OpenSource ADSL in South Africa

brucedav

Active Member
Joined
Jan 8, 2014
Messages
32
Reaction score
0
Location
Cape Town
One of my projects for this year is to launch a new consumer ISP service in South Africa with an OpenSource ideology. What I mean by this is it'll be community supported and driven.

Our focus, offer the most competitive consumer adsl service in the country with the best connectivity. How? We'll keep our costs low by running the service as OpenSource wherever possible. We'll build a community of passionate and technically capable peeps who are willing to contribute their time to the project for the greater good of internet communication in South Africa and incentives such as reduced cost / free adsl based on the extent of their efforts.

Whats the project status currently? I've reached out to MTN and IS regarding possible shared use of their Telkom IPC circuit and Local / International Transits.

I'm also looking at the costs involved in getting a IP Connect circuit direct from Telkom.

What are your thoughts? Who's keen to join the team?
 
One of my projects for this year is to launch a new consumer ISP service in South Africa with an OpenSource ideology. What I mean by this is it'll be community supported and driven.

Our focus, offer the most competitive consumer adsl service in the country with the best connectivity. How? We'll keep our costs low by running the service as OpenSource wherever possible. We'll build a community of passionate and technically capable peeps who are willing to contribute their time to the project for the greater good of internet communication in South Africa and incentives such as reduced cost / free adsl based on the extent of their efforts.

Whats the project status currently? I've reached out to MTN and IS regarding possible shared use of their Telkom IPC circuit and Local / International Transits.

I'm also looking at the costs involved in getting a IP Connect circuit direct from Telkom.

What are your thoughts? Who's keen to join the team?

I've worked fairly high up in a 100% opensource ISP environment in this country, so I can offer a little bit of insight. You can certainly do routing/monitoring/authentication using opensource technology, but don't skimp on your issue tracking system. Customers need to know that you've got their issues handled in a clean and manageable manner. Right now, there is nothing openSource that does that properly. Yes, RT and Jira are good, but the closed source solutions are hundreds of times more efficient.

Otherwise, just bear in mind that support for OpenSource technologies is hard to come by, and so you'll need to spend a pretty penny on a team that knows routing and billing extremely well, and have been in the space before. Newbies are not going to cut it, OpenSource is hard when it comes to doing RCA. Tools are built by people who can track down the issues by hand, and do not come with much diagnostic abilities, as the Devs do not need frontend buttons in order to figure out what's going on.

I've got more thoughts on the matter, but I'll leave you with these two pointers to start with. Will watch the thread and pop in from time to time if it grows.

My thoughts? It's a big undertaking, and you'll find that at the beginning, you will probably have to run at a loss to compete effectively on bandwidth with the competition. Why? Because they get bulk bandwidth discounts, which at the start, you will be inelligable for. Your costs are going to exceed your income by a wide margin until you have purchasing power, or some sort of agreement with your upstream providers for a bandwidth discount.
 
Open source or community maintained and funded non profit ISP?
 
Open source or community maintained and funded non profit ISP?

OpenSource tech where possible and community maintained. It will be a for profit company however the profit margin would be clearly disclosed to the community. No secrets! The startup will be funded by myself.
 
I've worked fairly high up in a 100% opensource ISP environment in this country, so I can offer a little bit of insight. You can certainly do routing/monitoring/authentication using opensource technology, but don't skimp on your issue tracking system. Customers need to know that you've got their issues handled in a clean and manageable manner. Right now, there is nothing openSource that does that properly. Yes, RT and Jira are good, but the closed source solutions are hundreds of times more efficient.

Otherwise, just bear in mind that support for OpenSource technologies is hard to come by, and so you'll need to spend a pretty penny on a team that knows routing and billing extremely well, and have been in the space before. Newbies are not going to cut it, OpenSource is hard when it comes to doing RCA. Tools are built by people who can track down the issues by hand, and do not come with much diagnostic abilities, as the Devs do not need frontend buttons in order to figure out what's going on.

I've got more thoughts on the matter, but I'll leave you with these two pointers to start with. Will watch the thread and pop in from time to time if it grows.

