Why Max?
August 3, 2009Cyberpress held a forum Friday last week on the topic “Why WiMax?”. The question, in more detail: “Is this really the technology that will bring affordable and reliable Internet connection to more and more homes in the Philippines? What are useful qualities of WiMAX versus other wireless technologies?” This was held at Coffee Bean and Tea Leaf at A.Venue on Makati Avenue, with presenters from Intel Philippines, Globe Telecoms, and SMART Communications. Also present were executives from Lenovo, Acer, and HP (through their PR agency). I expect the proceedings to be posted on the Cyberpress Website “soon” so I’ll go straight to a question and two comments I made during the forum. I did warn especially the two telcos present that they might not like the second comment.
My question had to do with the state of the Internet Exchange in the Philippines. I’m on PLDT DSL, and if I wanted to visit a site on the Globe network, my packets would have to travel out of the country, likely to the US, and travel back to the RP and thence to the Globe site. That means I’d get faster response from, say, www.apache.org than from www.globe.com.ph — in fact, some traceroutes (at the end of this article) bear me out. In brief, www.apache.org is 17 hops away from my PC, and www.freesoftwaremagazine.com, in Australia, is 16 hops away, with my packets having to travel via HK, Tokyo, and Los Angeles before hitting Western Australia. I couldn’t get complete traceroute results with globe.com.ph, because it would time out starting from hop # 18. And oddly enough, traceroute wouldn’t complete on pldt.com.ph either. Furthermore, my Firefox manages to load www.globe.com.ph and www.pldt.com.ph so traceroute doesn’t mean much in terms of accessibility of sites in one network (Globe) from the other (PLDT). Nevertheless, the point is that we’re using up too much of our international connectivity. I realize that this may be a chicken-and-egg problem: local hosting is more expensive than US hosting, because the US market is much larger than the Philippine market, and local content providers are making the former market even larger and the latter smaller. However, I have reason to believe that if more of our traffic were local, we could enjoy better speeds than the carriers currently provide.
On to the comments: my first had to do with access devices. I believe we’re too enamored with the PC as the Internet access device. But let’s face it: the PC may be overqualified – and presents too steep a learning curve – for much of what most users will do. Of course, we have to determine what those “most users” want, although off-hand, we can think of email, social networking (i.e., FaceBook), and video and voice calls over the Internet. Let’s challenge hardware manufacturers to design alternative access devices in the mold of the Mobile Internet Device (MID). But that’s just the start: the software has to be simple enough to learn and use – think of Android and Chrome.
My second comment was in regard to the telcos’ revenue model, which I called “outmoded”. Look at the pricing table below and tell me what’s wrong with the picture:
|
Service |
Monthly Service Fee |
|
512kbps Internet-only |
795 |
|
512kbps Internet+landline |
995 |
|
1024kbps Internet-only |
995 |
|
1024kbps Internet+landline |
1295 |
WiMax, as an IP network, can carry voice data via Voice-over-IP (VoIP). Doesn’t this render the landline redundant? Or are the telcos doing this to protect their voice services? And to further protect those voice services, might the telcos block VoIP services such as Skype? But can they stop the voice- and video-capable instant messengers, e.g., GoogleTalk and Yahoo! IM, and other smaller services that may crop up? Are they going to be like the music “industry” that refused to adapt to the Internet and instead has been fighting a losing battle against peer-to-peer networks? The music “industry” is dying in its present form, and the sooner it´s gone, the better – for its own sake, so it can take on another, more 21st century form. (As an aside, their calling themselves an “industry” is a reflection of their Industrial Age thinking. We´re in the Information Age, folks, get with it.) Our WiMax providers can keep themselves in the game by adapting. And I leave that as an exercise for the reader.
