-
Ask a question and get free supportUnlock new privileges and see hidden discussionsWin prizes and earn reputation
-
Question of the day ~ How often do you braai?
-
Afristay giveaway.Win a 2-night stay for two people at an establishment of your choice with R1,000 in spending money. Enter here.
Sign up with GoogleIs bandwidth expensive?
- Thread starter desraid
- Start date
BTTB
Executive Member
- Joined
- Feb 6, 2004
- Messages
- 8,195
Good question.
This has pondered many of the ADSL users on this forum. How can people using ISDN who consistently download during the call more time period and who download many gigabytes during a month. It's easy to download 2 gigs over a weekend period.
So how does Telkom equate this with the puny 3 gigs they offer ADSL users for a whole month. It becomes apparent that we are been ripped off by Telkom with their bogus excuses about expensive bandwidth. Well we knew that already. ISDN users don’t have port shaping either.
It's almost like Telkom expected abuse on ADSL, but rather than compromising the situation and targeting the abusers they went and throttled the whole product. Of course this is only one way they could have sorted abuse out they chose rather to blanket the whole ADSL pool with caps and port shaping. I’m afraid that Telkom have other reasons for doing this. Ask anyone with a business in SA and you will soon discover that plus minus 70% of Telkom's income comes from the businesses in SA and the state. Telkom is protecting its business market as if businesses would potentially migrate to ADSL like other businesses across the world, Telkom would be hit heavily by loss of income from the Diginet support base, which business people use. This won’t certainly hamper its 178% increase in earnings that they brag about.
In my experience in life, never brag Parliament asks questions about how well you are doing as you can bet your bottom dollar the tide will turn the other way.
<b><hr noshade size="1"></b><font size="2"><font color="red"><b>You can take Telkom out of the Post Office but you can't take the Post Office out of Telkom.</b></font id="red"></font id="size2">
This has pondered many of the ADSL users on this forum. How can people using ISDN who consistently download during the call more time period and who download many gigabytes during a month. It's easy to download 2 gigs over a weekend period.
So how does Telkom equate this with the puny 3 gigs they offer ADSL users for a whole month. It becomes apparent that we are been ripped off by Telkom with their bogus excuses about expensive bandwidth. Well we knew that already. ISDN users don’t have port shaping either.
It's almost like Telkom expected abuse on ADSL, but rather than compromising the situation and targeting the abusers they went and throttled the whole product. Of course this is only one way they could have sorted abuse out they chose rather to blanket the whole ADSL pool with caps and port shaping. I’m afraid that Telkom have other reasons for doing this. Ask anyone with a business in SA and you will soon discover that plus minus 70% of Telkom's income comes from the businesses in SA and the state. Telkom is protecting its business market as if businesses would potentially migrate to ADSL like other businesses across the world, Telkom would be hit heavily by loss of income from the Diginet support base, which business people use. This won’t certainly hamper its 178% increase in earnings that they brag about.
In my experience in life, never brag Parliament asks questions about how well you are doing as you can bet your bottom dollar the tide will turn the other way.
<b><hr noshade size="1"></b><font size="2"><font color="red"><b>You can take Telkom out of the Post Office but you can't take the Post Office out of Telkom.</b></font id="red"></font id="size2">
Usually, an ISP will not buy a line where they pay for bandwidth as a resource, for example, per megabyte, but instead, they will purchase a line where they pay a fixed rate for bandwidth, a high fixed rate.
This is why most South African ISPs offer absolutely horrid service at high prices. For some time, I performed system administration duties for a local ISP who wish remain nameless.
At the time I was working there, they were servicing about 600 dial-up customers, 100 ISDN customers and about 50 "broadband" customers. The "broadband" referred to illegal CAT5 connections between buildings in the CBD of the town where they were operating, serving about 20 customers with 10MBps or 100MBps ethernet links. The other 30 "broadband" customers were served by similarly illegal 802.11b wireless links provided by a "hop-along" ad-hoc network of access points with high-gain Yagi type directional antennae and some with high gain flat-panel "sector omni" antennae. This provided customers with 11MBps links, of course, shared with anyone using the same access point.
I'm very sure I need not explain the bandwidth and usage implications of servicing all the users on such a network. Yet, our intrepid ISP chose to serve all these users with only thirty phone lines for dial-up and ISDN and one continuous ethernet and wireless network. Bandwidth for the network was provided by one 128kbps line to a first tear ISP, luckily at 1:1 contention ratio, but still nowhere near enough.
If only dial-up and ISDN customers were being serviced, the ISP relied on a delicate balance created by the ISP's users who would normally only dial in for short periods of time. Also, there would rarely be more than 10 users dialed in. Most of these users would only check e-mail accounts so external bandwidth was not needed. Effectively, if only two or three of the customers where downloading at one time, the 128kbps would seem to be enough to service them.
This delicate balance was thrown off completely by the introduction of the "broadband" services. Suddenly, customers who had previously stayed away from such activities as downloading and forex trading, put off by the cost of the phone call, flocked to the "broadband" service, which was being sold at around R500 per month, relatively cheap in terms of what you would generally pay for permanent internet access.
The strain on the 128kbps line was incredible. It would be at virtually 100% usage for almost the entire day. Latency for real time applications shot up through the roof. Even ISDN users would have to contend with latencies of up to 5000ms to both local and overseas sites. Online gaming and forex trading are OUT at these latency levels. Throughput for any transfer would rarely be able to exceed about 14400bps, meaning you would download at a maximum of 1KB per second and surfing most of today's graphics intensive sites is absolutely impossible.
Even after designing an extremely sophisticated bandwidth management system and installing a transparent proxy with a stored object size limit of 700MB (meaning it would even store most ISOs) and 200GB of total storage space, there was absolutely nothing that could be done to stem the tsunami of bandwidth usage the new "broadband" services created for our much loved ISP.
It was at this point that I left the employee of said ISP. There wasn't much else to do since I installed the radical new firewalling and bandwidth management system that kept the network JUST within usable bounds, if the speed you get from a 14.4kbps modem can be considered usable and was thus being blamed by the management of said ISP for the bandwidth problems causing all their customers to want to leave.
They later upgraded to a 256kbps line, which provided just enough relief to keep most of their customers from leaving and changed from advertising "fast" internet to "permanent" internet, with each user being limited to the most miniscule portion of bandwidth.
They don't like the 256kbps set-up though, as the line is MUCH more expensive and this has caused a dramatic decline in their profit margin, since the "broadband" service that created the problem is being sold much too cheap to be commercially viable.
It would be evident from the description above that the massive over subscription rate of the small infrastructure, in terms of dial-up and ISDN users, was never really a problem, since these users didn't use the service enough, or simply didn't have the throughput capabilities to cause problems. As soon as the "broadband" services were added, everything went completely pear shaped, for everyone.
To get back to the topic, the point I am trying to make with my short memoir of ISP life, is that ISPs will not buy high throughput lines with bandwidth as a resource that has to be payed for by megabyte, but instead, will buy smaller lines with flat rate bandwidth. Then, reselling broadband service causes a problem since the flat rate lines can not effectively serve more than a handful of broadband customers.
Bandwidth is certainly not expensive for Telkom, as they own their infrastructure and the SAT3 cable. All of this infrastructure does cost money though and instead of seeing their portion of SAT3 as a long term investment, Telkom want to make money off it within a few years.
