC-C2C

EAC-C2C is a 36,500 km submarine cable system forming a ring around East Asia, with sixteen landing stations across seven countries and territories: Japan, South Korea, China, Taiwan, Hong Kong, the Philippines, and Singapore. Ready for service in 2002 and consolidated under Telstra ownership in 2011, it is one of the longest contiguous submarine cable systems ever built — and it works precisely because no single data stream traverses its full length.

The name EAC-C2C is itself a corporate consolidation artefact. C2C (City-to-City Cable System) was the original 17,000 km intra-Asia ring built by Asia Netcom in 2002; EAC (East Asia Crossing) was a parallel 19,500 km system. Both landed at overlapping stations, both targeted the booming intra-Asia wholesale market of the early 2000s, and both found themselves competing for the same capacity buyers. Telstra acquired Reach (the Telstra-PCCW joint venture that had previously consolidated them) and brought both systems under a single operational umbrella. Today they are managed as one cable — with redundant paths that deliver precisely the fault tolerance ring systems were designed for.

36,500 km on paper, ~7,000 km in use for any pair

Our monitor samples EAC-C2C between Changi North (Singapore) and Ajigaura (Japan). This is one of many possible pairs on the ring, but it happens to be a heavily-used route: both landings sit in major financial data hubs, and the Singapore–Tokyo corridor is one of the most commercially important intra-Asia links. Across thirty days we collected 43 samples:

The minimum of 69.83 ms corresponds to roughly 6,800 km of fibre one-way, round-trip. Against the full cable length of 36,500 km, our observed RTT is 0.195× the theoretical physics floor for the whole system. That ratio is not a measurement glitch — it is the ring architecture made visible. Light traversing between Singapore and Japan uses only the segment of the ring that connects those two landings directly, not the entire perimeter.

Why the ratio tells you the architecture

A point-to-point cable like Marea has one endpoint pair; its physics floor matches its physical length almost exactly. On our measurements, Marea sits at 1.95× floor because the full cable length is in use.

A coastal cable like AMX-1 has many landings but monitored flows traverse most of them; its floor ratio of 1.66× the great-circle reflects that AMX-1's "in-use" length is close to its physical length, just routed along the coast rather than across open ocean.

A ring cable is different. The "in-use" length for any flow is the arc between the chosen endpoints, which can be much shorter than the full perimeter. On EAC-C2C, Singapore–Japan uses perhaps 5,000–7,000 km of fibre — the western arc that hugs the coast of Vietnam and swings up through the Philippines and Taiwan to Japan. The other 30,000 km of the ring carries different traffic: Hong Kong–Korea, China–Taiwan, Singapore–China, Philippines–Japan, and so on. Every pair of landings uses its own arc, and the ring serves all those pairs simultaneously.

Expressed as a fraction of the full cable, any single pair is using 15–25% of the total fibre length. The 0.195× number we measure on Singapore–Japan is a direct visualization of that geometry: 0.195 ≈ 1/5.13, which is roughly the fraction of the 36,500 km ring that happens to connect those two specific landings.

Sixteen landings, seven countries

The sheer number of landings — sixteen — is characteristic of ring cables built in the 2000s. Point-to-point transoceanic cables of the 2020s land at two, three, or four stations; ring cables of the early 2000s landed at a dozen or more because their whole purpose was to serve regional connectivity, not to deliver a single route between two endpoints. Every landing adds a shore approach, a cable station, a slice of permitting paperwork, and a service offering to local carriers. Sixteen landings means sixteen independent regional markets served by the same piece of infrastructure.

Who uses a 2002 ring in 2026

At 24 years old, EAC-C2C is approaching the end of the design life of its wet plant. The original design target was 25 years of operational service on the repeaters and fibre; upgrades to the dry plant (coherent transponders at landing stations) have kept capacity growing over that period — from the ~7.68 Tbps it was lit with in 2002 to 17.92 Tbps of reported capacity today according to Submarine Networks' specs. That is a modest capacity by 2020s standards; modern trans-Pacific cables deliver 100–300 Tbps per system.

But capacity is not the only reason a cable remains in service. EAC-C2C's value proposition in 2026 is its dense landing footprint. A carrier offering wholesale intra-Asia connectivity needs to deliver at many cities, not just two — and building a new ring cable around the same set of stations would cost billions. The incumbent infrastructure remains cost-effective as long as the fibre pairs keep lighting and the landing stations keep operating. Telstra's global wholesale division uses the system to offer managed capacity across its Asian footprint; regional carriers lease fibre pairs or wavelengths to build their own domestic offerings on top.

The cable's eventual successor will not be another ring. The industry model has shifted decisively toward point-to-point cables with fewer landings, higher per-fibre-pair capacity, and coherent-compatible wet plant. SJC2 (2025) and Apricot (2025) are the intra-Asia successors, each serving a subset of what EAC-C2C's ring covers with higher per-pair capacity. But they do not replace the breadth of landings — that is still a property unique to the older ring systems, and the reason EAC-C2C continues to carry paying traffic.

What our data proves

• Singapore → Japan at 69.83 ms minimum over 11 hops. A clean measurement on the Changi North – Ajigaura arc of the ring, using a small fraction of the full cable.
• The 0.195× floor ratio is the ring architecture. Full cable is 36,500 km; any landing pair uses only an arc. Singapore–Japan is one-fifth of the perimeter, and our measurement shows exactly that geometry.
• Reverse direction tighter than forward. Japan → Singapore has standard deviation under 3 ms across six samples; Singapore → Japan varies by 24 ms. Likely a difference in the IP-layer routing, with the Japan-outbound path settling on a single carrier while the Singapore-outbound path has several options actively used.

Try it yourself

Live measurements on the EAC-C2C cable page. Compare the ring architecture with point-to-point cables like SJC2 (direct Singapore–Japan, 1.05× floor on its full 10,500 km length) and Apricot (Japan–Singapore–Guam–Philippines, closer to a hub-and-spoke than a ring). The three cables together make the architectural choices visible: point-to-point optimises latency between two specific landings, rings optimise coverage across many landings at the cost of carrying any single flow over only part of the cable.