mobile internet access never to improve unless tariffs go upward. Fiber to become a nightmare for inexperienced telcos?
To me it seems telcos here are very inexperienced in data services and in the process have burnt their fingers to offer reliable data services. Why? Simply, they modelled the data services similar to voice services. While on voice they could compress a number of voice channels per BTS capacity to say eg 8kbps, data is a different monster all together. TCP/IP alone has massive overheads and trying to squeeze it to work on 8-16kbps channels simply cannot work. The costs/profits to a mobile telco are based on the number of BTS channels utilized and congestion models that have predictive call behaviour. The congested internet backbone has until now helped them keep control on BTS overheads and channel allocations. For example, we can say that for an 8k channel allocation the business model for voice call is 1 Ksh and timed per second. During a voice call, the caller will naturally hang up asap and this allows another caller to use the same time slot. On data, the same 8k channel is timed per Mbyte. This Mbyte may take upto 10 minutes to clear. Therefore during this period, the channel cannot be used by anyone else and so queues build up proportionally to congestion. Once the fiber goes live, these BTS models will not be able to handle the demands. In many ways, not having the fiber has been a big advantage to telcos as the backbone self congested and slowed down. In fact if safaricom 3G don't follow certain rules, they may see a network crash within the first few hours of the fiber being lit. They seem to have learnt a few things and thus the reason for blocking smtp and torrents. Foolish and amatuer ways of network management ( I blame the sales management, not engineers! ). Orange has also adjusted their rate bundles. Yu has some sort of primitive channel controllers that drop packets if the speed rate/ports is exceeded. I believe tariff rates will have to go upwards if telocs are to improve on data delivery. Somehow safaricom seems to be the culprit. It comes up with cheaper rates and then not able to offer reliable services. Others follow suit and none can offer any reliable services. Teclos have tried to compete with traditional ISPs and their infrastructure will never allow it. Inexperience shines like a light on a dark street. my thots....
Hello, Yes you are right about GSM but lets not forget 3G (safcom) is basically UMTS (WCDMA) this is not channelized per se GSM (2G, 2.5 blah) is the one that is channelized (time slots) due to the modulation scheme which will be frequency related, meaning a GSM sites capacity will be determined by the number of TRX (frequencies) x 8 ( channels/ time slots) UMTS on the other hand is using Code to modulate so two frequencies ( at least its designed to be like one freq for voice and the other for packet one can be used though) Safcom used to use one freq during trials I do not know if they added an extra freq. So in theory a 3G Cell does not have alimit to the number of guys who can be hooked to it at the same time calling but again the more the calls (be it PS or CS) the more the cell shrinks due to power blah blah.. . Their only solution is to have many 3G sites which is costly since its a completely new roll out .. // Watson 2009/6/27 aki <aki275@googlemail.com>
To me it seems telcos here are very inexperienced in data services and in the process have burnt their fingers to offer reliable data services. Why? Simply, they modelled the data services similar to voice services. While on voice they could compress a number of voice channels per BTS capacity to say eg 8kbps, data is a different monster all together. TCP/IP alone has massive overheads and trying to squeeze it to work on 8-16kbps channels simply cannot work. The costs/profits to a mobile telco are based on the number of BTS channels utilized and congestion models that have predictive call behaviour. The congested internet backbone has until now helped them keep control on BTS overheads and channel allocations. For example, we can say that for an 8k channel allocation the business model for voice call is 1 Ksh and timed per second. During a voice call, the caller will naturally hang up asap and this allows another caller to use the same time slot. On data, the same 8k channel is timed per Mbyte. This Mbyte may take upto 10 minutes to clear. Therefore during this period, the channel cannot be used by anyone else and so queues build up proportionally to congestion.
Once the fiber goes live, these BTS models will not be able to handle the demands. In many ways, not having the fiber has been a big advantage to telcos as the backbone self congested and slowed down. In fact if safaricom 3G don't follow certain rules, they may see a network crash within the first few hours of the fiber being lit. They seem to have learnt a few things and thus the reason for blocking smtp and torrents. Foolish and amatuer ways of network management ( I blame the sales management, not engineers! ). Orange has also adjusted their rate bundles. Yu has some sort of primitive channel controllers that drop packets if the speed rate/ports is exceeded. I believe tariff rates will have to go upwards if telocs are to improve on data delivery. Somehow safaricom seems to be the culprit. It comes up with cheaper rates and then not able to offer reliable services. Others follow suit and none can offer any reliable services. Teclos have tried to compete with traditional ISPs and their infrastructure will never allow it. Inexperience shines like a light on a dark street.
my thots....
_______________________________________________ Skunkworks mailing list Skunkworks@lists.my.co.ke http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks Other services @ http://my.co.ke Other lists ------------- Skunkworks announce: http://lists.my.co.ke/cgi-bin/mailman/listinfo/skunkworks-announce Science - http://lists.my.co.ke/cgi-bin/mailman/listinfo/science kazi - http://lists.my.co.ke/cgi-bin/mailman/admin/kazi/general
-- Watson wanjohi kambo
participants (2)
-
aki -
Watson Kambo