Also, Telkom's terrestrial infrastructure is not equipped to handle broadband internet access. Generally, SAIX PoPs are connected to the larger network backbone by only a 2MBps ATM line. While it is possible to lay 144MBps or 1GBps fiber connections over huge distances to service these PoPs, fiber is expensive and Telkom just won't make the kind of investment needed.
Consequently, Telkom's infrastructure, like that of our beloved ISP, is not set up to handle broadband internet services. Telkom will certainly not pay for bandwidth by megabytes, so ISDN users hauling huge amounts of traffic won't effect Telkom that much, since ISDN is a very slow connection, relatively speaking, lots of ISDN users hauling lots of traffic also won't do that much to the integrity of Telkom's infrastructure.
However, broadband users, left unchecked, on such bottlenecked infrastructure, can cause huge disruptions very easily. Imagine a few hundred ISDN users and about 50 Diginet users connected a SAIX PoP with only 2MBps of bandwidth to the backbone. Let's say very few of the ISDN users are connected at any one time and that most of the Diginet users only use the lines for e-mail and internet banking. Since there is no simultaneous heavy load on the network, bandwidth and latency will appear to be fantastic to all the intermittent users with slow connections.
Now, add 8 ADSL users to the PoP. Each user is limited by the DSLAM to 512kbps, but this still means that 4 users, hauling their entire throughput limit with massive, continuous transfers will completely exhaust the bandwidth of the PoP. With 8 ADSL users, each user only has to be pulling 256kbps of throughput (downloads of about 30KB/s) to completely saturate the PoP's line.
As most users purchasing broadband connections do so with the inclination that they are buying a divine right to download all the time, fast, at least four users continuously maxing out their lines on any given PoP is a real possibility. Eight users doing so is also a very realistic proposition.
In such a nightmare scenario, the ISDN and Diginet users on the PoP, who have guaranteed bandwidth and SLAs, would have almost no internet access, almost all the time, as the line from the PoP to the outside world is constantly being saturated by a small group of users with extremely fast "last mile" connections that allow them to easily take advantage of all the bandwidth the PoP has to offer.
For Telkom, this would be a real headache, as Diginet and ISDN, their prime money-making services, would become frustrating to use and users would start cancelling their contracts. This would mean a drop in profit margin and a likely end to the careers of the Telkom top management currently living such luxurious lives while turning a huge profit for their American masters, the share holders.
Hence, Telkom have introduced the 3GB volume cap. The perpose of the cap is extremely simple. It is designed as a deterrent to keep download-mad users at bay. After hitting 3GB, international downloading, which accounts for most downloading, becomes completely impossible, meaning the hungry ADSL user no longer poses a threat to the integrity of the PoP and the Diginet and ISDN customers.
This is also the reason that the new "unshaped" service completely cuts out after 4GB. The total loss of connection is an even greater deterrent to those users who insist on using KaZaa at the highest possible throughput and are willing to pay for this. If these users were left unchecked, the entire SAIX network would easily become saturated by P2P traffic and ONLY P2P traffic.
Hence, the cap is not there due to the "cost of bandwidth", but merely as a measure to protect Telkom from having to upgrade their infrastructure in order to maintain service to users with slower, less powerful "last mile" connections.
It is easy to deduce this, as all internet connections these days always seem to slow down dramatically during the first days of a new month and then return to relative normalcy for the rest of the month. This effect is the tidal wave caused by bandwidth-hungry users at the beginning of the month, sucking up their 3GB of bandwidth in days.
After all of the extremely active users have been rooted out by the cap, the fast parts of the infrastructure are left to the more conservative bandwidth users on broadband lines and to those users with smaller "last mile" connections.
Since all users left unchecked at this point are either conservative enough not to saturate the network, or, in the case of dial-up or ISDN users, do not have the kind of throughput capabilities that can cause saturation of the network, everything appears to be very fast to these users, even though it isn't.
Willie Viljoen
Web Developer
Adaptive Web Development
This is why most South African ISPs offer absolutely horrid service at high prices. For some time, I performed system administration duties for a local ISP who wish remain nameless.
At the time I was working there, they were servicing about 600 dial-up customers, 100 ISDN customers and about 50 "broadband" customers. The "broadband" referred to illegal CAT5 connections between buildings in the CBD of the town where they were operating, serving about 20 customers with 10MBps or 100MBps ethernet links. The other 30 "broadband" customers were served by similarly illegal 802.11b wireless links provided by a "hop-along" ad-hoc network of access points with high-gain Yagi type directional antennae and some with high gain flat-panel "sector omni" antennae. This provided customers with 11MBps links, of course, shared with anyone using the same access point.
I'm very sure I need not explain the bandwidth and usage implications of servicing all the users on such a network. Yet, our intrepid ISP chose to serve all these users with only thirty phone lines for dial-up and ISDN and one continuous ethernet and wireless network. Bandwidth for the network was provided by one 128kbps line to a first tear ISP, luckily at 1:1 contention ratio, but still nowhere near enough.
If only dial-up and ISDN customers were being serviced, the ISP relied on a delicate balance created by the ISP's users who would normally only dial in for short periods of time. Also, there would rarely be more than 10 users dialed in. Most of these users would only check e-mail accounts so external bandwidth was not needed. Effectively, if only two or three of the customers where downloading at one time, the 128kbps would seem to be enough to service them.
This delicate balance was thrown off completely by the introduction of the "broadband" services. Suddenly, customers who had previously stayed away from such activities as downloading and forex trading, put off by the cost of the phone call, flocked to the "broadband" service, which was being sold at around R500 per month, relatively cheap in terms of what you would generally pay for permanent internet access.
The strain on the 128kbps line was incredible. It would be at virtually 100% usage for almost the entire day. Latency for real time applications shot up through the roof. Even ISDN users would have to contend with latencies of up to 5000ms to both local and overseas sites. Online gaming and forex trading are OUT at these latency levels. Throughput for any transfer would rarely be able to exceed about 14400bps, meaning you would download at a maximum of 1KB per second and surfing most of today's graphics intensive sites is absolutely impossible.
Even after designing an extremely sophisticated bandwidth management system and installing a transparent proxy with a stored object size limit of 700MB (meaning it would even store most ISOs) and 200GB of total storage space, there was absolutely nothing that could be done to stem the tsunami of bandwidth usage the new "broadband" services created for our much loved ISP.
It was at this point that I left the employee of said ISP. There wasn't much else to do since I installed the radical new firewalling and bandwidth management system that kept the network JUST within usable bounds, if the speed you get from a 14.4kbps modem can be considered usable and was thus being blamed by the management of said ISP for the bandwidth problems causing all their customers to want to leave.
They later upgraded to a 256kbps line, which provided just enough relief to keep most of their customers from leaving and changed from advertising "fast" internet to "permanent" internet, with each user being limited to the most miniscule portion of bandwidth.
They don't like the 256kbps set-up though, as the line is MUCH more expensive and this has caused a dramatic decline in their profit margin, since the "broadband" service that created the problem is being sold much too cheap to be commercially viable.
It would be evident from the description above that the massive over subscription rate of the small infrastructure, in terms of dial-up and ISDN users, was never really a problem, since these users didn't use the service enough, or simply didn't have the throughput capabilities to cause problems. As soon as the "broadband" services were added, everything went completely pear shaped, for everyone.
