Telenor fixes/brands traffic huts in Yangon and Mandalay. Results: mixed
Localizing the internet in the Philippines
In this interconnected world, no economy can grow without the internet. Several studies have shown that a 10% increase in broadband penetration could improve gross domestic product (GDP) and labor productivity, especially in low– and middle-income countries. But apart from access, the quality of connection defines how the internet can bee exploited for development.
Internet exchange points (IXPs) are facilities that allow ISPs to connect directly to each other and exchange traffic using local infrastructure. Naturally, the shorter distance lessens latency, improves connection quality, and reduces the transmission cost of using third-party networks and expensive international links for routing traffic.
The Internet Society recently released a pioneering study that, for the first time, quantifies the positive impact of IXPs in emerging countries, like Kenya and Nigeria. The report attributes annual savings of at least $1 million for telcos, plus millions in additional revenues from new traffic, speeding local data exchange, and encouraging locally hosted content and services.
While developed nations in North America, Europe and Asia have long reaped the benefits of IXPs, countries like the Philippines have yet to exploit it. Internet users in the Philippines contend with a rather ridiculous situation. If A and B are subscribers of different ISPs, data from A will have to travel outside the country—making several hops to the US then to Hong Kong and Singapore—before it comes back and finally reach B. All this, even if A and B are just five kilometers apart!
Clearly, this sounds illogical. But the mere fact that major ISPs are not interconnected means that this makes (business) sense to at least one player.
Not that there is no IX facility available. In 2007, the Advanced Science and Technology Institute (ASTI), the research and development arm for ICT and microelectronics of the Department of S&T, developed the Philippine Open IX as an open peering facility for local ISPs. Several players have already connected to the IX, but without the major ISPs, its impact remains very limited.
In 2011, the National Telecommunications Commission issued a draft memorandum circular mandating all telcos with international private lines (IPL) to peer at PHOpenIX. The first public hearing was scheduled in July, but after barely a month, the regulator dropped the MC and instead ordered the telcos to “negotiate” peering with each other. This threatens to mimic a similar standoff over 10 years ago when major telcos (the same dominant ISPs today) failed to negotiate interconnection until the President “mediated.” While local ISPs are not connected, internet quality remains wanting.
Perhaps the Philippines can learn a thing or two from Indonesia, which successfully set up an IX in 1997 that now connects all local ISPs. Johar Alam Rangkuti, a Philippine-educated techie who co-founded the Indonesian IX (iIX), recently gave a talk in Manila and offered some interesting insights about their open IX.
First, peering is free of charge. The iIX facility is a collection of devices from various ISPs, so maintenance is covered. It is administered by a non-ISP so nobody doubts its neutrality. Its system is very secure; no hacker has successfully breached it. Finally, the iIX’s business model is market-driven. Instead of convincing the players to link up for the “common good,” Johar’s team showed the ISPs how connecting to the iIX switch is more financially advantageous than negotiating with each other.
Once the small ISPs connected, the big telcos saw that there was business to be made and soon followed suit. The exponential growth in the iIX traffic, today peaking up to 60 Gbit/s from a measly 2Mbit/s years ago, is proof of its success.
The PHOpenIX similarly has a huge potential, but needs to deal with some critical issues. Although its service is also free, locating at a data center of an ISP (GlobeQuest) compromises its neutrality and deters other players. It should also minimize direct costs, like the leasing of IP addresses, and show readiness to expand.
Most importantly, the PHOpenIX needs a viable business model. Should it be government-administered or would a private entity manage it better? Does it know its market - how many ISPs and potential clients are out there? What can it offer to attract the major ISPs? Is there money to be made and how much? Tough but necessary questions. Because in this multi-million dollar internet industry, the PHOpenIX can only serve the common good once it makes business sense.