Sat-Ed is proud to have been profiled by CNN and other major media outlets... (
Focus Broadband satellite where it's most needed
IPStar is best suited to 'IP multicasting'
Don Sambandaraksa
On the exact same day that IPStar (a.k.a Thaicom 4) blasted off into the heavens, my dependence on satellite broadband via Thaicom 3 ended when True finally provided an ADSL link to my little house just north of Bangkok.
Until last week, the only way for me to get fast Internet access was to shell out 10,000 baht a month, first to Loxinfo and later Samart, for their IPStar based satellite broadband service, which while branded as IPStar is actually IPStar emulation via Thaicom 3. Yes, 10,000 a month for something which urban residents would expect to pay 590 baht a month for. Such is the cost of living on the other side of the digital divide. While I could afford that kind of money, what of the typical rural villager?
I have said it before and I will say it again, without a proper regulatory regime (proper being the imperative word here), the way networks are being rolled under today's concession system only serves to reinforce the urban-rural divide, not bridge it.
Worse, it was 10,000 baht a month for 256Kbps access that does not work when it rains (apparently the frequency used is affected by raindrops, which just happen to be the same size as the microwave band used), does not work when there is solar flare activity, and often just does not work because something is wrong with the network somewhere.
Half the time, large packets were just dropped while small packets arrived fine (meaning that downloading or viewing images was out, but things like instant messaging were fine). Worse, even in the best of times, a ping time of between 600 and 900 milliseconds means that Voice over IP is impossible and many programs like Google's Picasa just keep throwing time-out errors because of the latency.
In a case of mastering the "Somebody Else's Problem Field," the problem was never one of my ISP, Samart. One of their engineers cheerfully tells me every time there is a problem that it was either because of Thaicom or with the Internet gateway, but that their systems were fine. Since that same senior engineer could not tell the difference between a 2Kb packet (size), and 2Kbps bandwidth (flow rate), I voted with my chequebook, cancelled their service and plodded back across the bridge to the other side of the digital divide, resigned to depending on GPRS as my primary means of access to the Internet.
And then, quite unexpectedly, True's ADSL service came along and ever since I have been enjoying a much better service than the previous satellite connection for only 6% of the price.
So, is the future of connectivity with IPStar? Would it have solved any of the problems I had when I was using Thaicom 3?
More capacity would help, but even with a parliamentary majority, one law that cannot be changed is the law of physics. The radio frequencies used will still be the same size as a raindrop and hence rain will continue to interfere with the signal.
Similarly, no amount of bureaucratic reform can make radio waves travel any faster than what they already do _ the speed of light. Even with IPStar, packets will still have to spend the better part of half a second travelling to and from the satellite.
While 40 gigabits per second of capacity (that's around 40,000 Mbps or 40,000,000 Kbps) may have been a lot when the satellite was designed years ago when most people were using 56Kbps dial-up, in today's world of 28.7Mbps ADSL2+ (which Thailand still does not have) 40Gbps only equates to just under 1,500 ADSL2+ users running at full speed. Not a lot.
The powers that be should have learned from the rollout of 3G in Japan and Korea. Video calling, being a point-to-point paradigm, eats up bandwidth like there is no tomorrow and, worse, the demand for bandwidth scales in a linear manner with increased numbers of users.
What the Koreans and Japanese have discovered _ and what Thailand should pay attention to _ is how they have turned their 3G infrastructure from a point-to-point to a multicasting affair. Many of the features we associate with 3G _ video advertisements and video clips for news or sports _ can be effectively multicast (ie, sent from one point to many using bandwidth just once) and cached at the point of consumption. This model is much more scalable. Just think _ news can be cached at the phone at hourly intervals, and if someone wants to view it (and pay for it), they have access to it instantly.
The strength of IPStar is, according to Hawker, IP multicast rather than IP point to point. Take e-learning content, for example. These could be broadcast and cached in boxes with limited interactive bandwidth and held until they are unlocked, perhaps by pay per use, perhaps by being called upon. Careful profiling of the user can lead to most things being cached correctly. This will take care of the vast majority of the bandwidth demand while what is not cached can then be downloaded via (relatively bandwidth-expensive) point-to-point transmission.
The risk is that in the euphoria of having the world's first broadband satellite, we will miss that point entirely and focus IPStar only on providing access to ADSL-deprived areas, areas which with 3G mobile technologies like EDGE and CDMA2000 are already serving. Whether IPStar turns out to be a white elephant or a bridge into a multicast future will soon be apparent.
Don Sambandaraksa is doing his bit to advance awareness of the difference between free speech and free beer in the corridors of power in Thailand.
