China Data Center Report

In March 2018, I made up my mind and published an article to share a closer look at landscape of submarine cables in India with 17 submarine cables landing in 15 cable landing stations (CLS) owned by 6 service providers. It was well received. Meanwhile 2 new submarine cable systems under construction – IOX and CACS, progressed steadily for RFS in Q4 2019. A new cable system from South Africa - SAEx2, that competes with IOX, also inched forward, though it remains to be seen if it gets funded eventually. A new cable from Singapore also appeared on the drawing board – SEAX-3.

Looking forward, every growth indicator that relates to cloud, content, subscriber and smartphone shows accelerated growth amidst highly competitive market with 4 leading players – Airtel, Vodafone Idea, Reliance Jio and Tata Communications, other than the incumbent BSNL, strengthens confidence that India does require more submarine cables, albeit with evolved mindset and rearchitected connectivity ecosystem. This article attempts to foresee the provider, network and market dynamics that will shape up in 2019-21 and make the new submarine cables that will land in India, more relevant for Digital India.

It’s not just new submarine cables, but also the cohesive interconnect ecosystem has to develop that integrates submarine cable systems, datacenters, metro-fiber and long-haul networks. Interconnected fiber networks from multiple providers with delightful reliability will facilitate new initiatives for multi-point capacity breakout, carrier-neutral interconnect hubs, datacenters architected ground-up or retrofitted to support direct submarine cable landing, pre-approved beach manholes developed with connectivity pathway built to datacenters interested to land submarine cables. This needs to be tale of two cities – Mumbai and Chennai that are the submarine cable landing hubs, but remain isolated when it comes to terrestrial transit for submarine cables. While there is sufficient raw fiber infrastructure available in these cities, a highly reliable service with metro dark-fiber and/or wavelength connectivity that connects datacenters, cable landing stations and business high-rises remains largely unavailable. This raises the wishful question – can we have a provider like Superloop that is committed to build dark-fiber network with 2000 fiber cables in 12-24 months and disrupt the market, can we have equivalent of NJFX and 1025Connect, can we have the equivalent of Seaborn Networks, Acqua Comms and RTI to build new submarine cables - in Mumbai and Chennai. It's possible provided the demand-supply dynamics at scale is stimulated, fiber infrastructure is monetized competitively and create market forces to challenge telco dominance.

It’s time submarines cable systems are developed by telco’s but without the legacy telco mindset, and with the openness of independent infrastructure developers. It shouldn’t be anything like the crowded fish-bone consortium cables, that more often than not have suffered historically from collective indecisiveness. They need to be focused cable systems to countries that are the global connectivity gateways. Some of the key routes are Mumbai-Marseilles and Mumbai-Dubai in west and Chennai-Singapore-Hong Kong in the east. Further in addition to selling sliced capacities of 10G, 100G and eventually 200G and 400G, there needs to be heightened focus to sell fiber pairs, physical and virtual along with managed spectrum. In order to facilitate such services, micro-datacenters need to be developed adjacent to CLS where the SLTE disaggregated from the rest of the dry plant allows multi-vendor SLTE provided and managed by service providers and OTT behemoths going for physical or virtual fiber pair to be hosted. This is how submarine connectivity has evolved in developed markets and India needs to catch up.

The new submarine cables must provide open access connectivity to democratize access to subsea connectivity and make it indispensable business enabler and growth driver. It is also needed to create greater competition in the market and make international connectivity more affordable, targeting audaciously to narrow the price gap with Singapore and London. For this to happen, the top 4 providers need to expand the number of submarine cables they land. Conservatively a reasonable number of submarine cable landing would entail, 3 cable systems in Mumbai and 3 cable systems in Chennai. Only Tata Communications meets this criteria for Mumbai. Airtel also meets it, with some caveats. If the criteria is diluted to 2 submarine cables in Mumbai and 2 in Chennai, other than Airtel and Tata Communications GCX qualifies for Mumbai and Airtel and Reliance Jio qualify for Chennai. This is clearly insufficient for commanding play in India market, in the next 5 years.

If the murmurs in the market are to be believed, 3-4 submarine cables are under discussion with public announcements expected in the course of 2019. For Chennai-Singapore 3,200 Kms 4/6/8 fiber-pair cable system with 24 Tbps per fiber pair, that is designed and negotiated in 2019, manufactured and deployed in 2020 reaching operational readiness in 2021 is expect to cost $100 million. Possibility of someone going solo seems farfetched at this time. Hence with tripartite collaboration, each party needs to bring $33 million to the table and unarguably that’s quite affordable if it brings 2 fiber pairs or 48 Tbps capacity which translates to raw price of $7000 per 100G. With Mumbai-Marseilles 9,000 Kms 4/6/8 fiber-pair cable with 24 Tbps per fiber pair is expect to cost $300 million, not including the branching units for landing in other countries enroute. Once again with tripartite collaboration, each party needs to contribute $100 million. If that’s for 2 fiber-pair ownership or 48 Tbps, it translates to $20,000 per 100G. Thus the war chest needs to have $120-150 million for participating in two cable systems in 2019. Given the capex crunch, the providers could negotiate 5-year financing options with TE SubCom, ASN/Nokia or NEC. Interestingly, there is precedence of some of these submarine cable system manufacturers retaining fiber-pair(s) as cash-equivalent and selling off subsequently.

