BlogsRSS

Grace Mirandilla-Santos

Crowdsourcing data on PH mobile internet

At the Philippines’ 1st Telecom Summit last February, a number of speakers, myself included, presented data on the quality of internet service. Most of the data came from third-party foreign organizations and some commercial entities.

With the absence of official data, these companies seem to be the only reliable source of up-to-date information on Philippine internet.

Akamai Technologies and OpenSignal publish their reports (and methodologies) every quarter. M-Lab’s network diagnostic platforms are open for anyone to scrutinize and improve on.

There are also several other online testing tools and apps that internet users can download and use to test their ISP’s performance.

To be clear, the National Telecommunications Commission issued Memorandum Circular 07-08-2015 in August 2015 on measuring fixed internet service. However, at the Telecom Summit in March, the regulator only had a few samples (about 105 of institutions and households in total) to show. It also had no data on mobile broadband. This compared to the millions of data points on internet speed collected directly from consumers by Akamai, OpenSignal, and M-Lab.

Despite the abundance of global internet measurement tools, the availability of a localized testing tool would be an important milestone.

For one, the app developers would be more readily available to respond to queries about the measurement tool from the end-users.

They would also be more accessible when it comes to receiving and seeking inputs about the effectiveness and accuracy of the measurement tool.

Lastly, and more importantly, the developers would be more familiar with the local network architecture and the situation on the ground.

Last March, a group of Filipino volunteers led by Wilson Chua launched BASS or the Bandwidth and Signal Statistics app. The app measures mobile internet download speed by crowdsourcing data from end-users who voluntarily conduct the tests themselves. BASS is available on both Android and iOS.

The BASS volunteers envision the app to be a tool to help improve Philippine internet by providing data to both the consumers and the service providers. “Users can choose the carrier that is best for their location and carriers can use the data to re-allocate bandwidth from surplus areas to those that need more bandwidth,” says their profile on Facebook. BASS targets to get 1 million data points in order to do Big Data analytics.

Barely just a month after its launch, BASS uncovered what it called “anomalous routing” by a major telco. In his column, Chua pointed out that local traffic was being routed abroad despite using PLDT to fetch a file at the Department of Science and Technology (DOST) servers in the Philippine Open Internet Exchange (PHOpenIX). PLDT is peered with the government and has a direct fiber optic link to the PHOpenIX. Hence, Chua explained, the file should take around 20 to 30 ms. But BASS volunteers found that PLDT was not using the direct link, which increased latency to 399 ms or 1200% more delay.

Almost immediately, PLDT responded and explained that its bilateral peering agreement with the PHOpenIX tasks the DOST to “provide the list of IP addresses that will be advertised or opened to PLDT subscribers.” According to the press statement, the IP address that BASS was trying to access was not on the list of IP addresses given by DOST to PLDT.

A few days after, DOST responded with a statement clarifying that “its entire IP address block (202.90.128.0/19) is being provided and advertised to all peering partners; and that information about this address block was already available to all its partners in 2016 as part of the peering contracts or membership agreement documents.”

The routing issue has been fixed since.

The BASS app is a laudable initiative, which deserves support from the government, service providers, and consumers. Perhaps with some funding from donors, research institutions or other disinterested parties, the BASS app could be further improved and its reach expanded.

Some of the improvements that can be done include:

  • Measuring other metrics such as upload speed, latency, packet loss, and jitter;
  • Getting volunteers to conduct tests at given days and times, taking into consideration peak and off-peak hours, weekdays and weekends, and actual speeds versus advertised speeds, similar to LIRNEasia’s methodology;
  • Adding a date and time stamp to the result of each test

With the government’s launch of its national broadband and free Wi-Fi programs, it would be good for the Department of ICT and the NTC to recognize and take advantage of localized speed testing tools like BASS to benchmark and regularly measure Philippine mobile internet performance.