To get back to the topic, the point I am trying to make with my short memoir of ISP life, is that ISPs will not buy high throughput lines with bandwidth as a resource that has to be payed for by megabyte, but instead, will buy smaller lines with flat rate bandwidth. Then, reselling broadband service causes a problem since the flat rate lines can not effectively serve more than a handful of broadband customers.
Bandwidth is certainly not expensive for Telkom, as they own their infrastructure and the SAT3 cable. All of this infrastructure does cost money though and instead of seeing their portion of SAT3 as a long term investment, Telkom want to make money off it within a few years.
Also, Telkom's terrestrial infrastructure is not equipped to handle broadband internet access. Generally, SAIX PoPs are connected to the larger network backbone by only a 2MBps ATM line. While it is possible to lay 144MBps or 1GBps fiber connections over huge distances to service these PoPs, fiber is expensive and Telkom just won't make the kind of investment needed.
Consequently, Telkom's infrastructure, like that of our beloved ISP, is not set up to handle broadband internet services. Telkom will certainly not pay for bandwidth by megabytes, so ISDN users hauling huge amounts of traffic won't effect Telkom that much, since ISDN is a very slow connection, relatively speaking, lots of ISDN users hauling lots of traffic also won't do that much to the integrity of Telkom's infrastructure.
However, broadband users, left unchecked, on such bottlenecked infrastructure, can cause huge disruptions very easily. Imagine a few hundred ISDN users and about 50 Diginet users connected a SAIX PoP with only 2MBps of bandwidth to the backbone. Let's say very few of the ISDN users are connected at any one time and that most of the Diginet users only use the lines for e-mail and internet banking. Since there is no simultaneous heavy load on the network, bandwidth and latency will appear to be fantastic to all the intermittent users with slow connections.
Now, add 8 ADSL users to the PoP. Each user is limited by the DSLAM to 512kbps, but this still means that 4 users, hauling their entire throughput limit with massive, continuous transfers will completely exhaust the bandwidth of the PoP. With 8 ADSL users, each user only has to be pulling 256kbps of throughput (downloads of about 30KB/s) to completely saturate the PoP's line.
As most users purchasing broadband connections do so with the inclination that they are buying a divine right to download all the time, fast, at least four users continuously maxing out their lines on any given PoP is a real possibility. Eight users doing so is also a very realistic proposition.
In such a nightmare scenario, the ISDN and Diginet users on the PoP, who have guaranteed bandwidth and SLAs, would have almost no internet access, almost all the time, as the line from the PoP to the outside world is constantly being saturated by a small group of users with extremely fast "last mile" connections that allow them to easily take advantage of all the bandwidth the PoP has to offer.
For Telkom, this would be a real headache, as Diginet and ISDN, their prime money-making services, would become frustrating to use and users would start cancelling their contracts. This would mean a drop in profit margin and a likely end to the careers of the Telkom top management currently living such luxurious lives while turning a huge profit for their American masters, the share holders.
Hence, Telkom have introduced the 3GB volume cap. The perpose of the cap is extremely simple. It is designed as a deterrent to keep download-mad users at bay. After hitting 3GB, international downloading, which accounts for most downloading, becomes completely impossible, meaning the hungry ADSL user no longer poses a threat to the integrity of the PoP and the Diginet and ISDN customers.
This is also the reason that the new "unshaped" service completely cuts out after 4GB. The total loss of connection is an even greater deterrent to those users who insist on using KaZaa at the highest possible throughput and are willing to pay for this. If these users were left unchecked, the entire SAIX network would easily become saturated by P2P traffic and ONLY P2P traffic.
Hence, the cap is not there due to the "cost of bandwidth", but merely as a measure to protect Telkom from having to upgrade their infrastructure in order to maintain service to users with slower, less powerful "last mile" connections.
It is easy to deduce this, as all internet connections these days always seem to slow down dramatically during the first days of a new month and then return to relative normalcy for the rest of the month. This effect is the tidal wave caused by bandwidth-hungry users at the beginning of the month, sucking up their 3GB of bandwidth in days.
After all of the extremely active users have been rooted out by the cap, the fast parts of the infrastructure are left to the more conservative bandwidth users on broadband lines and to those users with smaller "last mile" connections.
Since all users left unchecked at this point are either conservative enough not to saturate the network, or, in the case of dial-up or ISDN users, do not have the kind of throughput capabilities that can cause saturation of the network, everything appears to be very fast to these users, even though it isn't.
Willie Viljoen
Web Developer
Adaptive Web Development
Great stuff, Willie - at last a coherent and logical explanation of the technicalities behind the constrained service! I would venture to suggest, however, that equal import must be given to the revenue-generating aspects of the service - there is no question that Telkrap's decision-makers are more concerned with boosting revenue levels, as opposed to augmenting infrastructure to improve service levels, which is why they would offer insane packages like the 4Gb one. Like your un-named ISP experience, this is what happens when the bean-counters hold sway over an organisation...
Nice read, makes sense...
Of course the POPs in denser areas are equipped with much larger pipes than 2Mbps
The CAP would be more acceptable if international browsing speed was unaffected after the 3GB, ala everything else slows like Kazaa etc. but browsing the Net is still normal speed.
I download way over 6 gigs a month on my measly dial-up and my bill is never over R400 so for the moment, until someone implants a conscience into SBC management, I'm sticking to dial-up and InfinitCall. I would prefer not to have to though but I'm not spending R600 a month extra for a third of the amount of data.
<font color="navy"><font size="1"><b>Where others have progress, we have Telkom.</b>
Hellkom website - http://telkomsucks.0catch.com</font id="size1"></font id="navy">
Of course the POPs in denser areas are equipped with much larger pipes than 2Mbps
The CAP would be more acceptable if international browsing speed was unaffected after the 3GB, ala everything else slows like Kazaa etc. but browsing the Net is still normal speed.
I download way over 6 gigs a month on my measly dial-up and my bill is never over R400 so for the moment, until someone implants a conscience into SBC management, I'm sticking to dial-up and InfinitCall. I would prefer not to have to though but I'm not spending R600 a month extra for a third of the amount of data.
<font color="navy"><font size="1"><b>Where others have progress, we have Telkom.</b>
Hellkom website - http://telkomsucks.0catch.com</font id="size1"></font id="navy">
BTTB
Executive Member
- Joined
- Feb 6, 2004
- Messages
- 8,195
Great stuff their Podo.
Now if Telkom would explain it that way, even the hard nut cases on MyADSL would begin to understand the problem. The way you explain things definitely points to a more complicated problem that has nothing to do with expensive bandwidth that Telkom would like us to believe. In order for Telkom to give 24/7 full speed bandwidth they would clearly have to upgrade their network first.
Well considering they have bragged about making such great profits they should take the extra cash and get on with the job of providing a decent service and stop sidestepping the issues at hand.
<b><hr noshade size="1"></b><font size="2"><font color="red"><b>You can take Telkom out of the Post Office but you can't take the Post Office out of Telkom.</b></font id="red"></font id="size2">
Now if Telkom would explain it that way, even the hard nut cases on MyADSL would begin to understand the problem. The way you explain things definitely points to a more complicated problem that has nothing to do with expensive bandwidth that Telkom would like us to believe. In order for Telkom to give 24/7 full speed bandwidth they would clearly have to upgrade their network first.