Annex: traceroute results
traceroute to www.freesoftwaremagazine.com (216.75.55.208), 30 hops max, 40 byte packets
1 192.168.1.1 (192.168.1.1) 3.518 ms 3.477 ms 3.568 ms
2 58.69.88.254.pldt.net (58.69.88.254) 8.886 ms 7.782 ms 7.867 ms
3 58.69.255.25 (58.69.255.25) 8.581 ms 8.125 ms 8.575 ms
4 58.71.0.92 (58.71.0.92) 8.397 ms 8.642 ms 8.319 ms
5 if-1-1.core3.HK2-HongKong.as6453.net (116.0.82.37) 352.833 ms 364.221 ms 370.721 ms
6 if-5-0-0.core1.TV2-Tokyo.as6453.net (116.0.82.2) 356.307 ms 415.324 ms 348.061 ms
7 if-6-0-0-737.mcore4.LAA-LosAngeles.as6453.net (209.58.61.38) 346.821 ms 349.583 ms 357.116 ms
8 Vlan77.icore1.LAA-LosAngeles.as6453.net (216.6.85.46) 387.995 ms 360.331 ms *
9 192.205.35.129 (192.205.35.129) 351.747 ms 352.852 ms 355.628 ms
10 cr2.la2ca.ip.att.net (12.122.129.34) 316.948 ms 317.587 ms 311.435 ms
11 cr2.sd2ca.ip.att.net (12.122.31.9) 316.075 ms 364.693 ms 326.078 ms
12 gar2.sd2ca.ip.att.net (12.123.215.197) 309.769 ms 310.807 ms 309.864 ms
13 12.86.27.90 (12.86.27.90) 326.926 ms 312.618 ms 313.253 ms
14 216.98.153.53 (216.98.153.53) 364.147 ms 361.885 ms 350.056 ms
15 216.98.153.21 (216.98.153.21) 366.446 ms 346.448 ms 347.561 ms
16 centos555208.aspadmin.net (216.75.55.208) 351.809 ms 355.254 ms 368.857 ms
traceroute to apache.org (140.211.11.130), 30 hops max, 40 byte packets
1 192.168.1.1 (192.168.1.1) 3.578 ms 3.524 ms 3.602 ms
2 58.69.88.254.pldt.net (58.69.88.254) 10.003 ms 8.172 ms 7.905 ms
3 58.69.255.25 (58.69.255.25) 8.574 ms 8.396 ms 8.559 ms
4 58.71.0.155 (58.71.0.155) 8.580 ms 8.583 ms 8.671 ms
5 if-1-1.core3.HK2-HongKong.as6453.net (116.0.82.37) 354.193 ms 345.239 ms 341.687 ms
6 if-5-0-0.core1.TV2-Tokyo.as6453.net (116.0.82.2) 344.469 ms 348.179 ms 352.155 ms
7 if-12-0-0.mcore4.PDI-PaloAlto.as6453.net (216.6.86.25) 369.889 ms 372.858 ms 371.287 ms
8 if-0-0-0.core3.SQN-SanJose.as6453.net (216.6.33.5) 368.252 ms 361.201 ms 364.083 ms
9 if-7-2.icore1.SQN-SanJose.as6453.net (209.58.116.49) 366.115 ms 382.538 ms 378.449 ms
10 ix-2-0.icore1.SQN-SanJose.as6453.net (209.58.116.14) 368.494 ms 353.182 ms 352.961 ms
11 vlan89.csw3.SanJose1.Level3.net (4.68.18.190) 365.897 ms 367.731 ms 370.841 ms
12 ae-72-72.ebr2.SanJose1.Level3.net (4.69.134.213) 377.319 ms ae-62-62.ebr2.SanJose1.Level3.net (4.69.134.209) 374.601 ms ae-82-82.ebr2.SanJose1.Level3.net (4.69.134.217) 352.942 ms
13 ae-7.ebr1.Seattle1.Level3.net (4.69.132.50) 341.076 ms 342.467 ms 340.563 ms
14 ge-10-0.hsa2.Seattle1.Level3.net (4.68.105.7) 339.225 ms 337.153 ms 337.435 ms
15 UNIVERSITY.edge5.Seattle1.Level3.net (63.211.200.246) 340.270 ms 341.978 ms 340.707 ms
16 corv-car1-gw.nero.net (207.98.64.177) 342.889 ms 343.069 ms 365.365 ms
17 eos.apache.org (140.211.11.130) 343.045 ms 342.895 ms 342.694 ms
traceroute to globe.com.ph (203.177.23.57), 30 hops max, 40 byte packets
1 192.168.1.1 (192.168.1.1) 3.486 ms 3.541 ms 3.952 ms
2 58.69.88.254.pldt.net (58.69.88.254) 8.997 ms 8.030 ms 8.107 ms
3 58.69.255.25 (58.69.255.25) 10.140 ms 8.444 ms 8.680 ms
4 58.71.0.155 (58.71.0.155) 9.131 ms 8.769 ms 8.948 ms
5 if-1-1.core3.HK2-HongKong.as6453.net (116.0.82.37) 342.087 ms 350.317 ms 356.364 ms
