Technology cost report, Derek Beevor Sep 2004

Introduction

Last year I wrote a paper criticising the Governments LRUC for proposing over complex and expensive technology. That paper was called: "Lorry Road User Charge: a debacle in the making"

This paper suggests using standard Internet technology at a fraction of the cost of the Government's over complex solution.

The Government's Lorry Road User Charge is a proposal to change the way that lorries are taxed. The idea is to charge a fee for every mile / kilometre that a lorry drives on UK roads. The reason for this scheme is to attract revenue from foreign trucks that use the UK roads and, at present, do not pay for this use.

The scheme is due to start in 2008 and a pre tender analysis by INPUT, the public sector IT research and Consultancy Company, has estimated that the total value of the contract will be in the region of £4 billion over the next ten years.

At the time of writing (Sept 2004) no details of the proposed systems have been made available but proposals have been submitted to the Government by ten companies to provide the system. This list will be reduced to three companies by early 2005 and eventually one company will win the contract in mid 2005.

The function of this document is to investigate why the estimated cost of this scheme is so high and to open the debate into the costing before a contract is awarded.

One of the suspected reasons for the high cost is that the Government want to base the system on proprietary in vehicle computers that will use GPS satellite tracking, mobile phone communications technology and onboard mapping. This method of tolling has been tried In Germany and failed. The reason for the German failure was the use of over complex and proprietary technology even though the German scheme was on a much smaller scale (motorways only) than the proposed UK scheme. At the time of writing the German scheme, due to start in Aug 2003 is still not working.

This document is an investigation into using much simpler and existing technology to achieve the Government's aims of LRUC at a fraction of the cost.

An Alternative Method of Lorry Road User Charging

When looking into the technology cost of LRUC it is important to factor in Moore's law because we are designing equipment now for use in three years time. Moore's law made in 1965 by Intel founder Gordon Moore observed that the number of transistors on integrated circuits tend to double every 18 months, this law has been modified over the past two decades to include the cost and power of computer equipment. It is now generally accepted that computer equipment will half in price and double in power every 18 months.

Moore's law suggests that any LRUC onboard computer that is designed in 2005 at a cost of £1,000 should cost £250 by 2008 when the onboard units will be sold to the LRUC users.

Interestingly, the Government are currently negotiating the procurement of the LRUC equipment but are not involving the end users in the debate. No doubt the end users will end up paying for the LRUC equipment that the Government are procuring on their behalf so they will want to ensure that Moore's law is factored into the deal. Or is it possible that the Government is negotiating a price now that will not change over the next three years making the provision of LRUC equipment a very valuable contract for the provider and a very bad deal for the end users?

If the LRUC equipment is proprietary, the suppliers would end up in a monopoly position and the market forces that would normally follow Moore's law would not apply to them even though their manufacturing costs in 2008 would be 25% of today's prices.

This makes a very strong argument in favour of using standard PC equipment for LRUC that will follow Moore's law. Today's price for a rugged hand held PC is around £1,000, this price includes a standard wireless network card. Wireless "hot spots" are available all over the country, coffee shops and motorway service stations for example allow the user of standard PC equipment to log onto a broadband Internet connection for e-mail and remote connection to their offices.

The Government are very keen to encourage Internet systems for delivering services and a standard Internet system would be particularly relevant for delivering the technology for the Alternative Method of Road User Charging.

The proposed technology for vehicles entering and leaving the country would be to install standard wireless routers at each of the six major ro-ro ports and the channel tunnel. This will give broadband Internet connectivity to anyone with a wireless enabled PC. These PCs could be hand held units that are issued to Customs staff or mobile PCs used by the truck drivers.

The LRUC back office system could be an Internet site that all the users would log onto. As with many Internet sites it could easily be both multi lingual and secure. Of course, being Internet based, any Internet connected PC could also access the system.

The Internet site would have two main parts, an admin area for creating and maintaining vehicle accounts and a driver interface area for logging vehicle mileage and booking vehicles in to and out of the country.

All UK based vehicles would have a vehicle record on the system that included the odometer reading. A vehicle owner could update the odometer reading at any time to give them the accumulated road user charge for auditing. When a vehicle is sold the account will be closed and the new owner would open a new account for that vehicle. When a vehicle has its annual test the end of year odometer reading could be entered by the testing station to provide a final annual road user charge for that vehicle.

Because the system is Internet based a new vehicle record could be added by a foreign vehicle owner long before the vehicle turned up at the port, equally the driver or a member of the port staff could use a hand held computer to create the new account. Each new vehicle entered into the system would generate a unique password for that vehicle. Standard Internet credit card or Pay Pal type payments can be made to purchase credit for foreign vehicles.