Well considering they have bragged about making such great profits they should take the extra cash and get on with the job of providing a decent service and stop sidestepping the issues at hand.
<b><hr noshade size="1"></b><font size="2"><font color="red"><b>You can take Telkom out of the Post Office but you can't take the Post Office out of Telkom.</b></font id="red"></font id="size2">
Pr13St, I'm not a writer, not professionally anyway, but the kind complements from you all are always welcome. Thanks. [
]
As for solving the bandwidth problem, it's actually a very simple matter, Telkom need to upgrade the infrastructure. Considering the huge profits about which they bragged, I can't imagine how it could be impossible for them to embark on a national infrastructure improvement program. Such a program would cost a large amount of money, but is certainly not out of reach of Telkom.
The problem is that even a sober upgrade program, scaled to best serve each specific area and stretched over a time period of say, three years, according to my worst case estimates, would eat up two thirds of their annual net profit for the three years. SBC and Telkom's larger local shareholders are much too greedy to allow them to persue such a plan, even if they wanted to.
An annual net profit of R50 million would not be bad for a para-statal private company, but it is a far cry from the estimated R200 million they will be making this year. SBC and the shareholders would see a noticeable change in their dividends and questions would start to be asked.
Of course, in time, improved infrastructure would allow them to offer broadband service at improved prices and with greatly increased service levels. This, in turn, would attract new customers to the service, who are currently scared away by the high price and poor service quality levels.
A large new customer take-up would generate growth in the market that will eventually lead to a large increase in revenue for Telkom. Sadly, Telkom's strategic management do not believe in good long term strategy.
As a good example of this, I might point to the case of the new SAT3 cable. The costs of building the cable were incredibly high, as with all submarine cables. The costs for construction of the cable were carried by the consortium of African telephone operators along the West coast who collaborated on the project.
Most of the other operators have seen the project as a long term investment. The operators have set pricing for access to their portions of the cable at very reasonable rates. Access is resold to local telephone operators, where there are competing operators, cellular operators, internet service providers and anyone who might have a need for international frame relay services.
These operators do not envisage making a profit on their portion of the line for the first 10 to 15 years of use. With a design lifetime of 35 years, this means, the cable will likely only make a profit for these operators for 25 years.
The operators have kept pricing reasonable to attract customers to their line. Thus, while the line will take some time to pay for itself, they will have a huge customer base by the time it becomes profitable and will be able to make large long term profits from these customers.
Telkom have not chosen this route. Instead, pricing for access to Telkom's portion of SAT3 is unusually high, even by the worst international standards. By all accounts, the pricing has been set such that Telkom will start to turn a profit on their portion of SAT3 after only three years of use.
This is bad long term strategy. While their portion of the cable will have payed for itself very soon, the high pricing is keeping many potential customers away. While Telkom will be making a profit relatively quickly and will be able to make a profit out of the line for at least 30 years, the profits will remain small in comparison to those generated by the other operators over the envisaged 25 year period during which the line will be profitable for them.
Had Telkom set pricing for SAT3 access at reasonable rates, the market for the use of the cable may have been much larger. Currently, only Telkom itself and a small number of top tier ISPs make use of the South African portion of the SAT3 cable. Even the large ISPs have only limited use, as buying bandwidth on the cable is so expensive. With better prices, many smaller ISPs might have been enticed into taking up usage on SAT3. Larger ISPs might also have purchased more bandwidth than they currently use.
In time, the line would become profitable anyway. After that point, with a large customer base, Telkom would be able to generate very handsome revenues from the cable. As it stands now, they will probably turn a profit from the project as early as the end of this year, but all future profits will be much less than what might have been.
When examining this situation, it would seem to be very good short term strategy, as Telkom are turning a handsome profit for their shareholders and their American masters, SBC. Short term strategy is not good strategy though. In the long term, as possible competitors find their feet in the South African market, the current high pricing and bad service will immediately drive customers away from Telkom.
A mass customer exodus would have a very dramatic effect on Telkom's profit margin. I dare say, the effect will be much more dramatic than that of an infrastructure upgrade program. While SBC and Telkom's shareholder enjoy huge dividends from the enormous profits now, they will, in time, see a sudden, shocking decline.
This will be a bitter pill to swallow for the greedy investors, but it is the path that they and Telkom management have chosen. This massive strategic blunder also leaves Telkom itself in a very precarious position. If a competitor should enter the market, catching Telkom unprepared, the ensueing mass customer exodus would also mean a mass shareholder exodus.
This will leave Telkom without the support of share holder capital and without adequite infrastructure to compete in the open market.
As for MaD's suggestion that the cap should not effect international web surfing, but only traffic from third party applications, I think this is not only an excellent idea, but entirely feasible.
To implement it, Telkom would only need to allow ADSL users access to SAIX's main network of proxy caches and set up the proxy caches that serve the DSLAMs to use the main caches as parent systems. In this way, third party traffic can still be routed via the slow international connection, but web surfing traffic can pass through the proxies and be routed via the regular, faster connections.
I very much doubt that Telkom would go for such a set-up though, as it still leaves a door open to abusive users who might inundate the network and the proxies with large HTTP transfers, such as ISO image downloads, large pirated software downloads, pornography and streaming protocols that can be routed over HTTP.
Nevertheless, it is worth atleast investigating the idea and bringing it to Telkom's attention. To whom do we address the letter?
Willie Viljoen
Web Developer
Adaptive Web Development
]As for solving the bandwidth problem, it's actually a very simple matter, Telkom need to upgrade the infrastructure. Considering the huge profits about which they bragged, I can't imagine how it could be impossible for them to embark on a national infrastructure improvement program. Such a program would cost a large amount of money, but is certainly not out of reach of Telkom.
The problem is that even a sober upgrade program, scaled to best serve each specific area and stretched over a time period of say, three years, according to my worst case estimates, would eat up two thirds of their annual net profit for the three years. SBC and Telkom's larger local shareholders are much too greedy to allow them to persue such a plan, even if they wanted to.
An annual net profit of R50 million would not be bad for a para-statal private company, but it is a far cry from the estimated R200 million they will be making this year. SBC and the shareholders would see a noticeable change in their dividends and questions would start to be asked.
Of course, in time, improved infrastructure would allow them to offer broadband service at improved prices and with greatly increased service levels. This, in turn, would attract new customers to the service, who are currently scared away by the high price and poor service quality levels.
A large new customer take-up would generate growth in the market that will eventually lead to a large increase in revenue for Telkom. Sadly, Telkom's strategic management do not believe in good long term strategy.
As a good example of this, I might point to the case of the new SAT3 cable. The costs of building the cable were incredibly high, as with all submarine cables. The costs for construction of the cable were carried by the consortium of African telephone operators along the West coast who collaborated on the project.
Most of the other operators have seen the project as a long term investment. The operators have set pricing for access to their portions of the cable at very reasonable rates. Access is resold to local telephone operators, where there are competing operators, cellular operators, internet service providers and anyone who might have a need for international frame relay services.
These operators do not envisage making a profit on their portion of the line for the first 10 to 15 years of use. With a design lifetime of 35 years, this means, the cable will likely only make a profit for these operators for 25 years.
