By Wilson Chua
A-Ar Andrew Concepcion was the first to notice this. While testing the BASS (Bandwidth and Signal Strength) analyzer, Andrew noticed that he got unexpectedly low bandwidth speeds. BASS measures the bandwidth that a user actually receives on his mobile phone. This was strange. It did not make sense to get low speeds. We were expecting higher speeds since he was using PLDT to fetch a file at DOST servers in PHOpenix.
Note: PLDT is directly linked to the Department of Science and Technology (DOST) via high speed fiber optic facilities to the DOST’s Philippine Open Internet Exchange (PHOpenIX)”. Conceptually, this is what was supposed to be happening:
1.BASS requests a file from mirror.pregi.net
2.PLDT sends the request over to PHOpenIX
3.PHOpenix will notify the file server
4.File Server (mirror.pregi.net) will answer the request and send the file.
The request should take around 20 to 30 milliseconds.
However, our investigation showed that PLDT was not using this direct link. PLDT was anomalously re-routing this traffic overseas to the US and then back to the Philippines. This action added 1200% more delay against the original (30 vs 399)
Why is this finding important?
This ‘smoking gun’ proves that the slow internet is NOT due to “user-abuse” as is often claimed by carriers. Neither is it due to lack of cell sites/towers. At least in this instance, this is solid proof that PLDT is at fault for “inserting” delays to its own network.
If PLDT had used the direct fiber link, their end users would have faster and more reliable internet. It is also a lot CHEAPER for PLDT to do so. Not surprisingly, BASS results as of April 18, 2017 shows that Globe outperforms PLDT.