Europe India Gateway (EIG)

Europe India Gateway (EIG) is a 15,000 km submarine cable that connects twelve countries on three continents — the United Kingdom and Portugal at the European end, then Spain, France, Gibraltar, Monaco, Egypt, Saudi Arabia, Djibouti, Oman, the United Arab Emirates, and finally Mumbai, India. Ready for service in 2011, EIG has been in the ground for over a decade. Unlike the hyperscaler cables we have covered elsewhere on this site (Marea, Equiano, Nuvem), EIG was a classic carrier consortium cable — sixteen telecom operators pooling their capital to build a single system through the Mediterranean and across the Arabian Sea.

The cable has two fibre pairs with a design capacity of 3.84 Tbps. That was a sensible number for 2011. By 2026 it is modest — a single modern hyperscaler cable like JUPITER carries more than that on a single wavelength. EIG's value today is not raw capacity. It is its existence: 15,000 km of established fibre through some of the most geopolitically complex waters on Earth.

What our measurements show is that the cable has a latency personality split cleanly into two halves.

Westbound hits the physics floor

Over the past 30 days, our probe near the Sesimbra landing in Portugal pinged a Mumbai target 15 times with clean returns. The results are astonishingly consistent.

Light in a submarine fibre travels at about 204,500 km/s, so the absolute minimum round-trip for 15,000 km of glass is 146.7 ms. Our Sesimbra → Mumbai measurement comes in at 166.6 ms — 1.13× the physics floor. That is a best-in-class number for a 15-year-old cable, comparable to what we measured on the brand-new Marea (1.95×) and noticeably tighter than Equiano (2.5×). EIG, traversed westbound, is delivering latency within 20 ms of what glass physics allows, along a 15,000-kilometre path.

Eastbound takes a completely different route

The reverse — Mumbai toward Sesimbra — measures 299.7 ms minimum, 320 ms average, with frequent spikes past 400 ms. Twenty-three IP hops vs fourteen to twenty westbound. This is not the same path.

At 299 ms round-trip, the fibre-equivalent distance is about 30,600 km. That is more than twice the length of EIG itself. The return packet is almost certainly not using EIG at all — most likely it is routed eastward through Southeast Asia (SEA-ME-WE 4 or 5, or one of the newer Asia-Europe systems), through Singapore or Hong Kong hubs, and then westward through a second set of cables to reach Portugal. The 23-hop count corroborates: a packet crossing multiple independent cables accumulates more routing decisions and therefore more hops.

Why does the return take the long way? Three usual suspects:

• Cost-based BGP preference. The Indian transit provider hosting our Mumbai-side probe has commercial contracts and private peering arrangements. Those may send Europe-bound traffic through a Singapore or Hong Kong hub operated by a partner carrier, regardless of whether EIG offers a shorter physical path.
• Capacity or reliability preference. EIG has only two fibre pairs and a 15-year-old repeater chain. A carrier that owns or leases capacity on a larger, newer cable may route production traffic through the modern system and reserve EIG for overflow or backup.
• Peering topology. The target address on the Portuguese side is behind a network that peers extensively with Asian carriers but less directly with European ones — so Indian operators reach it via Asia-Pacific relationships, not Mediterranean routes.

What our data cannot tell us from outside is which of the three dominates. What it can tell us is that from Mumbai's perspective, the 166 ms one-way latency that EIG could deliver is untaken. Europe-bound traffic pays an extra 130 ms to go the other way around the world.

Twelve countries on one consortium

EIG's landing list reads like a map of twenty-first century geopolitics:

The cable crosses the Mediterranean, lands in Egypt, crosses the Sinai on land (the cable has to traverse Egyptian territory between Mediterranean and Red Sea landings), passes through the Red Sea, exits via the Bab-el-Mandeb strait between Djibouti and Yemen, crosses the Gulf of Aden, and arrives in the Arabian Sea via Oman and the UAE before terminating in Mumbai. Every stretch of that route is geopolitically sensitive.

The Bab-el-Mandeb strait is perhaps the most notable. It is the bottleneck of Red Sea cable traffic — almost every cable between Europe and South Asia passes through those narrow waters, along with tankers, container ships, and several military fleets. Cable faults in this area have knocked substantial capacity off the global internet on more than one occasion. EIG is one of many cables that share this vulnerability.

Two fibre pairs in the age of twenty-four

EIG was designed with two fibre pairs. That was a strong spec for 2008 when construction started — SEA-ME-WE 4, its contemporary, has three. But modern cables built for the hyperscaler era run four, eight, twelve, and now twenty-four fibre pairs through the same conductor. The more recent Medusa, a 2026 Mediterranean cable of similar length, has 24 pairs and a design capacity of 480 Tbps — more than a hundred times EIG's rated 3.84 Tbps.

Why did EIG's designers pick just two pairs? Same answer as why early transatlantic cables had fewer still: cost. Fibre pairs are not the bulk of a cable system's cost — the cable body, power feed, and repeater amplifiers are. But each additional pair means more wavelengths to light on each end, and in 2008 the coherent modulation and electronics to exploit a high fibre-pair count did not yet exist commercially. Two pairs was a deliberate choice matching the economics of the time.

The system has been upgraded multiple times since commissioning. The underlying glass hasn't changed, but the transponders at each end have stepped through 10 Gbps, 40 Gbps, 100 Gbps, and now 200 Gbps per wavelength. Each upgrade extends the cable's commercial life without laying a single new kilometre of fibre. That is the quiet genius of submarine infrastructure: the hardware in the ocean is meant to last 25 years, and the only part that actually advances is the dry-side electronics.

What our data proves

Three takeaways from 60 measurements over 30 days:

• EIG still hits physics floor. Sesimbra → Mumbai at 166.6 ms is within 20 ms of the theoretical light-speed minimum for 15,000 km of fibre. Fifteen years on, the cable delivers exactly what it was designed to.
• Return traffic doesn't use it. Mumbai → Sesimbra at 300 ms is physically double the one-way latency EIG could deliver — the Indian-to-European return path is using a 30,000 km detour through Asia.
• The asymmetry is stable. Both patterns have held steady for at least 30 days. This is not a temporary incident — it is the steady-state routing policy of the carriers involved.

EIG is the kind of cable that no one writes about. It was a second-tier investment at its commissioning, overshadowed by newer and longer trans-continental systems. But a decade and a half later, our measurements show it is quietly delivering near-theoretical performance along one of the most geopolitically dense routes on the planet. That is, itself, a story worth telling.

Try it yourself

Live data on the EIG cable page. Compare with our coverage of Marea (Atlantic, 1.95× floor), JUPITER (Pacific, asymmetric), and the brand-new Medusa in the Mediterranean.