The functionality of the booking in to and out of the country could be very simple, identify the vehicle by entering the unique vehicle password and confirm the booking by entering the vehicle registration and the vehicles odometer reading.

Police and enforcement staff, anywhere in the country could check that a vehicle had registered by using standard lap top computers with an Internet connection.

The technology set up costs for this system is very reasonable. Setting up wireless "hot spots" is very cheap. A home wireless router costs around £70 but the sort of equipment used by hotels, service stations and coffee shops is around £1,200 per site, we should increase this to £2,500 per site to add weather proof casing and lightning conduction. A broadband internet connection cost around £39 per month but this is a typical home user or coffee shop type connection that has a contention ratio (shared) a professional web hosting un-contended connection would cost around £850 per month per site.

The back office system would be an Internet server and back office database. The cost of the back office system will once again be very reasonable because the record sizes are quite small. See Appendix A

Based on 6000 trucks entering the UK every day and 6000 exiting every day we would only be generating around half a gigabyte of data per year. The owner records are only 5 megabytes of data and the vehicle records are 1.5 gigabytes. These are relatively small amounts of data, for example nowadays the average PC has 120 gigabyte of storage, in other words enough for around 60 years of data.

The back office system would be a standard Internet server able to deliver 20,000 pages per day and run a database able to store and report on around two gigabytes of data. Along with a firewall and backup system the capital cost of the hardware would be well under £20,000.

The cost of the database software would largely depend on the database chosen. For a database of this size it is very common for web developers to use standard open source software such as "My SQL" which is free of charge. However, we will assume that one of the well known commercial database package was chosen that was priced on a per user basis and the cost could be around £50,000.

The development of the secure web site, password system, payment system and database interface would cost around £80,000.

The records holding UK vehicle mileages and owner details would be very similar to the system that your local garage would use to book your car in for a service.

The port booking system is actually a simple Internet based warehouse system. You will be booking vehicles into your store / country and booking them out again. The enforcement staff around the country are just carrying out a stock check. Once a vehicle leaves your store / country a payment is made.

I haven't included an accounting package in this report because any of the well known packages could easily do the job and the prices vary so much between them. As a rough guide the hardware, software and interface into the database should cost around £100,000.

The fuel rebate system

This is a straight forward task that is well within the scope of a competent accounting package.

The LRUC operator has a product just like any other commercial company; the product is the use of a kilometre of UK roads. The accounting system can read the LRUC database and see that the customer (vehicle owner) has used (x) number of kilometres, this amount goes into the sales ledger.

The customer (vehicle owner) is entitled to a fuel rebate (credit note) based on the amount of fuel that has been purchased. The accounting system just reads the sales ledger, deducts the credit note and produces an invoice.

What we don't know yet (because the Government hasn't told us) is how we work out the credit note amount. Once the Government develop a formula for the fuel rebate scheme we can tackle that question.

Another reason for not including an accounting system in this cost report is that a separate accounting system may not be necessary. Let's not forget who is running the LRUC scheme: its HM Customs and Excise; the VAT man, and they already have one or two systems in place to handle rebates. Collecting money from almost every business in the UK (output tax) and then rebating some of it (input tax) is the main function of VAT and is a routine that is well known by every company accountant.

System costs

Cabling costs are just a guess for the sake of this pricing exercise, wireless routers need a site survey for positioning which is normally £850 per site.

Maintenance cost are around 13% per year for both hardware and software so the annual maintenance would cost £50,440, Broadband connection to 7 sites would cost £71,400 per year

Let's say that there are six ro-ro ports and the channel tunnel to equip. Each of the seven sites will need four wireless routers to cover the area and 20 hand held computers for their staff.

Possible improvements

A local server could be located at each port and still update a central sever. This would give a high degree of back up and allow the system to run smoothly via its wireless network even if the Internet connection was down. Running local servers allows us to monitor the wireless hand held computers MAC address for added security.

Internet café installed in port terminals for drivers to use for booking themselves in and out of the country.

A UPS/Battery power system could be added to support Port based equipment. Ports have a nasty habit of using some very powerful electric motors, and are not known for having a nice clean power supply

All licensed HGV drivers in Europe, will be issued cards and driver ID's to use with the new digital tachograph systems. It would be nice to tie a LRUC driver ID in with this system.

Enforcement

The police or other enforcement staff that are away from the ports and are not near wireless "hot spots" would need to use standard GPRS data connected PCs for Internet access. A lap top wireless enabled PC costs around £1,000 and a GPRS data contract would cost around £8 per month per PC.

