Friday, February 11, 2011


Upgrading infrastructure

Neelie Kroes, Vice-President of the European Commission responsible for the Digital Agenda, recently called for more action to deliver access to superfast broadband in pursuit of that Agenda.

Bemoaning the decline in rate of new connections she said that "Maximising broadband access obviously takes a mix of players and a mix of technologies, and the role of fibre is central."

It is good to see both mix of players and mix of technologies given prominence. The UK's current public sector broadband procurement process seems to be working against the former, and if your tender results in a choice between hammer providers you're going to get a solution involving only nails which works against the latter.

Personally I would like to see more diversity at the heart the BDUK pilots and funded projects. It should be a specific objective to try at least 5 different approaches in both technology and delivery organisation - otherwise BDUK and the County Councils are left trying to predict the best medium to long term solution without giving the other solutions a decent run for the money.

It would also be good to see more tolerance and acceptance in the debate around broadband delivery. If fibre to the cabinet is adequate for 10 years in appropriate areas then should we not welcome the provision of extra bandwidth as part of Neelie's Digital Agenda and the economic hopes riding on it ? Why do we spend so much time decrying solutions other than our own favourite - a sales tactic I have never liked. Can we not celebrate any increase in bandwidth or coverage if it helps someone in some way, especially if it pays its own way ?

The last twenty years has seen bandwidth grow in steps by changing the technology at each end of existing copper infrastructure. From 9600 baud modems through to 56k modems, then ADSL from exchanges, ADSL2+ and now VDSL2 from cabinets. All have been relatively low cost stepwise upgrades that have seen us get 18 million broadband users in the UK. The money spent on this has not been wasted, and investing in FTTC is not a waste of money if it pays off the investor within the useful life of the assets.

If universal FTTH will cost £5, 10 or £30 billion then the interest we save by deferring that investment by a year is a very substantial amount of money. So it makes sense to me to upgrade the infrastructure as and when it ceases to deliver adequate services to a location, and not get in a huge rush to spend money we don't have to meet a demand that maybe isn't actually here yet.

For some locations the time for an upgrade was 8 years ago. 10km of twisted pair copper isn't going to deliver an ADSL broadband service of any speed so there is a case to re-engineer such locations with either fibre optic connections, wireless or other solution that provides the needs of the user economically. These locations may be predominantly "rural"but some will be in built up areas and "electrically remote from the exchange" is probably the guiding principle.

If we draw an 800m radius around a BT cabinet the area covered is 2 which with 80,000 cabinets gives us a coverage of 160,000 That's 66% of the UK land area, but only if we're lucky to have no overlaps and an even distribution of cabinets across the land - which clearly isn't the case. We are still left with at least a "final third" that is out of range of the "cabinet circles", plus a good number of customers that aren't connected to a cabinet at all. Even adding in VDSL2 from the exchange only adds another 5500 circles so less than 7% extra.

Where I'm heading with this is that there are places which have had no broadband for 8 years, like Ashby-de-la-Launde, and we do not appear to be addressing them as a specific target. Rolling them up into a county wide scheme may still see them left out as "the uneconomic part" as councillors seek to deliver maximum connections per public expenditure. Should we not identify them and have a specific plan to solve their problem ?

Do we need a BT Openreach product that is a single run fibre optic upgrade from exchange to premise that could be ordered in the same way as a new telephone line, perhaps. This would start the incremental upgrade to our telecoms infrastructure in the areas that are currently the least well served.

If we don't have a plan for BT to upgrade its infrastructure, or for Virgin Media to expand its, then we need a plan for an alternative infrastructure in areas where the existing services are inadequate. This probably means a plan to specifically not spend public money within 2 or 3km of a BT telephone exchange, other than in exceptional cases with demonstrable infrastructure problems.

Labels: , , , , ,

Tuesday, February 08, 2011


IPv6 Part 1

It seems about time to start looking at and understanding IPv6 ahead of the inevitable transition. One of the ISPs resold by South Witham Broadband is Entanet and they are active in supporting IPv6 so I bit the bullet and bought a £10 Speedtouch 546v6 off Ebay to provide a bridged connection to my ADSL line.

The Speedtouch is used to provide a transparent bridge between the BT provided ADSL link to my router, in other words it acts as an ATM to ethernet bridge and allows the router to activate a PPPoE session and log in to Entanet. My router isn't IPv6 yet, but I can use a PC directly for starters. My initial objective was to get the bridge mode and PPPoE working, rather than change everything at once and then struggle to understand why it wasn't working.