The operators have kept pricing reasonable to attract customers to their line. Thus, while the line will take some time to pay for itself, they will have a huge customer base by the time it becomes profitable and will be able to make large long term profits from these customers.
Telkom have not chosen this route. Instead, pricing for access to Telkom's portion of SAT3 is unusually high, even by the worst international standards. By all accounts, the pricing has been set such that Telkom will start to turn a profit on their portion of SAT3 after only three years of use.
This is bad long term strategy. While their portion of the cable will have payed for itself very soon, the high pricing is keeping many potential customers away. While Telkom will be making a profit relatively quickly and will be able to make a profit out of the line for at least 30 years, the profits will remain small in comparison to those generated by the other operators over the envisaged 25 year period during which the line will be profitable for them.
Had Telkom set pricing for SAT3 access at reasonable rates, the market for the use of the cable may have been much larger. Currently, only Telkom itself and a small number of top tier ISPs make use of the South African portion of the SAT3 cable. Even the large ISPs have only limited use, as buying bandwidth on the cable is so expensive. With better prices, many smaller ISPs might have been enticed into taking up usage on SAT3. Larger ISPs might also have purchased more bandwidth than they currently use.
In time, the line would become profitable anyway. After that point, with a large customer base, Telkom would be able to generate very handsome revenues from the cable. As it stands now, they will probably turn a profit from the project as early as the end of this year, but all future profits will be much less than what might have been.
When examining this situation, it would seem to be very good short term strategy, as Telkom are turning a handsome profit for their shareholders and their American masters, SBC. Short term strategy is not good strategy though. In the long term, as possible competitors find their feet in the South African market, the current high pricing and bad service will immediately drive customers away from Telkom.
A mass customer exodus would have a very dramatic effect on Telkom's profit margin. I dare say, the effect will be much more dramatic than that of an infrastructure upgrade program. While SBC and Telkom's shareholder enjoy huge dividends from the enormous profits now, they will, in time, see a sudden, shocking decline.
This will be a bitter pill to swallow for the greedy investors, but it is the path that they and Telkom management have chosen. This massive strategic blunder also leaves Telkom itself in a very precarious position. If a competitor should enter the market, catching Telkom unprepared, the ensueing mass customer exodus would also mean a mass shareholder exodus.
This will leave Telkom without the support of share holder capital and without adequite infrastructure to compete in the open market.
As for MaD's suggestion that the cap should not effect international web surfing, but only traffic from third party applications, I think this is not only an excellent idea, but entirely feasible.
To implement it, Telkom would only need to allow ADSL users access to SAIX's main network of proxy caches and set up the proxy caches that serve the DSLAMs to use the main caches as parent systems. In this way, third party traffic can still be routed via the slow international connection, but web surfing traffic can pass through the proxies and be routed via the regular, faster connections.
I very much doubt that Telkom would go for such a set-up though, as it still leaves a door open to abusive users who might inundate the network and the proxies with large HTTP transfers, such as ISO image downloads, large pirated software downloads, pornography and streaming protocols that can be routed over HTTP.
Nevertheless, it is worth atleast investigating the idea and bringing it to Telkom's attention. To whom do we address the letter?
Willie Viljoen
Web Developer
Adaptive Web Development
Two excellent posts podo! Thanks for the explanations.
Edit: Can a wireless infrastructure be implemented more easily and cheaper than laying down fibre? What would the effects on latency be? Dunno if you can answer this, but I just thought it would be interesting to know. Wireless just always seems to make more sense in Africa.
Edit: Can a wireless infrastructure be implemented more easily and cheaper than laying down fibre? What would the effects on latency be? Dunno if you can answer this, but I just thought it would be interesting to know. Wireless just always seems to make more sense in Africa.
BTTB
Executive Member
- Joined
- Feb 6, 2004
- Messages
- 8,195
I agree Wireless makes more sense in Africa, but can work both ways in South Africa, because the Telecommunication Infrastructure is far more advanced and their are dense areas of urbanisation for Antennas. However take that with a pinch of salt. As "advanced" is limited by Telkom's stranglehold over that infrastructure. I think the capacity for a larger ADSL support base is a more feasible option. I'm a bit sceptical of the Wireless Alternative at this stage. Perhaps it just needs to mature a bit.
But once again Podo, I take my hat off to you with your clear and understandable posts on this forum. You should join RPM at the ICASA hearings.
<b><hr noshade size="1"></b><font size="2"><font color="red"><b>You can take Telkom out of the Post Office but you can't take the Post Office out of Telkom.</b></font id="red"></font id="size2">
But once again Podo, I take my hat off to you with your clear and understandable posts on this forum. You should join RPM at the ICASA hearings.
<b><hr noshade size="1"></b><font size="2"><font color="red"><b>You can take Telkom out of the Post Office but you can't take the Post Office out of Telkom.</b></font id="red"></font id="size2">
martin,
Wireless is certainly cheaper than fiber optics, but does come with some disadvantages, especially in the case of a national phone operator.
The problem here is that in most cases, telecommunications infrastructure must not only carry data traffic, but also television signals, internal signaling and switching traffic and voice calls.
Building a communications infrastructure on wireless connections would mean that all of these services would need to be acomadated within the wireless service. This has great disadvantages, especially for analogue signalling technologies. Telkom already use wireless networking to bridge their infrastructure to areas where cable coverage is poor or cable theft is rampant.
In these areas, using an analogue modem or fax machine is virtually impossible. This is because the analogue signal must first be compressed into the relatively small bandwidth provided by the wireless communications links and then converted back from digital to analogue at the other end.
The drop off in signal quality is so severe that analogue modems and fax units generally can not traverse these connections at speeds of greater than 9600bps. Sometimes the maximum possible speed is as low as 2400bps.
Because fiber provides much better bandwidth than wireless connections, it can be used to build back-bone infrastructure. Voice calls or analogue data transmissions must still be digitised before being transmitted across a fiber link, but since very little or no compression is needed, the signal can be converted back to analogue at the other end with virtually no loss in signal quality.
Also, any wireless networks still need to be connected to the terrestrial network at some point in order to be effective. Usually, this still means running fiber or another form of terrestrial cable connectivity to the wireless access point. The technology does exist to allow wireless transmitters to "repeat" for each other, meaning the network only needs to be linked to the terrestrial infrastructure at a single point, but this has its own disadvantages.
Wireless networks need to transmit an immense amount of data at very high frequencies. When repeating for other transmitters over very large distances, wireless networking usually suffers as transmitters become tied up by the load caused by repeating for other transmitters and eventually the throughput and quality of service to other devices dependant to the transmitter begins to degrade.
One solution is to link each wireless access point to the terrestrial network at one point by first bridging the connection via satellite. However, this has huge disadvantages for both voice and data services on the telecommunications infrastructure.
Since satellites used for fixed communications are suspended in a very high geostationary orbit, approximately 36 000 km above the Earth's surface, any signal traveling via satellite will incur a huge delay upon itself.
This will mean massive latency for data applications and a noticeable delay period in voice calls. Leasing bandwidth and frequency space on satellites is also extremely expensive, mitigating the savings of installing wireless networks in place of terrestrial cable infrastructure.
Long range wireless links also suffer from unreliability due to atmospheric and space conditions. Changes in the ionisphere due to electrical storms, hail storms and dust storms can seriously degrade the propagation of high frequency radio signals over long distances.
Solar flares of high magnitudes cause massive geomagnetic storms which can also completely disrupt high frequency radio transmissions over any distance and put satellites out of commission for a day or more.
Obviously, this would cause massive disruptions to telecommunications infrastructure based entirely upon wireless connections. For this reason, wireless has generally been reserved as a last resort for areas where providing cable service is not possible.
Wireless networking also suffers from frequent packets loss due to multi-path attenuation and atmospheric conditions. Packet loss on a backbone network has an adverse effect on frame relay services and voice or analogue data transmissions traversing the backbone.
If some packets in a digital stream carrying a compressed analogue data transmission were to be lost, the analogue signal resulting at the other end would be full of incomplete modulations and the analogue devices on both ends would struggle to sustain a connection.
In terms of reliability and performance, terrestrial infrastructure based on fiber and copper still offers the best solution for building large telecommunications networks.
Wireless is certainly a viable alternative for last mile internet connectivity though, as repeating in this scenario generally isn't necessary and only data traffic needs to be carried.
Since the atmospheric and geomagnetic conditions that generally cripple long range wireless transmissions have a much less dramatic effect over small distances, wireless is also a relatively reliable last mile internet access solution.
Willie Viljoen
Web Developer
Adaptive Web Development
Wireless is certainly cheaper than fiber optics, but does come with some disadvantages, especially in the case of a national phone operator.
The problem here is that in most cases, telecommunications infrastructure must not only carry data traffic, but also television signals, internal signaling and switching traffic and voice calls.
Building a communications infrastructure on wireless connections would mean that all of these services would need to be acomadated within the wireless service. This has great disadvantages, especially for analogue signalling technologies. Telkom already use wireless networking to bridge their infrastructure to areas where cable coverage is poor or cable theft is rampant.
In these areas, using an analogue modem or fax machine is virtually impossible. This is because the analogue signal must first be compressed into the relatively small bandwidth provided by the wireless communications links and then converted back from digital to analogue at the other end.
The drop off in signal quality is so severe that analogue modems and fax units generally can not traverse these connections at speeds of greater than 9600bps. Sometimes the maximum possible speed is as low as 2400bps.
Because fiber provides much better bandwidth than wireless connections, it can be used to build back-bone infrastructure. Voice calls or analogue data transmissions must still be digitised before being transmitted across a fiber link, but since very little or no compression is needed, the signal can be converted back to analogue at the other end with virtually no loss in signal quality.
Also, any wireless networks still need to be connected to the terrestrial network at some point in order to be effective. Usually, this still means running fiber or another form of terrestrial cable connectivity to the wireless access point. The technology does exist to allow wireless transmitters to "repeat" for each other, meaning the network only needs to be linked to the terrestrial infrastructure at a single point, but this has its own disadvantages.
Wireless networks need to transmit an immense amount of data at very high frequencies. When repeating for other transmitters over very large distances, wireless networking usually suffers as transmitters become tied up by the load caused by repeating for other transmitters and eventually the throughput and quality of service to other devices dependant to the transmitter begins to degrade.
One solution is to link each wireless access point to the terrestrial network at one point by first bridging the connection via satellite. However, this has huge disadvantages for both voice and data services on the telecommunications infrastructure.
Since satellites used for fixed communications are suspended in a very high geostationary orbit, approximately 36 000 km above the Earth's surface, any signal traveling via satellite will incur a huge delay upon itself.
This will mean massive latency for data applications and a noticeable delay period in voice calls. Leasing bandwidth and frequency space on satellites is also extremely expensive, mitigating the savings of installing wireless networks in place of terrestrial cable infrastructure.
Long range wireless links also suffer from unreliability due to atmospheric and space conditions. Changes in the ionisphere due to electrical storms, hail storms and dust storms can seriously degrade the propagation of high frequency radio signals over long distances.
Solar flares of high magnitudes cause massive geomagnetic storms which can also completely disrupt high frequency radio transmissions over any distance and put satellites out of commission for a day or more.
Obviously, this would cause massive disruptions to telecommunications infrastructure based entirely upon wireless connections. For this reason, wireless has generally been reserved as a last resort for areas where providing cable service is not possible.
Wireless networking also suffers from frequent packets loss due to multi-path attenuation and atmospheric conditions. Packet loss on a backbone network has an adverse effect on frame relay services and voice or analogue data transmissions traversing the backbone.
If some packets in a digital stream carrying a compressed analogue data transmission were to be lost, the analogue signal resulting at the other end would be full of incomplete modulations and the analogue devices on both ends would struggle to sustain a connection.
In terms of reliability and performance, terrestrial infrastructure based on fiber and copper still offers the best solution for building large telecommunications networks.
Wireless is certainly a viable alternative for last mile internet connectivity though, as repeating in this scenario generally isn't necessary and only data traffic needs to be carried.
Since the atmospheric and geomagnetic conditions that generally cripple long range wireless transmissions have a much less dramatic effect over small distances, wireless is also a relatively reliable last mile internet access solution.
Willie Viljoen
Web Developer
Adaptive Web Development
BTTB,
I thank you for the kind complements, it means a lot to me that my posts are appreciated by other forum members.
Prior to reading your post, I had considered offering to join RPM at the hearings, even if only to serve as a second representative. I am a bit limited by a contractual obligation that I am tied in to at the moment, so any possible participation would depend on the time at which the hearings occur.
Also, I am not entirely sure where the hearing would happen, but I am fairly certain that I will need to travel as I live in Bloemfontein. Traveling to Cape Town or Pretoria is not such a problem, as I can do so by bus. Any other form of travel is out as I am not allowed to drive a car due to a visual impairment and airfares to and from Bloemfontein are incredibly high.
If the hearings are to occur at a time when I am able to travel and some of the forum members in the area might be able to help out with acomadation and transportation, I would most certainly be available to assist RPM at the hearings.
With your support, I will contact RPM and discuss the possibilities of this with him. I will welcome any help any forum members could provide in terms of acomadation and transportation. I will endeavor to cover the travel costs myself, as bus fares are generally quite reasonable.
Willie Viljoen
Web Developer
Adaptive Web Development
I thank you for the kind complements, it means a lot to me that my posts are appreciated by other forum members.
Prior to reading your post, I had considered offering to join RPM at the hearings, even if only to serve as a second representative. I am a bit limited by a contractual obligation that I am tied in to at the moment, so any possible participation would depend on the time at which the hearings occur.
Also, I am not entirely sure where the hearing would happen, but I am fairly certain that I will need to travel as I live in Bloemfontein. Traveling to Cape Town or Pretoria is not such a problem, as I can do so by bus. Any other form of travel is out as I am not allowed to drive a car due to a visual impairment and airfares to and from Bloemfontein are incredibly high.
If the hearings are to occur at a time when I am able to travel and some of the forum members in the area might be able to help out with acomadation and transportation, I would most certainly be available to assist RPM at the hearings.
With your support, I will contact RPM and discuss the possibilities of this with him. I will welcome any help any forum members could provide in terms of acomadation and transportation. I will endeavor to cover the travel costs myself, as bus fares are generally quite reasonable.