Three groups of consortium members or co-investors are expected to participate in the new submarine cable system builds. The fourth group of independent infrastructure developers remains a wishful proposition. The first group is Indian providers as landing party with appetite to make investment of the order of $50-150 million. Reliance Jio is expected to take the lead to fuel its march to wireless broadband dominance and recent mega launch of FTTH rollout, in addition to being the party expected to acquire GCX submarine cable assets. Airtel and Vodafone could join hands to slow the march of Reliance Jio as they are exploring currently for fiber and tower infrastructure, provided the submarine cable has forked landing in their respective CLS, the way IMEWE lands with Airtel and Tata Communications in Mumbai. Tata Communications is also a likely participant as its inventory of 100G's in TIC is likely to get exhausted sometime soon and also needs diverse routes to Singapore.

The second group of co-investors would be the expected cloud and content hyperscalers - Google, Facebook, Microsoft and Amazon. Their much publicized and eyepopping investment in transpacific, transatlantic and Latam-US routes, indicates they will go for heavy investment in connectivity to India in order to have wholly owned submarine cable assets round the globe. Hence they are expected to invest in Chennai-Singapore and Mumbai-Dubai-Marseilles. As anyone’s guess Google will take the lead, followed by Microsoft and Amazon, and Facebook thereafter.

The third co-investors are the regional or global telco carriers. This group could have interesting co-investors, including China Telecom, China Unicom or China Mobile, other than more expected participants SingTel, NTT Com, Telstra and PCCW. There could be more for some of the cables depending on number of countries where the cables land.

The new submarine cables must land in datacenters directly for PoP-to-PoP connectivity and not the tiresome CLS-to-CLS with the struggle to arrange backhaul connectivity. The submarine cables scheduled to be operational in 2021 and beyond will hopefully follow this approach, particularly those being built by independent infrastructure developers. The two submarine cable landing hubs - Chennai in south east and Mumbai in north west are perfectly suited for this objective, as are also hubs for growth of datacenter. The challenge is for one or more provider to take the lead and drive cable landing in its datacenter, even if other submarine cables land in cable landing stations. 

The Mumbai-Chennai terrestrial corridor needs to be positioned or rather repositioned as major cost saving opportunity for Asia-Europe submarine cables. It was proposed for some of the submarine cables in the past, but the consortium dropped the proposal due to concerns with reliability. Mumbai to Chennai overland route is 1,200 Kms. However Mumbai to Chennai on sea route tromboning Sri Lanka is 4,400 Kms. Considering $30,000/Kms as construction cost, its $130 million. Comparatively two fiber pair for diverse paths is expected to be available for $15 million or even less. However the reliability of the route must be improved significantly with improved O&M. There are well known precedence of such terrestrial transit but for much shorter span in Egypt between Zafarana-Abu Talat (400 Kms) and in Thailand between Satun-Songkhla (123 Kms). It would be interesting for providers to conduct pilot runs with SMW4 and BBG to uphold merit of the proposition and monetize low latency routes.

For connectivity to Southeast Asia, a great Public-Private-Partnership (PPP) opportunity is set to unfold with the Chennai-Andaman Submarine Cable, that is contracted by the incumbent provider BSNL with NEC in March 2018 with completion timeline of Q4 2019. Submarine cable built from Chennai to Andaman Islands will reduce the distance to 3 countries -Thailand, Malaysia and Singapore by more than 50%. It will not be surprising, if in the next 3 years, improvements in submarine cable systems will enable distance of 1000-1200 Kms to be covered by comparably inexpensive unrepeatered submarine cables with some tradeoff with capacity.

A collaborative initiative to develop terrestrial route from the eastern state of West Bengal in eastern India through Bangladesh, Myanmar, Thailand, Malaysia to Singapore could open up a new connectivity corridor. Understandably a cross-border terrestrial route through 5 emerging-market countries would cast doubts on reliability. However with new road transport corridors being built across these countries with governmental collaboration, laying fiber along the highway does look feasible. Its for the providers in these countries to join hands and demonstrate unprecedented reliability of cross-border terrestrial connectivity. Also with new privately built cables like SIGMAR (Singapore-Myanmar) coming up, a combination of terrestrial and subsea routes could be combined to develop new route to Singapore.

With several options and possibilities, the new submarine cables to be announced in 2019 and 2020 are expected to catalyze a new connectivity ecosystem of seamless interconnect with harmonious multi-provider network play where SDN enabled network automation and automated workflows will lower end-to-end capacity provisioning times to few days. While aspirations with new submarine cables remain high, it also needs to be practical. Mumbai and Chennai are not the kind of intercontinental gateways that Fortaleza, Long Island, Singapore, Hong Kong or Tokyo are. With that said, let's stay tuned to exciting times ahead and await new submarine cable announcements to go public revealing consortium members and target timelines.