I must admit, when Sentech first announced their MyWireless product I was quite happy. Of course, their product was a dud.
Now iBurst launched, and again, I became excited about an alternative to Telkom. But it will all end in tears.
Now, I will make explain why they can't live up to expectations. There might be some small technical errors in my explanations below, but the general gist of it will be accurate.
The biggest problem is the fact that they cannot offer a committed international rate for cheaper than they can get it. This makes sense, they resell international bandwidth, only supplying the local loop infrastructure, and obviously, they cannot make a loss when reselling it. Therefore, there are two mechanisms to deal with it, one is to provide only a fraction of the international bandwidth to you at any one time, and the other is to put caps on the amount you can download. Sentech uses the first method, while iBurst uses both. The effect is that Sentech limits you to a fraction of your local loop capacity, while providing a consistent service – with the overall effect of being always slow. iBurst limits the amount of data you can download in order to balance their contention ratio’s.
The iBurst solution will not work in its current form, for a few reasons. The first is the fact that they give everybody their allocated block of data at the beginning of the month. The problem with this is, is that their international connection’s bandwidth should large enough to cater for peak network activity – distributing the allocation of the caps will allow them to cater for the average network activity (ignoring hourly fluctuations). The second problem is – I doubt if they will be able to supply a guaranteed 64kbps service after everybody’s cap takes effect. I am not aware of what the cost of a 45Mbps Telkom leased line connection to London cost (The UUNet international link) – but based on Telkom’s rates for smaller leased lines, it would be impossible to supply 64kbps CIR at a decent cost.
This would be the main problem – the cost of any broadband experience in South Africa is limited to the CIR rates on leased lines by Telkom. This place Telkom in a very good position: They can get CIR bandwidth for cheaper, and they can destroy competition by messing with prices of leased lines (legally, by only stopping to offer “discounts”).
This argument also destroys any argument for services that promise life after the cap: They can only resell bandwidth purchased from Telkom. The cost of their bandwidth would be higher than the cost for Sentech or iBurst, because these two companies get a discount for buying in bulk.
The only way these two broadband operators will be able to provide the services that the user expect, is for cheaper international bandwidth to become available – either through SNO or government regulation.
IMHO there are a few ways that they will be able to make their service better:
• Create “duty free” download sites, maybe with help of…
• ...create peering links to other networks in South Africa
• Distribute cap allocations over the month instead of giving it on a specific day of the month.
• Instead of caps, or plain user contention ratio’s, implement a system that weighs the priority of that user’s data by the amount that user has downloaded, on a month by month basis. This means users that download 24/7 would be penalized more, but not stopped dead in their tracks, and users that never use the system have blistering fast access, the average case might require balancing – this would be the most democratic way of using resources as limited by telkom. Obviously, P2P applications will be put in their place.
• If there really is a need for port shaping, do adaptive port shaping based on current international contention.
• “Duty-free” streaming from national sites like 5FM, TuksFM etc. (err.. Radio RSG anyone?)
There are of course a few ways the users can help:
• If at all possible, download from local sites, choosing the sites with the lowest ping time. This would make your download faster, save on international bandwidth.
• And this one I know I will get knocked for: Limit the use of peer to peer software. It eats up twice the international bandwidth. Remove files from the list to prevent uploads as soon as possible. Anyway, most of it are illegal anyway AND I HAVE NO SYMPATHY FOR PEOPLE WHO COMPLAIN about P2P. You are just like the guys who broke my car window to steal my radio – first you take something not belonging to you, but in the process cause a lot of additional damage as well! Movies are only R12/R14 in the cinemas okes and the economic damage caused to South Africa for downloading a single movie is greater than that!
Now iBurst launched, and again, I became excited about an alternative to Telkom. But it will all end in tears.
Now, I will make explain why they can't live up to expectations. There might be some small technical errors in my explanations below, but the general gist of it will be accurate.
The biggest problem is the fact that they cannot offer a committed international rate for cheaper than they can get it. This makes sense, they resell international bandwidth, only supplying the local loop infrastructure, and obviously, they cannot make a loss when reselling it. Therefore, there are two mechanisms to deal with it, one is to provide only a fraction of the international bandwidth to you at any one time, and the other is to put caps on the amount you can download. Sentech uses the first method, while iBurst uses both. The effect is that Sentech limits you to a fraction of your local loop capacity, while providing a consistent service – with the overall effect of being always slow. iBurst limits the amount of data you can download in order to balance their contention ratio’s.
The iBurst solution will not work in its current form, for a few reasons. The first is the fact that they give everybody their allocated block of data at the beginning of the month. The problem with this is, is that their international connection’s bandwidth should large enough to cater for peak network activity – distributing the allocation of the caps will allow them to cater for the average network activity (ignoring hourly fluctuations). The second problem is – I doubt if they will be able to supply a guaranteed 64kbps service after everybody’s cap takes effect. I am not aware of what the cost of a 45Mbps Telkom leased line connection to London cost (The UUNet international link) – but based on Telkom’s rates for smaller leased lines, it would be impossible to supply 64kbps CIR at a decent cost.
This would be the main problem – the cost of any broadband experience in South Africa is limited to the CIR rates on leased lines by Telkom. This place Telkom in a very good position: They can get CIR bandwidth for cheaper, and they can destroy competition by messing with prices of leased lines (legally, by only stopping to offer “discounts”).
This argument also destroys any argument for services that promise life after the cap: They can only resell bandwidth purchased from Telkom. The cost of their bandwidth would be higher than the cost for Sentech or iBurst, because these two companies get a discount for buying in bulk.
The only way these two broadband operators will be able to provide the services that the user expect, is for cheaper international bandwidth to become available – either through SNO or government regulation.
IMHO there are a few ways that they will be able to make their service better:
• Create “duty free” download sites, maybe with help of…
• ...create peering links to other networks in South Africa
• Distribute cap allocations over the month instead of giving it on a specific day of the month.
• Instead of caps, or plain user contention ratio’s, implement a system that weighs the priority of that user’s data by the amount that user has downloaded, on a month by month basis. This means users that download 24/7 would be penalized more, but not stopped dead in their tracks, and users that never use the system have blistering fast access, the average case might require balancing – this would be the most democratic way of using resources as limited by telkom. Obviously, P2P applications will be put in their place.
• If there really is a need for port shaping, do adaptive port shaping based on current international contention.
• “Duty-free” streaming from national sites like 5FM, TuksFM etc. (err.. Radio RSG anyone?)
There are of course a few ways the users can help:
• If at all possible, download from local sites, choosing the sites with the lowest ping time. This would make your download faster, save on international bandwidth.
• And this one I know I will get knocked for: Limit the use of peer to peer software. It eats up twice the international bandwidth. Remove files from the list to prevent uploads as soon as possible. Anyway, most of it are illegal anyway AND I HAVE NO SYMPATHY FOR PEOPLE WHO COMPLAIN about P2P. You are just like the guys who broke my car window to steal my radio – first you take something not belonging to you, but in the process cause a lot of additional damage as well! Movies are only R12/R14 in the cinemas okes and the economic damage caused to South Africa for downloading a single movie is greater than that!
Last edited: