Telkom blasts MWEB’s transit cut decision

[boerseun]

Telkom: MWEB is trying to stop us F@cking the consumer; Bad MWEB!!!

 

[boerseun]

Telkom:"Our first responsibility is to ensure that we protect the quality and integrity of our paying customers,” read:" Our first responsibilty is that we contitnue to rape our paying customers!"

 

[Steelclaw]

Shame... poor BEE okes does not want to lose a source of revenue... they don't want to lose their padded salaries...

 

[Deckert]

It's so simple, all Telkom and MTN have to do is setup a link to JINX and problem solved.

MWeb already has massive links to JINX, as does MTN, IS, Neotel, Tenet and many other players. You only have to look at the stats page to see all the different players:

http://stats.jinx.net.za

The issue here is that, even with the JINX and CINX peering points, ISPs (like MTN) want to profit from peering by not allowing transit via JINX, but rather insisting on a private peering link. You'll also notice that SAIX/Telkom is glaringly absent from both CINX and JINX, because you have to be an ISPA member to peer there, and SAIX/Telkom has historically always refused to become an ISPA member. Or maybe they just want to force the issue of private peering.

--deckert

 

[Nickste]

I don't think you guys truly understand why the big ISP don't want to peer with the likes of MWEB, sure they charge for interconnect rates, but its simple, in the industry there would most likely be a standard rate for that (you cant sell for more than you buy unless you have all the chips), Telkom uses MTN's network and pays for its use. Then vice- verse, so at the end of the day companies only gain any "profit" if their own network gets used more than what it uses other companies networks.
...

Let me start by saying that I'm 100% behind Mweb and open peering.

However, I believe it's a waste of time to argue over the ethics of Telkom/MTN not peering for free - I'm sure by now we all know that ethics don't come into decisions made by these larger companies - profits are what drives everything. Instead, it is a lot more useful to try and determine how much Telkom/MTN stand to lose by supporting open peering with Mweb. If we can work this out, it will be a lot easier to predict whether Mweb's gamble will pay off.

So Telkom/MTN make money off the balance of traffic requested from within their network by Mweb, i.e. if Telkom customers request 1TB of data from Mweb servers and Mweb customers request 2TB of data from Telkom servers, then Mweb has to pay Telkom for 1TB of excess data. The question that we need answered is how much more data does Mweb request, and what does this cost Mweb (i.e. what do Telkom profit out of it).

Next, one has to identify the actual cost to Telkom of having to carry this balance of Mweb data. If infrastructure was infinitely scalable at no additional cost, it wouldn't cost Telkom anything. Unfortunately this isn't the case. The following example may be helpful:

Imagine you operate a wifi network from your house and this wifi network in turn connects to your neighbour's wifi network via a wireless link. Your neighbour happens to be a massive apartment block, with 100 people living in it. You operate a server on your wifi network, which distributes hi-definition movies (legally of course) to anyone on the network. If all of the 100 neighbours decide to watch movies on your server at the same time, this would place a heavy load on the wifi network and it would very quickly become so saturated that you wouldn't even be able to watch movies yourself, sitting at home on your own wifi network access point.

There are two ways in which you could deal with this problem. Firstly, you could upgrade your wifi infrastructure by purchasing a few more access points and connecting them up to your server via wired connections. This however, would require a fairly substantial capital investment that would offer no real benefit to you, other than being able to reach a larger number of users. What happens further if another apartment block were to want to connect their 100 users to your wifi network? If we were to apply the principles of open peering here, we would further say that you wouldn't be allowed to charge the apartment users for this cost of infrastructure.

Secondly, you could start to charge the apartment users for bandwidth used on your network. This way, you would be able to ensure that they are incentivised not to use too much data on your network - thereby resulting in you not having to shell out for more infrastructure - and would mean that you could use this additional cash for your own profit. This is effectively what Telkom/MTN have been doing.