Willie Viljoen
Web Developer
Adaptive Web Development
Exellent posts Podo, you have answered many of my questions. I still think that true broadband means having a 24/7 connection at the stated speeds, with no port shaping (or at least that's what I'm used to in Europe) -
"I very much doubt that Telkom would go for such a set-up though, as it still leaves a door open to abusive users who might inundate the network and the proxies with large HTTP transfers, such as ISO image downloads, large pirated software downloads, pornography and streaming protocols that can be routed over HTTP.
"
However I don't think it's anyones' place to judge who are ''abusive users' or not whether or not they are downloading porn or doing their homework - future applicaiions are going to be ever more bandwidth hungry and if ZA is to progress at all we'll need al the infrastrtucture / banfdwidth we can get laid on.
btw - Do you think the proposed east coast cable will have a signifiicant effect on the available bandwidth ?
"I very much doubt that Telkom would go for such a set-up though, as it still leaves a door open to abusive users who might inundate the network and the proxies with large HTTP transfers, such as ISO image downloads, large pirated software downloads, pornography and streaming protocols that can be routed over HTTP.
"
However I don't think it's anyones' place to judge who are ''abusive users' or not whether or not they are downloading porn or doing their homework - future applicaiions are going to be ever more bandwidth hungry and if ZA is to progress at all we'll need al the infrastrtucture / banfdwidth we can get laid on.
btw - Do you think the proposed east coast cable will have a signifiicant effect on the available bandwidth ?
I'm not saying anyone using the bandwidth are abusive users. Certainly, if you are paying for the service, it's your own business what you download.
Perhaps I should clarify, when I refer to "abusive users" above, I refer to users who commit what would be considered about from Telkom's viewpoint. For instance, if the proxies were to be left open to capped users, as a way to allow them to still enjoy web surfing regardless of the cap, you might find that some users will use proxy services that allow the tunneling of other protocols over HTTP.
This would create a situation where extremely bandwidth hungry users could misuse the back channel to web surfing provided by the proxies and effectively evade the cap. In the case of extremely greedy users, this might even lead to an excess of stress on the proxies which would spoil the experience for everyone.
In truth, the only real solution to this problem is for Telkom to upgrade its infrastructure. There's just no other way around this problem. With improved infrastructure, bandwidth would become more affordable to ISPs and Telkom could also offer a better service, without the need for capping and with reduced port shaping.
There is currently a possibility that I might join RPM at the public hearing into the ICASA complaint, if and when it takes place. I have contacted him and offered my assistance, I will update you all on this when I hear from him.
If I do get the opportunity to help RPM in representing the forum at the hearing, I will be sure to raise these points with the regulators and media as far as it is possible, to get exposure for the problem to the media and the public.
As for the proposed East coast cable, I doubt very much if it would have any real effect on the available bandwidth in South Africa. In fact, if not correctly handled it might act as a contributing factor to our bandwidth problems.
Presently, the proposed cable will run up the East coast and have endpoints connecting to coastal states. It will also offer a connection to the Middle East.
Sadly, not much international bandwidth exists to the Middle East at the present time, with few exceptions. That being the case, it is most likely that operators connecting to the cable will be buying their international bandwidth from Telkom, to be routed via the SAT3/WASC/SAFE cable system.
This would, in effect, mean that East African countries and possibly even some Middle Eastern countries would be piggy-backing on us for bandwidth.
In my opinion, the proposed cable does present an opportunity for growth in the continent and is generally a very good idea, however, I can see no immediate benifit to the South African consumer or business markets from the construction of the cable. I may be wrong though, we will have to wait and see.
Willie Viljoen
Web Developer
Adaptive Web Development
Perhaps I should clarify, when I refer to "abusive users" above, I refer to users who commit what would be considered about from Telkom's viewpoint. For instance, if the proxies were to be left open to capped users, as a way to allow them to still enjoy web surfing regardless of the cap, you might find that some users will use proxy services that allow the tunneling of other protocols over HTTP.
This would create a situation where extremely bandwidth hungry users could misuse the back channel to web surfing provided by the proxies and effectively evade the cap. In the case of extremely greedy users, this might even lead to an excess of stress on the proxies which would spoil the experience for everyone.
In truth, the only real solution to this problem is for Telkom to upgrade its infrastructure. There's just no other way around this problem. With improved infrastructure, bandwidth would become more affordable to ISPs and Telkom could also offer a better service, without the need for capping and with reduced port shaping.
There is currently a possibility that I might join RPM at the public hearing into the ICASA complaint, if and when it takes place. I have contacted him and offered my assistance, I will update you all on this when I hear from him.
If I do get the opportunity to help RPM in representing the forum at the hearing, I will be sure to raise these points with the regulators and media as far as it is possible, to get exposure for the problem to the media and the public.
As for the proposed East coast cable, I doubt very much if it would have any real effect on the available bandwidth in South Africa. In fact, if not correctly handled it might act as a contributing factor to our bandwidth problems.
Presently, the proposed cable will run up the East coast and have endpoints connecting to coastal states. It will also offer a connection to the Middle East.
Sadly, not much international bandwidth exists to the Middle East at the present time, with few exceptions. That being the case, it is most likely that operators connecting to the cable will be buying their international bandwidth from Telkom, to be routed via the SAT3/WASC/SAFE cable system.
This would, in effect, mean that East African countries and possibly even some Middle Eastern countries would be piggy-backing on us for bandwidth.
In my opinion, the proposed cable does present an opportunity for growth in the continent and is generally a very good idea, however, I can see no immediate benifit to the South African consumer or business markets from the construction of the cable. I may be wrong though, we will have to wait and see.
Willie Viljoen
Web Developer
Adaptive Web Development
I read in one of the articles questioning Steve White that he said during the trial user abused the service so much that they decided the cap was unavoidable. IMO if everyone had chilled out during the trial period we would have a higher cap. A cap is acceptable only if it is reasonable - and these days 3GB is very much inadequate... 1GB per day would be more than enough for anyone.
Of course downloading movies and mp3's all day is just mindless, a waste of time and technology, not to mention bandwidth. Greediness has cost us a 3rd rate service.
One fellow apparently downloaded a record 90GB in under a month during the trial.
<font color="navy"><font size="1"><b>Where others have progress, we have Telkom.</b>
Hellkom website - http://telkomsucks.0catch.com</font id="size1"></font id="navy">
Of course downloading movies and mp3's all day is just mindless, a waste of time and technology, not to mention bandwidth. Greediness has cost us a 3rd rate service.
One fellow apparently downloaded a record 90GB in under a month during the trial.
<font color="navy"><font size="1"><b>Where others have progress, we have Telkom.</b>
Hellkom website - http://telkomsucks.0catch.com</font id="size1"></font id="navy">
You can't exactly blame the trial users though - who would pass up the oppurtunity to download all the stuff they've never been able to before? 90GB is quite excessive, but I don't think it's fair to blame the trial users just because they did what most people would probably do. They weren't to know that Telkom would impose a cap, after all.
1GB a day is perhaps asking a bit too much, but even if the cap was doubled or something like that it would be more adequate, and would at least allow people to do the downloading they want without having to worry so much about hitting the cap.
1GB a day is perhaps asking a bit too much, but even if the cap was doubled or something like that it would be more adequate, and would at least allow people to do the downloading they want without having to worry so much about hitting the cap.
TheVoice,
That is a very valid suggestion, but as is the nature of most users, it would most probably cause problems under the current infrastructure conditions.