First I put the Speedtouch in circuit as a conventional PPPoA NAT router, as it comes out of the box, and established that it worked. Then I switched it to PPPoE NAT mode retaining the login username / password on the Speedtouch. That worked too, so then I set about switching to bridge mode.

The instructions at the above entanet page didn't work for me, nor did several others on the internet. Using the command line interface on the Speedtouch is very powerful but some prior knowledge is helpful. I know now that the multiple levels of interfaces have to be dissembled and assembled in order, and most online guides don't do this. The stack of layers looks something like:


So you can't remove or edit an ATM layer if the Ethernet layer that runs on it is still active. The command sequence from entanet missed out the command for the ethernet layer :-

ppp ifdelete intf=Internet
atm ifdetach intf=atm_Internet
To get the second command above to work you have to precede it with
eth ifdetach intf=ethoa_Internet
otherwise you get an obscure error message suggesting you typed the interface name wrong !

A Norwegian called Roy has the simplest approach - tear the lot down and start again - posted on his blog I tweaked this for the UK VPI/VCI settings of 0,38 and reproduce it below :-

ppp relay flush
eth flush
atm flush
ppp flush
atm phonebook flush
atm phonebook add name=BrPPPoE_ph addr=0.38
atm ifadd intf=BrPPPoE_atm
atm ifconfig intf=BrPPPoE_atm dest=BrPPPoE_ph ulp=mac
atm ifattach intf=BrPPPoE_atm
eth bridge ifadd intf=BrPPPoE_br
eth bridge ifconfig intf=BrPPPoE_br dest=BrPPPoE_atm
eth bridge ifattach intf=BrPPPoE_br
This works fine, the Speedtouch becomes a brick with a phone input and an ethernet output and my WRT54GL router running dd-wrt firmware does the PPPoE log in to Enta.

Zen Internet have a template to download to the Speedtouch 5x6 routers, making bridge mode a drop down option. I haven't tried this, but it looks a good place to start if you prefer a point and click solution.

Next time we'll try to use an IPv6 capable device to log in. I have a new Mikrotik router ordered and I'm also waiting for Entanet to confirm they've switch my IPv6 on.

Friday, February 04, 2011


How Scaleable is Localism ?

The decision by Lancaster City Council to abandon a pilot Next Generation Access (NGA) broadband project in favour of a County Council scheme raises some interesting questions about the appropriate scale of localism.

Both projects were seeking public funding, £0.75 and £20m respectively, and both were obviously addressing the same issue. There is no reason however why the smaller scheme could not have continued and had its area excluded from the County scheme.

The local scheme was aimed at addressing known notspots (areas with either no broadband or less than 2M downstream) and was a full NGA solution with Fibre to the Home (FTTH). As part of the project preparation a lot of groundwork had been done getting small businesses on board with the proposal. An experienced local expert (Barry Forde) had contributed to the design which was innovative and also depended on community contributions to reduce cost.

The county wide scheme will not be able to rely on achieving the same level of local engagement. It also requires a minimum £20m investment contribution and by adopting a single partner for the whole county restricts itself to large telecoms companies or groups of companies acting together (as in S Yorks Digital Region). By nature of its size it will be slower to deliver and not solve the immediate problem of some notspots.

If we have a "localism" agenda, what is the scale of project we should attempt ? How far does "community" stretch and at what point does it fragment into multiple smaller groups ?

I don't have any answers, but a county is clearly too big a unit. The Cumbria Broadband Project is promoting "parish champions" and the BDUK backed pilot is targeting the Eden Valley rather than the whole county, both of which sound like good ideas.

By going for large schemes we rule out a single coherent community of interest in the project and we restrict the eligible solutions and solution providers to those that can handle large projects and have access to large investment budgets. Is that necessary ? or desirable ?

Perhaps we can't use County Councils to deliver local solutions, unless they are specifically charged with developing solutions made up of manageable sized units that can be individually tendered and tailored to suit an appropriate sized community ?

Wednesday, February 02, 2011


County Wide NGA for Lancashire

The Leader of Lancashire County Council has made a decision to "select a strategic partner" in order to "maximise the benefits from the deployment of superfast broadband across Lancashire". Impressive sounding stuff, subject to getting the nod by Cabinet on 3rd February.