Wireless "hot spots" are very cheap to install and could be rolled out to all major truck parks making enforcement even easier and cheaper. These sites will not need the dedicated 2 Meg line that I have recommended for the ports, a much cheaper connection would do. These wireless enabled truck parks would also allow the Government to offer some incentives to drivers for using the LRUC system. Voice over IP is now readily available on any broadband connected PC. It would not cost the Government anything to allow drivers to make free phone calls home (to broadband connected PCs) from their handheld PC over the LRUC broadband connection. This would be particularly valuable for foreign drivers.

Fraud detection

There is a much higher risk of fraud with the Governments proposed system based on in vehicle computers. Computers need a power supply which can, quite simply be switched off and GPS satellite signals can be easily blocked. It won't be of benefit to a UK haulier to switch off his computer because he will want to claim his refund for miles driven, it will however be a benefit to foreign drivers who don't buy fuel in the UK and will not be claiming refunds. For these foreign vehicles, manually checking mileages at the port will prevent having to set up elaborate anti fraud systems around the country.

Advantages

Using standard Internet and wireless network technology from the start completely eliminates compatibility problems across Europe. For example a vehicle operator in Germany could buy a Panasonic handheld computer in France and be able to log onto the LRUC web page in Dover just by using the standard browser software that comes with the PC. The only development the LRUC scheme needs is to write a secure web page with a back end database.

All of the prices mentioned here are today's (2004) prices. Moore's law will not apply to the software development (£80,000) or the maintenance charges but will apply to everything else.

Using standard Internet technology is in line with the e-government strategic framework for public services

All equipment can be purchased in the high street allowing the users to buy and maintain their own systems.

Standard wireless networking means there is no need for mobile telecoms costs and no need for complicated set up routines.

We are avoiding the three most complicated and unreliable elements of a LRUC system. GPS satellite tracking, mobile telecom data and onboard mapping.

This method starts LRUC off and can be used whether vehicles have a mobile computer or not. At a later date, when perhaps all vehicles are fitted with a mobile PC and the digital tachograph is installed as standard, a digital feed from the tachograph can be read by the PC to obtain time and mileage data. Wireless hot spots at the vehicles home base can then be used to update the main LRUC servers automatically.

Conclusion

The main difference between this alternative LRUC technology and the Government's proposed system is that this system uses standard Internet technology and the Government's version will use proprietary onboard computers. This simpler Internet version will rely on people manually entering mileage readings but the Government's system will do this task automatically. The first assumption from non IT people is that the automatic system will be much more reliable than a manual entry system. They will assume that because it's a computer it will automatically get it right every time. People in the IT industry will know that this assumption is wrong and relies on 420,000 onboard computers working with near perfect reliability for the whole of their working life in the harsh environment of a commercial vehicle.

The reality is that both systems will have a degree of inaccuracy, the manual entry system will have the human error factor of entering the wrong digits and the automatic system will suffer from system failures. Which version will be the most reliable is a matter of personal opinion but even the most ardent fan of automatic systems will have a hard job justifying the £4 billion price difference just to enter digits automatically.

We can do quite a lot to ensure as much accuracy as possible on the manual entry of mileage into the hand held computers. Just like many online bank accounts or commercial sites we can ask the user to enter the digits twice to confirm that they match. We can also check the last reading that the user entered to make sure that the number entered is larger that the last entry. If the system prevents mistakes that are lower than the last entry then any mistake made will be to the LRUC benefit and not the user. It will then be the user's responsibility to find the mistake and claim a refund.

There is a very real danger that this alternative Internet based system will be dismissed out of hand because of the system cost, it is far too cheap. The LRUC scheme is seen as a large Government IT project and everyone involved is price conditioned to around the £4 billion level. The ten companies who are currently bidding for the contract will have budgeted very large sums of money (possibly more than this proposals entire cost) just to enter the bidding process. There are over 80 staff at HM Customs and Excise working full time on this project, they won't like a proposal that satisfies the aims of LRUC and costs far less to buy than their office costs to run.

£4 billion is an extraordinary amount of money. This project will make the Scottish Parliament building and the Millennium Dome look like good value for money; at least they were under £1 billion each.

The reality is that this LRUC proposal is a credible alternative and it is priced realistically for a commercial Internet site. There are thousands of businesses that use the Internet to trade who, very obviously, don't have or need billions of pounds to start up. .LRUC is just another business that is particularly suited to Internet trading and let's face it, the function is hardly complicated, collecting truck mileages.

Appendix A

A rough guess at the data size would be to record:

Assume 10,000 vehicle operators. Double the data size for indexing, Total data size 5.03 Meg

Assume 10,000,000 vehicles. Double the data size for indexing Total data size 1,487 Meg

Assume 6000 vehicle entry and 6000 vehicles exit the country per day. Double the data size for indexing. Total annual data size 509,60 Meg