Natas
Expert Member
So do think its time, or do I have to wait a while longer before I start phoning Telkom and telling them that either they stop charging me for the phone line AND my ADSL line OR i go over to another ADSL service provider!?!?!?!
-
Join the MyBroadband community
South Africa’s biggest forum. Discuss, discover, and connect with thousands of members.
-
Question of the Day: Worst movie or series plot hole?
Vodacom ADSL here in weeks
- Thread starter rpm
- Start date
DavidJ
Well-Known Member
I imagine that price is for the 'capacity' not the 'bandwidth' .. ?
Capacity and Bandwidth are the same thing. Data Usage, i.e. volumes of data transferred in a given time frame, is something entirely different.
HavocXphere
Honorary Master
Finally a company with some credibility announces a move against Telkom. I think we might see VC adsl before we see Neotel adsl.
ld13
Honorary Master
vodacom3g
Vodacom Representative
We tend to use both these words to describe different things, at different times, and it confuses many people, especially those who gave fire to the people.
A transmission link is spec'd and payed for in units of speed, say 155Mb/s.
Once you paid for this link, the provider does not care how you use (or not use it).
The amount of data you can transfer (measure in bytes) is a function of the speed x the time you have access to the link. So a 155Mb/s link can transfer around 1.6TB per day.
Where it gets confusing is when you don't buy an access 'speed' (in bits/s) from a provider but rather a certain amount of capacity (in bytes/month). This is the case in any 'capped' environment like we have locally.
So we buy in bits/second but sell it off in chunks of bytes/month. In theory this means we could sell a 1G bundle and provide an access speed of 3.5kb/s. This will still give you what you bought, i.e. 1GB of data over the month. Somehow I don't think people are going to be happy with this.
We have to provide you with your 'cap' but also must give you the ability to download it at a decent speed, WHENEVER you have the need. In theory this means that if you want to download your 1GB at 3.5Mb/s, we have to over-engineer the network by factor of 1000! Think what that will do to the costs.
It therefore becomes an balancing act where we have to find a sweet spot between having the right speed links to give a decent level of performance, if and when you need it, without making it so expensive no-one can afford it. This is called the contention ratio and although the industry standard used to be 50:1, Vodacom currently sits between 1:1 and 3:1, depending on your access device.
So when a company like Vodacom states they're "upgrading capacity", they're actually increasing the link speed, i.e. the bandwidth of the links measured in bits/second.
To answer DavidJ's question the price (673Mbps/month) is for the bandwidth. This translates in a monthly capacity of around 218 000 TB/month!
HavocXphere
Honorary Master
Cool info.vodacom3g said:
Just curious though: If you guys buy in bytes/second and the contention ratio is close to 1:1, then it would be possible to offer true uncapped @ a slow-ish speed without incurring additional costs on VCs side. (Ignoring the fact that new subscribes will then flood the network) Not so?Second Q: Is the 1:1 and 1:3 ratio you mentioned the contention ratio to intl access or to the network itself?
Roman4604
Executive Member
Hi V3G, I think it would be wise to clarify this statement as its quite easy to misconstrue.
For example you have 350,000 data cards in the market, and by product mix (e.g. 3G, HSDPA 1.8 & HSDPA 3.6) it works out that the average performance you're providing (promising) is say 700Kbps, then you would need 117 Gbps of capacity on SAT3 to claim that your provisioning your service/network at 2:1 end-to-end. Now to my knowledge only ~20Gbps of SAT3's capacity reaches SA.
See how misleading it can get.
vodacom3g
Vodacom Representative
It can, you're right. The important thing to remember is that mobile data users don't have a 24x7, max. bitrate profile (ic excluded). So the traditional calculations collapse. Also, I'm referring end-user experience and not a specific link.
I find it easier (from an end-user perspective) to look at it this way:
If I can get my maximum bit rate, at any time, then it stands I'm not contending with anyone else for MY bandwidth. Thus I'm getting my full service and my contention ration must, by definition, be 1:1. To me it looks like I'm the only person on the network.
If I'm on a 384/64 service (R99-3G) and I'm getting it (or damn close to it, if you take overheads into account, say 365/57), you're still on an effective 1:1 contention ratio or very close to it.
If I'm on a HSDPA 1.8 service and I get, say 1400/360, I'm on a 1:1.3 or better ratio.
And so on.