My thoughts? It's a big undertaking, and you'll find that at the beginning, you will probably have to run at a loss to compete effectively on bandwidth with the competition. Why? Because they get bulk bandwidth discounts, which at the start, you will be inelligable for. Your costs are going to exceed your income by a wide margin until you have purchasing power, or some sort of agreement with your upstream providers for a bandwidth discount.

Thanks so much for the advise, :). I'm recently back in South Africa after a 7 year stint in the UK working as a technical consultant with two separate small ISP's so I'll be doing much of the initial technical configuration and 3rd line support myself with a view to allow trusted experienced community member to assist down the road.

Very good point re support. Completely agree its very important to get right. My thoughts are to mimic a firm in the UK who offer extremely competitive mobile services by relying on a trusted community to help them in support / ideas for future growth ect. Check out this excerpt from their website re support:

'We do things slightly differently; we don't have a call centre, we handle all of our support online so that we can save you money and because we think it's better. There are several ways of getting help on the website; searching the Knowledge Base is the fastest, followed by asking the community.'

I believe this support model can work with the right community ethos and clarity in communicating this style of support up front to new customers.
 
Example of how payback is likely to be structured for community members who help out, :) *Borrowed from a firm in the UK using a similar structure in the mobile space..

What's Payback?

It's simple. Recommend us to your friends and family or help your fellow members in our forums and we'll give you Payback points. Here are a few ways to earn points:

Recommend friends with your member name when they order an adsl package to earn x points
Help out giffgaff members in our forum. Points are added to your "My Payback" widget on the 5th of every following month.
Our Payback currency is pennies (1 point = 1 penny).

We collect all the points you earn and send monthly email statements. Twice a year (June and December) choose how to use your points:

Cashback via PayPal (only eligible if you've earned more than 1,000 points)
Data allowance credit
Donate it to charity
 
'We do things slightly differently; we don't have a call centre, we handle all of our support online so that we can save you money and because we think it's better. There are several ways of getting help on the website; searching the Knowledge Base is the fastest, followed by asking the community.'

Sounds great... but when there is a problem with your ISP caching DNS entries incorrectly, or I have a serious problem that I need an operator to assist with, nothing beats picking up the phone and calling someone. The time taken waiting for a possible response from the community would not be justified... not to mention that if my ADSL is not working... how would I access said community? From my phone? I have a feeling that 90% of the population would feel similarly.

With margins what they are on ADSL already, i'm not sure that you'd be able to compete on price without dropping man-power by a large margin... and if you pick up to the volumes of people that are required to negotiate volumes with suppliers, i'm pretty sure a large number of those people would be of the variety that appreciate someone on the end of a phone line. There's a small percentage of users that would be happy just with a knowledge base or forum.

I think it's an interesting idea for sure. Would be interested to see how you get through the market barriers.
 
I am unlikely to really use such a service - my ADSL needs pretty much well covered by Telkom and Afrihost

However I am very willing within my resources (time and energy really) prepared to chip in (join the team) if it is a genuine exercise in an internet community project - the startup funding and profit issue is quite murky in what is presented.
 
Sounds great... but when there is a problem with your ISP caching DNS entries incorrectly, or I have a serious problem that I need an operator to assist with, nothing beats picking up the phone and calling someone. The time taken waiting for a possible response from the community would not be justified... not to mention that if my ADSL is not working... how would I access said community? From my phone? I have a feeling that 90% of the population would feel similarly.

With margins what they are on ADSL already, i'm not sure that you'd be able to compete on price without dropping man-power by a large margin... and if you pick up to the volumes of people that are required to negotiate volumes with suppliers, i'm pretty sure a large number of those people would be of the variety that appreciate someone on the end of a phone line. There's a small percentage of users that would be happy just with a knowledge base or forum.

I think it's an interesting idea for sure. Would be interested to see how you get through the market barriers.