I think the above example paints a clearer picture as to the situation that Telkom finds itself in. The only way I can see Mweb succeeding is if they are able to get more content hosted on their own network - thus resulting in Telkom/MTN users requesting more data from their network and in turn evening out the balance of bandwidth usage between the two parties.

Edit: Just found MTN's view backing this up from here: http://mybroadband.co.za/news/telecoms/16213-MWEBs-transit-cuts-storm-teacup-MTN-Business.html

According to Thompson, the information he has suggests that any traffic sharing between MTN Business and MWEB will be very skewed, which means that it will not be mutually beneficial to both parties to peer.

Thompson however added that if MWEB boosts the content on their network – which in turn would result in more equitable traffic flow – they will definitely look at free peering between the two ISPs."

 

[Gatecrasher]

There are two ways in which you could deal with this problem. Firstly, you could upgrade your wifi infrastructure by purchasing a few more access points and connecting them up to your server via wired connections. This however, would require a fairly substantial capital investment that would offer no real benefit to you, other than being able to reach a larger number of users. What happens further if another apartment block were to want to connect their 100 users to your wifi network? If we were to apply the principles of open peering here, we would further say that you wouldn't be allowed to charge the apartment users for this cost of infrastructure.

Secondly, you could start to charge the apartment users for bandwidth used on your network. This way, you would be able to ensure that they are incentivised not to use too much data on your network - thereby resulting in you not having to shell out for more infrastructure - and would mean that you could use this additional cash for your own profit. This is effectively what Telkom/MTN have been doing.

But neither Telkom nor MTN are streaming their own content. They have content providers already paying vast sums of money to them for hosting the content. These content providers rely totally on the custom of the "apartment dwellers" so to speak.

What Telkom and MTN are doing is double charging; leeching off the traffic that content providers and consumers have already paid for.

In Telkom's case is it even more diabolical given the triple charge of line rental, ADSL rental, and IP Connect, even before we get to the additional transit charges.

 

[Nickste]

Sure, fair point. I suppose the question then is whether this cost that the content providers pay is sufficient in covering the cost of skewed traffic levels created by transporting all this data to ISPs such as Mweb?

 

[Gatecrasher]

Sure, fair point. I suppose the question then is whether this cost that the content providers pay is sufficient in covering the cost of skewed traffic levels created by transporting all this data to ISPs such as Mweb?

And SA has some of the most expensive hosting charges in the world...

 

[davemc]

<snip>This however, would require a fairly substantial capital investment that would offer no real benefit to you, other than being able to reach a larger number of users.<snip>

Well, reaching a larger number of users IS A MASSIVE BENEFIT.
More users = more money for the movies that you rent = your business is growing.
Or, am I missing the bus by another beer?

 

[BeVonk!]

But what if - like in the case of Telkom - you gained much of your equipment for free (for all practical purposes) due to an unfair Government protection scheme? The other players (like MWeb) had no such protection and therefore had to buy everything and build a system fom scratch. Now the skewness is mirrored, not so?

 

[MickeyD]

But what if - like in the case of Telkom - you gained much of your equipment for free (for all practical purposes) due to an unfair Government protection scheme? The other players (like MWeb) had no such protection and therefore had to buy everything and build a system fom scratch. Now the skewness is mirrored, not so?

For free? Explain that one... (in the context of this thread).

 

[ibanez]

"not in the spirit of the internet community"? .. not in the spirit of hellkoms current scr3w the subscribers, business ethos ..

 

[smokey]

For free? Explain that one... (in the context of this thread).

I think he's referring to the fact that Telkom inherited the copper system built by the government pre-privatization. It didn't pay for it/make it with money made (soley) from operations.

 

[MickeyD]

I think he's referring to the fact that Telkom inherited the copper system built by the government pre-privatization. It didn't pay for it/make it with money made (soley) from operations.

Telkom purchased all the assets from the Government when they split from the Post Office in 1991 and also took over some of the PO's debt. They did not get it for free. Besides, all the assets relevant to this issue (peering) were definitely built post '91.