If this was a corporate system where you had to guarantee maximum bitrate (under SLA) at all times, you calculation would hold. But because of the nature of consumer, mobile data, the only way to calculate contention ratios is to actually measure it in real-time.
Of course my whole argument above is actually illegal as, according to the ASA, we're not allowed to make any speed claims. Thus whatever speed you're getting is OK and therefore your contention ratio (by ASA law) is always 1:1 !!
Last edited:
Roman4604
Executive Member
I understand what your getting at, but by strict definition (as accepted in the industry) contention ratio is ...
The amount of capacity sold (by virtue of product specification) vs. the amount of capacity actually provisioned (in the network)
... the capacity usage does not come into it.
The problem with your methodology is each provider could manipulate the figures (usage patterns) to suit their own needs, and therefore each could claim (Telkom etc.) a near 1:1 ratio for their circumstances.
Without a scientifically calculatable basis for determining contention, comparisons become largely irrelevant.
The amount of capacity sold (by virtue of product specification) vs. the amount of capacity actually provisioned (in the network)
... the capacity usage does not come into it.
The problem with your methodology is each provider could manipulate the figures (usage patterns) to suit their own needs, and therefore each could claim (Telkom etc.) a near 1:1 ratio for their circumstances.
Without a scientifically calculatable basis for determining contention, comparisons become largely irrelevant.
I think v3g was pointing out that ASASA's decision to disallow speed claims in adverts, effectively means that all of the broadband network operators can claim 1:1 by virtue of not having specified any speed in their adverts - if no speed is specified, it means that whatever speed you get, is by definition better than what was advertised - something like that...
PS: @v3g, what's this I hear about a maximum bitrate profile? - sounds like a conspiracy to get me to buy bigger data bundles
.
Last edited:
vodacom3g
Vodacom Representative
Yup, that's why the proof of the pudding is in the eating, i.e. what does an independent user get when using his service.
Now it's not a Vodacom, Telkom, MWeb or anyone else's number, it's Joe's number at that point in time when Joe was using the service. And it's the only number Joe cares about.
I agree capacity does not come into it, it's merely a function of bit-rate.
Pity ASA stopped us from making bit-rate claims. This effectively makes our contention ratio 0, or at least, tending to it!
Guess in the ASA view of the world we should calculate "capacity contention", i.e (total capacity sold as a function of the bundles) / (total capacity available on the network). But you'll agree it'll again be less than 1.
vodacom3g
Vodacom Representative
Roman4604
Executive Member
True, but thats not really the contention ratio and it cannot be used to make service comparisons, for example to say Telkom's specification of a 20:1 contention ratio on the access layer of their ADSL network (e.g. backhaul bandwidth from DSLAMs is 1/20th the sum of all customer port speeds provisioned).
The whole point of (quoted) contention ratios is so that Joe can get some indication of comparitive performance potential between differing 'best effort' services (on paper) without having to go throught the expense of physically trying them all.
vodacom3g
Vodacom Representative
Do you know if they quote the number on specific links, specific areas of the network (say local traffic) or user end-to-end experience, i.e from client to server?
In a know, fixed network I can see how this can work (using a comparative measurement), especially if it's a Diginet type of service. Or WiMax.
Question is, does it actually mean anything useful to the person doing the browsing? How would one measure end-to-end user experience in a wireless, mobile environment where the network itself changes dynamically?
Simplistically contention ratio tells Joe with how many others like him, he shares the network. 1:10 says he's got a 10% change of getting maximum throughput, or on average he'll only get 10% performance as he's sharing the network with 10 others. 1:1 means his seeing no 'contention' i.e. it appears as if he is alone.
I guess the question is if it's still a valid method of comparison between dissimilar networks?
I see an interesting discussion in the near future
. I give you a shout when I'm back in Jozy 
Not sure, but I suspect it's meant to be the ADSL contention ratio after DSLAM ports - presumably on the ATM backbone, although knowing Telkodemonopolies it could be the contention ratio on a 100Mbits/s ethernet cable 
basically Telkodemonopolies' ADSL contention ratio should not be believed anyway.
BTW, there is zero contention on the copper pairs between a customer's ADSL modem|router and the corresponding DSLAM Port Card.
basically Telkodemonopolies' ADSL contention ratio should not be believed anyway.BTW, there is zero contention on the copper pairs between a customer's ADSL modem|router and the corresponding DSLAM Port Card.
Last edited:
Vodacom will initially be reselling SAIX-based ADSL ISP accounts.