At the present time, there is a noticeable slowdown to all South African internet users near the beginning of the month, the tsunami of usage I described in my earlier post to this thread. This effect is caused by greedy users who simply will not listen. These users take advantage of the first 3GB by going all out and downloading immense volumes of data.
The mass charge on the bandwidth has an effect on other users that very clearly degrades the service and causes a dramatic slowdown in international traffic, even for uncapped users.
Once these extremely greedy users hit their cap, the network seems to normalise somewhat, at least to the point where speeds are mostly acceptable to conservative users.
With an increased cap, these greedy users would cause even more disruptions and slowdowns as the effect of mass downloading near the beginning of the month will certainly become more dramatic if the users are left unchecked for longer periods of time.
A higher cap would also attract more greedy users who have previously been scared away from ADSL by the cap. These users would further contribute to network saturation, spoiling the experience for everyone else.
With the introduction of the "unshaped" service, the effect was much more dramatic this month than it had been in previous months. Users hit their 4GB cap quite quickly, but before getting there, the network was hit by a huge slowdown due to P2P traffic saturating the slow links holding the infrastructure together.
There is only one way to solve this problem in a manner that would be acceptable to all users. Telkom must upgrade their infrastructure. On the present infrastructure, an increased cap or no cap at all would almost certainly mean a disaster in terms of service quality, comparable to what Sentech MyWireless users have been forced to put up with in the past few weeks.
In my opinion, campaining for a less strict cap or no cap at all, without upgrading the infrastructure, would be shooting ourselves in the foot.
Willie Viljoen
Web Developer
Adaptive Web Development
That is a very valid suggestion, but as is the nature of most users, it would most probably cause problems under the current infrastructure conditions.
At the present time, there is a noticeable slowdown to all South African internet users near the beginning of the month, the tsunami of usage I described in my earlier post to this thread. This effect is caused by greedy users who simply will not listen. These users take advantage of the first 3GB by going all out and downloading immense volumes of data.
The mass charge on the bandwidth has an effect on other users that very clearly degrades the service and causes a dramatic slowdown in international traffic, even for uncapped users.
Once these extremely greedy users hit their cap, the network seems to normalise somewhat, at least to the point where speeds are mostly acceptable to conservative users.
With an increased cap, these greedy users would cause even more disruptions and slowdowns as the effect of mass downloading near the beginning of the month will certainly become more dramatic if the users are left unchecked for longer periods of time.
A higher cap would also attract more greedy users who have previously been scared away from ADSL by the cap. These users would further contribute to network saturation, spoiling the experience for everyone else.
With the introduction of the "unshaped" service, the effect was much more dramatic this month than it had been in previous months. Users hit their 4GB cap quite quickly, but before getting there, the network was hit by a huge slowdown due to P2P traffic saturating the slow links holding the infrastructure together.
There is only one way to solve this problem in a manner that would be acceptable to all users. Telkom must upgrade their infrastructure. On the present infrastructure, an increased cap or no cap at all would almost certainly mean a disaster in terms of service quality, comparable to what Sentech MyWireless users have been forced to put up with in the past few weeks.
In my opinion, campaining for a less strict cap or no cap at all, without upgrading the infrastructure, would be shooting ourselves in the foot.
Willie Viljoen
Web Developer
Adaptive Web Development
podo, thank you for your effort in writing your great articles on the thread. They are highly valuable and it made a sense in understanding from the technical point of view!
In my own opinion, Telkom is trying their best effort in earning more revenues from every opportunies before their 'monopoly license' expires.
Most users would probably expecting to see a few more broadband companies to appear in South Africa to create some competition for Telkom which will make Telkom to consider upgrading their infrastructure and lower down their cost and to improve their service quality which their customers will be receiving the benefits from it. Yes, it will be good for both Telkom and their customers!
No one really can't say that Sentech has brought a competition for Telkom, because Telkom has such huge user base of Diginet ISDN, and ADSL. But I believe Telkom received some stimulation from them, even it is very small. I noticed that Telkom had released their new 4GB ADSL product (even it's not worth), and if a person signs a 24-months ADSL contract (Standard 3GB cap product), that person receives FREE ADSL MODEM (Telkom claims that it is valued at R1500).
It is quite funny that '24-months ADSL contract' product has the the same usage cap and port shaping as the ADSL product that were released during the 'Commercial Beta Testing Phase in Gauteng'. How could they combine their 'beta products' with the 24-months contract?
Sentech also gives a modem away with their 24-months MyWireless contract. But, Sentech has a option that if you purchase the modem upfront, you can get reduced monthly subscription. Which I think is bit fair in terms of the cost. Would Telkom provide reduced cost to an ADSL subscriber who owns the ADSL modem?
Finally, I think 'Telkom ADSL' needs so much improvements (as mentioned by podo) to be made to become 'true ADSL'. Telkom ADSL just don't meet the standard.
In my own opinion, Telkom is trying their best effort in earning more revenues from every opportunies before their 'monopoly license' expires.
Most users would probably expecting to see a few more broadband companies to appear in South Africa to create some competition for Telkom which will make Telkom to consider upgrading their infrastructure and lower down their cost and to improve their service quality which their customers will be receiving the benefits from it. Yes, it will be good for both Telkom and their customers!
No one really can't say that Sentech has brought a competition for Telkom, because Telkom has such huge user base of Diginet ISDN, and ADSL. But I believe Telkom received some stimulation from them, even it is very small. I noticed that Telkom had released their new 4GB ADSL product (even it's not worth), and if a person signs a 24-months ADSL contract (Standard 3GB cap product), that person receives FREE ADSL MODEM (Telkom claims that it is valued at R1500).
It is quite funny that '24-months ADSL contract' product has the the same usage cap and port shaping as the ADSL product that were released during the 'Commercial Beta Testing Phase in Gauteng'. How could they combine their 'beta products' with the 24-months contract?
Sentech also gives a modem away with their 24-months MyWireless contract. But, Sentech has a option that if you purchase the modem upfront, you can get reduced monthly subscription. Which I think is bit fair in terms of the cost. Would Telkom provide reduced cost to an ADSL subscriber who owns the ADSL modem?
Finally, I think 'Telkom ADSL' needs so much improvements (as mentioned by podo) to be made to become 'true ADSL'. Telkom ADSL just don't meet the standard.
As of March 31, 2003, Telkom had 197 switches in the asynchronous transfer mode network. The present available bandwidth between the core switches on the asynchronous transfer mode network is 76 STM-1s or 11.78 Gbit/s, while the available bandwidth between the core switches and the services access switches is 420 STM-1s or 65.1 Gbit/s.
Access to the asynchronous transfer mode network is primarily provided via copper wire.
An STM-1 is also referred to as a OC-3 and runs at 155Mbps. You also get STM-0, 4, 16 and 64 so this is at the bottom rung.
<font color="navy"><font size="1"><b>Where others have progress, we have Telkom.</b>
Hellkom website - http://telkomsucks.0catch.com</font id="size1"></font id="navy">
Access to the asynchronous transfer mode network is primarily provided via copper wire.
An STM-1 is also referred to as a OC-3 and runs at 155Mbps. You also get STM-0, 4, 16 and 64 so this is at the bottom rung.
<font color="navy"><font size="1"><b>Where others have progress, we have Telkom.</b>
Hellkom website - http://telkomsucks.0catch.com</font id="size1"></font id="navy">