In part 1 of this blog post, I talked about how peering has helped me understand the beauty of the internet and the first reason why consumers need to be more aware whether their ISPs are peered or not.
The second reason is somewhat more difficult to prove. Does peering improve an ISP’s performance? Does it translate to better internet service to the end-user? What type of peering would be the most beneficial to the whole of Philippine internet?
To answer these questions, I collaborated with Mr. Wilson Chua, a Big Data analyst, and Measurement Lab or M-Lab, a consortium of research, industry, and public-interest partners. M-Lab is a partnership between New America’s Open Technology Institute, Google Open Source Research, Princeton University’s PlanetLab, and other supporting partners.
M-Lab provides the largest collection of open internet performance data on the planet using openly available measurement platforms and tools, which anyone, with the time and skill, can review and help improve. What is the difference between M-Lab and other broadband testing tools? Unlike some of the more popular internet performance measurement platforms, M-Lab’s networking diagnostic tool (NDT) places servers in an internet exchange point, outside of local ISPs, hence approximating consumer experience rather than an ISP’s core network performance. More detail on NDT’s methodology here.
The data set of M-Lab is mostly fixed broadband. There were almost 6 million tests in the Philippines, which resulted in over 22 million records from 2009 to February 2017 included in the analysis.
Here’s what we found:
Philippine internet service was a laggard from 2009 until end-2014, but started improving in 2015.
In 2009, Philippine internet had an average download throughput of 1.48 Mbps, but it started to pick up in 2015, reaching 3.39 Mbps by end-2015 and 7.64 Mbps as of February 2017.
Towards the end of 2016, the Philippines’ bandwidth began pulling up and away from other ASEAN neighbors like Viet Nam and Indonesia.
However, this improvement is associated with wired broadband, which makes up less than 5% of the Filipinos’ internet connection.
Public institutions showed the most significant improvement in bandwidth.
Beginning 2014, publicly funded institutions like Preginet, the Department of Science and Technology (DOST), the Department of Finance (DoF), the International Rice Research Institute, and the University of the Philippines showed exponential growth in bandwidth.
This improvement could be attributed to two things: (1) the connection of the government’s fiber optic network and (2) peering through the Philippine Open Internet Exchange (PHOpenIX), the only open and neutral IX in the country.
These public sector peers are multilaterally peered and have been beefing up their links to the IX over the years. This translates to lower cost and more efficient use of bandwidth, and should be encouraged, if not required by the government for all public offices.
Peered networks perform better than non-peered networks.
After peering, major telcos are shown to perform more optimally. Small ISPs and other peered networks are also performing very well. In fact, it is small companies that are leading the pack of commercial ISPs with high download throughput.
Given these crystal-clear benefits of peering, a few thoughts come to mind.
Will PH internet performance improve more for everyone if all peering were done multilaterally? Instead of a few companies increasing bandwidth, would it help the whole Philippine internet community if all networks take advantage of the open and multilateral peering arrangement that PHOpenIX has to offer?
If peering increases ISP performance, should peering be required by government? Of course government cannot dictate business decisions on private entities. But what if certain business decisions, like routing of what is otherwise local traffic to servers abroad, negatively affect the Philippine internet ecosystem and consumers in general?
With the national broadband plan setting the direction for Philippine broadband development and improvement, which entities and activities will drive bandwidth in the country? The NBP identifies open access and peering policy as priority reforms. But if ISPs are allowed to tweak the open peering policy of the PHOpenIX, would this dampen efforts to improve broadband service for everyone?
Finally, we need to raise awareness about peering. As consumers, we should ask our ISPs if andhow they are peered. And thanks to efforts of civic-minded Filipinos like Wilson Chua and his BASS (Bandwidth and Signal Statistics) volunteers, this is already happening now (more on this in my next post).