LUI(Leeds Unbundled ISP) - part I Backhaul
Whilst discussing life’s rich tapestry and the shortcomings of generic ISP economic models with my cyberbuddy, Jeremy Penston of IPDev.net, we rapidly came to the conclusion that what was really needed was a real life example and so the Leeds Unbundled ISP was born.
As well as being a wonderful place to live, Leeds has the distinct modelling advantage of being home to a manageable number of BT exchanges (34), a decent mix of residential homes (328k) and business premises (17k) and most importantly a dense urban sprawl with smaller conurbations on the outskirts of the city and good fibre connectivity with the rest of the universe. The ever useful Google Maps presents a geographical view of the city and the BT Exchanges.
The first design consideration for our ISP was that any BT exchange serving more than 5k homes would be unbundled – therefore we plan for only 21 out of the 34 BT exchanges to be unbundled. This straightforward size variable led us to unbundle Ilkey, Otley and Wetherby which although being a considerable distance from the centre of Leeds have the advantage of having high average household disposable income and very few unbundlers currently present.
Upon closer examination of the map it is pretty apparent that in phase II we would look to expansion into Bradford and Wakefield and some of the exchanges would probably cost us less than the outer regions of the Leeds Postal region. We also might consider daisy chaining some of the outlying exchanges.
Another design consideration was that we would use 1-gig links back to a central hub using the relatively new Openreach BNS1000 product. The Central Hub is the main Leeds Exchange which fortunately is only 430m from our main PoP which is situated next to the Train Station at Leeds where a surplus of fibre is currently available to connect with the main UK backbones.
The main link from the BT exchanges to our PoP is a redundant 2.5-gig link which ensures that this key link will keep our customers happy with its availability. This provides a Main Link Aggregation Factor of 8:1 from the 20 spoke exchanges with 1-gig links into the main BT exchange.
The maximum length of these spoke links is 22.7km out to Otley and we have 6 links under 5km with a minimum distance from Chapeltown at 2.5km. We have an average link length of 7.9km and a total radial distance of fibre used of 158.258km.
The price of backhaul is proportional to spoke bandwidth, spoke distance, hub distance from PoP and redundant bandwidth to PoP. In order to minimize prices we have taken the maximum contract length which is five years and leads to the following bills from Openreach:
It is also worth noting that our PoP is extremely close to the Leeds Exchange and this reduces our ongoing bill dramatically. A link of only 15km would push up the Annual Running Cost to £200k per annum.
All this bandwidth without a shedload of customers is pretty much a wasted resource and therefore we have carefully examined our demographics to provide a model of future demand. Although our ISP will undoubtedly be the best in the universe, we have tried to be realistic with our estimates and think we will attract users in proportion to the current number of unbundlers at each exchange + BT + Virgin Media.
Without going into the detail of the model we think that after a period of build-up we expect to finish with a residential market share of 16% of total homes or 47k customers. We have no plans to enter the business market although realize that is an upside to the model. In terms of per exchange data our share varies between 12.5% and 35%. The maximum number of homes who become customers at an individual exchange are 3.6k, the minimum are 1.3k and the average is 2.2k.
The important use of all this demand modelling is to calculate that the average capex per customer is £9.71 and opex per customer per month is £0.35. Of course our sugar daddy is the archetypical manic depressive financier and has insisted that we halve our forecasts and even with capex of £20 per customer and opex of £0.70 per month we think our backhaul costs are under control certainly compared to Openreach quotes of only 12 months ago.
Over the coming weeks, both blogs will be building on this initial model, hopefully with a little help from our friends, to provide insight to other key areas of ISP operations.
Over to you Jeremy…
As well as being a wonderful place to live, Leeds has the distinct modelling advantage of being home to a manageable number of BT exchanges (34), a decent mix of residential homes (328k) and business premises (17k) and most importantly a dense urban sprawl with smaller conurbations on the outskirts of the city and good fibre connectivity with the rest of the universe. The ever useful Google Maps presents a geographical view of the city and the BT Exchanges.
The first design consideration for our ISP was that any BT exchange serving more than 5k homes would be unbundled – therefore we plan for only 21 out of the 34 BT exchanges to be unbundled. This straightforward size variable led us to unbundle Ilkey, Otley and Wetherby which although being a considerable distance from the centre of Leeds have the advantage of having high average household disposable income and very few unbundlers currently present.
Upon closer examination of the map it is pretty apparent that in phase II we would look to expansion into Bradford and Wakefield and some of the exchanges would probably cost us less than the outer regions of the Leeds Postal region. We also might consider daisy chaining some of the outlying exchanges.
Another design consideration was that we would use 1-gig links back to a central hub using the relatively new Openreach BNS1000 product. The Central Hub is the main Leeds Exchange which fortunately is only 430m from our main PoP which is situated next to the Train Station at Leeds where a surplus of fibre is currently available to connect with the main UK backbones.
The main link from the BT exchanges to our PoP is a redundant 2.5-gig link which ensures that this key link will keep our customers happy with its availability. This provides a Main Link Aggregation Factor of 8:1 from the 20 spoke exchanges with 1-gig links into the main BT exchange.
The maximum length of these spoke links is 22.7km out to Otley and we have 6 links under 5km with a minimum distance from Chapeltown at 2.5km. We have an average link length of 7.9km and a total radial distance of fibre used of 158.258km.
The price of backhaul is proportional to spoke bandwidth, spoke distance, hub distance from PoP and redundant bandwidth to PoP. In order to minimize prices we have taken the maximum contract length which is five years and leads to the following bills from Openreach:
Whilst this may seem like a lot of money, it is a dramatic reduction from previous Openreach products and certainly should give us a competitive edge over some of our so-called competitors who unbundled early and alledgely have overloaded 100-meg links, especially at the Moortown exchange.
- Capex - £457,090
- Annual Running Cost - £167,791
It is also worth noting that our PoP is extremely close to the Leeds Exchange and this reduces our ongoing bill dramatically. A link of only 15km would push up the Annual Running Cost to £200k per annum.
All this bandwidth without a shedload of customers is pretty much a wasted resource and therefore we have carefully examined our demographics to provide a model of future demand. Although our ISP will undoubtedly be the best in the universe, we have tried to be realistic with our estimates and think we will attract users in proportion to the current number of unbundlers at each exchange + BT + Virgin Media.
Without going into the detail of the model we think that after a period of build-up we expect to finish with a residential market share of 16% of total homes or 47k customers. We have no plans to enter the business market although realize that is an upside to the model. In terms of per exchange data our share varies between 12.5% and 35%. The maximum number of homes who become customers at an individual exchange are 3.6k, the minimum are 1.3k and the average is 2.2k.
The important use of all this demand modelling is to calculate that the average capex per customer is £9.71 and opex per customer per month is £0.35. Of course our sugar daddy is the archetypical manic depressive financier and has insisted that we halve our forecasts and even with capex of £20 per customer and opex of £0.70 per month we think our backhaul costs are under control certainly compared to Openreach quotes of only 12 months ago.
Over the coming weeks, both blogs will be building on this initial model, hopefully with a little help from our friends, to provide insight to other key areas of ISP operations.
Over to you Jeremy…
<< Home