In the report from the Chief Exec sets out the benefits of 'superfast broadband' or 'Next Generation Access' and observes that "About 34% of premises across Lancashire will not have superfast broadband by 2015" - hence the decision to proceed with a project to do something about it. The report also identifies key features of the project and potential partners :

  • Partners are expected to invest at least £20m to match the ERDF bid by the council.
  • The private sector partner will own the superfast broadband network
  • The private sector partner will be required to ensure open, equitable and transparent access to the network open to all service and communications providers
  • Partners to identify the extent to which existing technology in Lancashire, e.g. the Cumbria and Lancashire Education Online (CLEO) network, can be utilised
  • The completion of the network within 2.5 years from commencement of deployment
  • Delivers optimum coverage for the available investment funding, aiming for close to 100% coverage in Lancashire, including rural, remote and sparsely populated areas
  • The County Council will not prescribe the technology.
A study by Analysys Mason for the NWDA last year titled "Making NGA a Reality in the Northwest" looked at the costs of providing NGA to areas in the region (not just Lancashire) likely to be left out of a commercial provision and concluded that the cost of providing FTTC to those properties would be about £164m with the cost of FTTH some six times higher and just short of a billion. Given the current state of finances the FTTC solution is somewhat inevitable and it was also the route taken by the South Yorkshire Digital Region project.

I will put money on BT winning this tender, as they did the CLEO operating contract by forming a JV with Lancashire County Council. They even have a JV company up and running to do it. They may be the only company big enough to deliver it in the timescale, and certainly they have a track record of providing a competitive set of retailers unlike other "open access" attempts that have failed to impress.

So what will a BT technical solution offer, and how can a tender document pre-empt some of the likely shortfalls ?

BT will use FTTC fibre to the cabinet to connect a large proportion of the population to NGA. VDSL from the cabinets will deliver the sort of bandwidth seen as necessary in the next 5 years at around 18-20Mbits/s sustained demand for streaming HD video.

Those not connected to a cabinet will not get FTTC, unless BT undertake to re-engineer their copper network to make interconnection points at cabinets available. This I think is unlikely. So a different solution is required for those close to the telephone exchange and for those far away but not connected via a cabinet.

For those close to the exchange without a cabinet a universal upgrade to ADSL2+ services would be a start, but it isn't NGA. This means either fibre from the exchange to the home (FTTH), more "artificial" cabinets to create break-in points, or adoption of exchange based VDSL - currently not practised in the UK but potentially a solution.

Those further from the exchange but cabinet-less could again be addressed by FTTH or by fixed wireless broadband dimensioned to deliver a next generation service, as described in a recent report to the Broadband Stakeholders Group.

The final group of customers will be far from the exchange on long lines, possibly in isolated or sparsely distributed properties, and possibly in a current "not spot". For this group there are the same options - fixed wireless or FTTH. Copper based solutions are ruled out by the line length and consequent signal loss, and there is no concentration point to deploy FTTC.

In odd cases a satellite solution or non-NGA 2 Mbits/s broadband over copper might be used for properties that are otherwise "awkward". Remember we're looking for "close to 100%" not actually 100% with no exclusions. These folks will get the "Universal Service Commitment" rather than "Next Generation Access".

So what should the tender document say ?
  • It should specify a minimum sustained downstream bandwidth for the service, perhaps 18-20 Mbits/s on the basis of future bandwidth needs.
  • The above speed should be achievable for at least 90% of peak hours 8am - 10pm.
  • Upstream speed shall be at least 5 Mbits/s or 25% of the minimum downstream.
  • NGA solutions are required for >97% properties close to and connected directly to telephone exchanges, that are not served by cabinets.
  • NGA solutions are required for >90% of more distant properties that are not served by cabinets.
  • Every single property shall receive at least a 2 Mbits/s downstream / 500 kbits/s upstream service to meet the universal service commitment, by whatever technology.
Such a specification is technology neutral and could be met by BT, by an HFC cable network operator building out or by a brand new FTTH infrastructure provider. The numbers can be adjusted to individual taste, but the above might fit the bill without the invoice being too big.

The biggest risk in my view is that we create "NGA doughnuts" of FTTC around exchanges with nothing in the middle and nothing outside of the ring of cabinets around the exchange.

This page is powered by Blogger. Isn't yours?