6 if-5-0-0.core1.TV2-Tokyo.as6453.net (116.0.82.2) 359.165 ms 342.600 ms 339.228 ms
7 if-12-0-0.mcore4.PDI-PaloAlto.as6453.net (216.6.86.25) 351.313 ms 364.258 ms 371.727 ms
8 if-0-0-0.core3.SQN-SanJose.as6453.net (216.6.33.5) 372.939 ms 353.266 ms 358.625 ms
9 Vlan55.icore1.SQN-SanJose.as6453.net (66.198.97.10) 368.588 ms 381.125 ms 378.234 ms
10 ix-2-0.icore1.SQN-SanJose.as6453.net (209.58.116.14) 382.484 ms 354.898 ms 353.943 ms
11 ae-42-99.car2.SanJose1.Level3.net (4.68.18.196) 353.301 ms 356.985 ms ae-32-89.car2.SanJose1.Level3.net (4.68.18.132) 365.246 ms
12 4.79.42.166 (4.79.42.166) 373.015 ms 363.964 ms 352.772 ms
13 120.28.0.81 (120.28.0.81) 528.787 ms 533.328 ms 518.633 ms
14 * 120.28.0.6 (120.28.0.6) 545.573 ms 545.733 ms
15 120.28.0.150 (120.28.0.150) 539.235 ms 537.570 ms 523.007 ms
16 120.28.0.226 (120.28.0.226) 527.255 ms 531.247 ms 537.259 ms
17 203.177.69.134 (203.177.69.134) 547.955 ms 546.394 ms 537.037 ms
18 * * *
19 * * *
20 * * *
21 * * *
22 * * *
23 * * *
24 * * *
25 * * *
26 * * *
27 * * *
28 * * *
29 * * *
30 * * *
Previous Comments
Hi Jed. VoIP calls within the same network should be fine so QoS isn’t be an issue, or at least it shouldn’t be. When those VoIP calls though cross over to the other network — e.g., Globe to Smart or Smart to Globe — then their argument may have some merit. But this goes back, at least partially, to “local” IP traffic not being so local, since Globe-to-Smart or Smart-to-Globe traffic has to use up the international lines. I posit that the more local content we host locally, and the more we visit those local sites, the better our connection to sites outside the country.
Am I mixing the two arguments? You bet 
— but then I believe they *are* connected, in more ways than one 
As an aside, I got more reactions at the forum from the hardware vendors than from the cellular carriers. My sense is that they assert that the PC is THE Internet access device of choice. My suggestion is for them to design alternate devices that are easier to use and maintain (no registry or UAC nonsense, for one).
Posted by Daniel Escasa at August 6, 2009, 3:27 pmAll comments are moderated. Your comments will not appear here unless approved by the blog owner. Thank you.
Hi Danny…couldn’t the telcos always counter that with the Internet, there are bound to be QoS performance issues what with it being a public network outside their control, such as latency, congestion, etc? I’m sure they’d say at least with the traditional circuit-switched landline network, the whole network is theirs and thus completely under their control, so they can fix any problem anywhere it may arise there.
And besides, given the pricing packages above, they could always argue their revenue models aren’t so outdated…they derive most of their income from the Internet service already, not the traditional landline component…so what’s happening now is that revenue-wise, the landline part is just a basic fee to get Internet access at all.
Just thinking out loud.
Posted by Jonathan Domingo at August 5, 2009, 10:21 am