The support element is definitely going to be a challenge but I'm determined to get it right! I by no means want to sacrifice quality of service (support service) but rather focus a lot of attention on perfecting a model that requires less hired man power thereby reducing cost (for both us the ISP and the end user). Main focus would be to reduce the number of 'standard' support cases that in my experience take the most time on an ISP support desk. i.e. New installation assistance, router doa, dhcp issues ect. I'd still look to offer manned / employed support to deal with 3rd line 'real' issues that truly lie within the ISP infrastructure. Haven't quite figured out how but I do envisage giving some level of access (to be determined) to 'elite' proven tech community members on the core equipment to allow them assist in troubleshooting. Also as a way of improving visibility of our infrastructure in the spirit of community and no secrets.

All in all, as much as is possible I would like this to be a service for the people, managed by the people, loved by the people and grown by the people.
 
I am unlikely to really use such a service - my ADSL needs pretty much well covered by Telkom and Afrihost

However I am very willing within my resources (time and energy really) prepared to chip in (join the team) if it is a genuine exercise in an internet community project - the startup funding and profit issue is quite murky in what is presented.

Ahhh, let me clarify, :).. I'll fund the startup myself and to be clear I do look for it to be a for profit company. I've decided to start this project not out of a desire to make money however but rather because I'm passionate about the technology and improving connectivity in South Africa so the focus isnt insane profits but rather fun and the challenge of the task at hand.

It'd be great to have you as part of the team. Perhaps you could drop me a PM with some details around your expertise?
 
Ahhh, let me clarify, :).. I'll fund the startup myself and to be clear I do look for it to be a for profit company. I've decided to start this project not out of a desire to make money however but rather because I'm passionate about the technology and improving connectivity in South Africa so the focus isnt insane profits but rather fun and the challenge of the task at hand.

It'd be great to have you as part of the team. Perhaps you could drop me a PM with some details around your expertise?

That's the part that most people are going to have a problem with. Everyones keen to help out a free project, and to some extent a startup. Their limits to assistance stop pretty abruptly when the business starts making a profit, unless they are getting something out of it. How you tackle this part will be the most tricky part of the business plan. The ISP has to be for profit to some degree to pay for staff etc. How you keep things transparent is going to be key.

I hope you have thought this through very carefully.
 
That's the part that most people are going to have a problem with. Everyones keen to help out a free project, and to some extent a startup. Their limits to assistance stop pretty abruptly when the business starts making a profit, unless they are getting something out of it. How you tackle this part will be the most tricky part of the business plan. The ISP has to be for profit to some degree to pay for staff etc. How you keep things transparent is going to be key.

I hope you have thought this through very carefully.

Absolutely, :). The rewards for community members who assist would be in the form of data allowance on their personal service / discounts on their bill / and possibly cash payments (their choice of these reimbursement options). How much would be clearly defined. i.e. 500 points per support incident ect (Points will be clearly defined as a rand value. Something like 10cents per 1 point for example).

Another advantage for contributing community members will be that involvement gives them access to the platform (RO access to config initially I'm thinking, controlled through the likes of TACACS). Quite nice to have that level of visibility into the platform, :).
 
In meeting MTN today I'm hoping they will be open to providing a solution where we can share their IPC and Transit links while still giving us a level of access and control of the environment. i.e. our own authentication realm, control over routing, co-location space (to allow us the option of offering shared services to adsl users). And all this priced in such a way that allows us the flexibility to start small and grow as our user base increases. Big ask, Fingers crossed, :).. Meeting is at 12.
 
Absolutely, :). The rewards for community members who assist would be in the form of data allowance on their personal service / discounts on their bill / and possibly cash payments (their choice of these reimbursement options). How much would be clearly defined. i.e. 500 points per support incident ect (Points will be clearly defined as a rand value. Something like 10cents per 1 point for example).

Another advantage for contributing community members will be that involvement gives them access to the platform (RO access to config initially I'm thinking, controlled through the likes of TACACS). Quite nice to have that level of visibility into the platform, :).

Interesting concept. Wish you luck with it, and your meeting today.
 
anything to report back?

Meeting went really well in that they like the idea and are open and enthusiastic about providing the services we need to get started (sharing their IPC capacity, co-location, local and international transit). They've promised to come back to me by close of play tomorrow with preliminary pricing, O_O. Fingers crossed its competitive!
 
Top
Sign up to the MyBroadband newsletter
X