 

[Nickste]

Well, reaching a larger number of users IS A MASSIVE BENEFIT.
More users = more money for the movies that you rent = your business is growing.
Or, am I missing the bus by another beer?

I wasn't particularly clear about this in my example. Telkom doesn't own the content, their clients who host on their network do. So for their clients, there is obviously a benefit to being able to reach a wider audience (at a good speed), but there is very little incentive for Telkom, who own the infrastructure.

 

[simonbuerger]

I wasn't particularly clear about this in my example. Telkom doesn't own the content, their clients who host on their network do. So for their clients, there is obviously a benefit to being able to reach a wider audience (at a good speed), but there is very little incentive for Telkom, who own the infrastructure.

Ah but the clients PAY Telkom depending on how much traffic they get, no? If fewer people can reach their content (or it's at a slower speed) they'll pay less for a lower hosting package.

 

[Shayd]

Sure, fair point. I suppose the question then is whether this cost that the content providers pay is sufficient in covering the cost of skewed traffic levels created by transporting all this data to ISPs such as Mweb?

Once you have trenched fibre or have a microwave link into the INX's the cost of peering is basically the cost of a couple of routers a couple of staff and a couple of meters of UTP. This is already in place so the net cost is pure profit that is lost.

Ultimately this profit is lost anyhow as MWEB will not pay you, so what exactly are you holding out for? Oh wait, you can still bully the smaller ISP's into paying you to send data from one router to another inside a data centre.

Screw you Telkom, Mtn, Vodacom.

 

[antowan]

FFS TELKOM stop raping people for no good reason!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Telkom has also challenged MWEB’s decision to cut its paid peering links, saying that it is a pity that MWEB is choosing to let Telkom's and other South African Internet subscribers take the brunt. “MWEB’s unilateral measure is not in the spirit of the Internet community,” said Telkom.

“MWEB is deciding not to exchange traffic directly with Telkom’s internet network namely AS 5713. As a result, MWEB is portioning the Internet and disrupting the flow of traffic between MWEB and Telkom customers,” said Telkom.

“While this has a negative impact on some users of the Internet, this effect will be the result of MWEB’s decision and is unfortunately beyond Telkom’s control. Telkom does not believe that it is constructive to air a potential commercial dispute in the public domain or to negotiate through the media.”

This arrogance just pisses me off!!!!!!! :twisted:

 

[antowan]

“Telkom has an internationally benchmarked peering policy. The policy follows best practice principles that are applied by network providers in fully competitive, commercially functional and efficient Internet markets globally,” Telkom explained.

Translated as... "Why prosecute us... look there are plenty of other companies raping their customers with costs that have no reason to exist!!"

 

[Nickste]

Ah but the clients PAY Telkom depending on how much traffic they get, no? If fewer people can reach their content (or it's at a slower speed) they'll pay less for a lower hosting package.

Yeah, I hadn't thought of that (I wonder if Telkom has?). Obviously, however, the incentive is still there for Telkom to try and keep the majority of this traffic on their own network, as it costs them less because they don't have to maintain infrastructure to other networks.

Once you have trenched fibre or have a microwave link into the INX's the cost of peering is basically the cost of a couple of routers a couple of staff and a couple of meters of UTP. This is already in place so the net cost is pure profit that is lost.
...

No, I disagree. Telkom need the infrastructure at JINX/CINX, but also further down the line as that increase in traffic from Mweb moves deeper into their network. i.e. if you were downloading something off my ADSL line, and I were on a SAIX account, the traffic would move over my copper line (Telkom), into the exchange (Telkom), into the MetroEthernet backbone (Telkom), and across all the infrastructure that connects that to JINX/CINX. There is a lot more infrastructure that gets utilised that would be at the cost of Telkom.

I still think Telkom should bite the bullet and peer openly. I just don't see them doing so from a profit point of view.