A distributed ledger as the future of ticketing?

[AngusH]

In another thread, there was discussion about whether tickets should be stored in a database and the difficulties that this would cause for issuing tickets on trains.


Instead, I'd like to put forward the idea of using a distributed ledger as the future of storing ticket information.

The crucial idea here is that each authorised issuing machine can independently create records which are valid, and the devices transfer the records between themselves whenever they can to create a single consensus.

There are some issues obviously, and this isn't something that is going to get deployed tomorrow, but it is an interesting thought as to how things might work in the future.

(It's also rather fashionable, so avoiding hype is important)

This uk gov introduction: https://www.gov.uk/government/uploa...972/gs-16-1-distributed-ledger-technology.pdf

 

[35B]

My immediate reaction is that this is still bleeding edge technology, and the use case for railway ticketing needs to support very disparate levels of data connection and kit.

 

[takno]

Yup, with bitcoin transactions currently taking an average of over 50 minutes to complete and power usage just on blockchain systems starting to have a real impact on global energy usage I think the time has absolutely come to roll this hopeless technology bubble out over the low value ticketing market

 

[Abpj17]

erm....there are already a number of misunderstandings on this thread

- blockchain is where things are starting to take off. bitcoin has a number of well known flaws.
- the length of time thing is only relevant where the nodes aren't trusted and so you need sufficient consensus or proof of work. if the nodes are trusted (like ticket offices, TOC websites, ticket machines) the length of time for the transaction is near-immediate (and consensus isn't a factor for the validity of the transaction). The time gap is then time taken to reach other nodes/part of the distributed ledger.
- the energy usage is a feature of proof of work systems. That's complete irrelevant here. That's for systems with a finite supply such as currencies. Not for ticketing systems.
- blockchain systems already exist for some use cases and readily support different types of kit. There needs to ultimately be a single protocol somewhere but not necessarily identical hardware

https://www.imeche.org/news/news-article/hs2-rail-future-blockchain-li-fi-and-much-more is an informed article about some of the possibilities - benefits and challenges

and https://media.consensys.net/the-future-of-tickets-a729ea4e9c95

 

[takno]

erm....there are already a number of misunderstandings on this thread

- blockchain is where things are starting to take off. bitcoin has a number of well known flaws.
- the length of time thing is only relevant where the nodes aren't trusted and so you need sufficient consensus or proof of work. if the nodes are trusted (like ticket offices, TOC websites, ticket machines) the length of time for the transaction is near-immediate (and consensus isn't a factor for the validity of the transaction). The time gap is then time taken to reach other nodes/part of the distributed ledger.
- the energy usage is a feature of proof of work systems. That's complete irrelevant here. That's for systems with a finite supply such as currencies. Not for ticketing systems.
- blockchain systems already exist for some use cases and readily support different types of kit. There needs to ultimately be a single protocol somewhere but not necessarily identical hardware

https://www.imeche.org/news/news-article/hs2-rail-future-blockchain-li-fi-and-much-more is an informed article about some of the possibilities - benefits and challenges

and https://media.consensys.net/the-future-of-tickets-a729ea4e9c95

Okay, so if we're moving beyond proof of work, and don't need a decentralised trust mechanism (which we absolutely don't), what benefits are blockchains bringing to the table here? We're definitely entering the phase where the tech is being suggested as a solution to an endless series of problems people don't actually have

 

[Abpj17]

I'd suggest you read the two articles which contain multiple examples

 

[takno]

I'd suggest you read the two articles which contain multiple examples

I got through the imeche one just about. It's drivel, starting from lifi and semantic Web, which were both going nowhere ten years ago, and then going onto blockchain. various examples are provided of data being useful, which it obviously is, but in most cases it's not data where absolute accuracy or definitive history is even important. This tenuous talk about data is then used by a series of futurologist wonks as a justification for some barely-related technology, and we're all just supposed to nod and assume they're smarter than us. There are no problems with middlemen that are likely to be solved with a whole network of middlemen, and if you take the leccy use out then really there isn't a lot in any blockchain technology I've seen that is particularly new or innovative

 

[AngusH]

The original problem was much simpler:

1) There are advantages to storing all tickets in a database then merely issuing references to that database (much as airlines do). It removes a number of problems that sometimes occur with the various railway ticket machines, print at home and m-ticket systems.

2) There are problems with (point 1) with respect to on-train ticket machines and machines in locations where data connections are intermittent or unreliable.


A distributed ledger is one approach that allows records to be recorded accurately and securely at the time of purchase in final form and then synchronised together at a future point, then the result distributed back to all of the fixed and mobile devices, either in whole or in part.

It would hopefully allow things that aren't currently so easy, like invalidating paper tickets that have been lost then reissuing them, or moving m-tickets to paper tickets. Because each of these would merely be a reference to an entry in the ledger.

Edit: It would be essential though to keep an accurate record of who made modifications to the records and where. Which is an inherent advantage of a ledger approach.

Yes, there are assorted difficulties, but I think the end result would be rather useful and would reduce both cost and customer service issues.

This is at some point in the future though, I'm not suggesting implementing the thing tomorrow.

Also, this is wholly separate although technical similar to the bitcoin system.

edit reasons: tone and word choice

 

[Abpj17]

really there isn't a lot in any blockchain technology I've seen that is particularly new or innovative

You're right. In a lot of ways, blockchain isn't new. Companies have been operating large central databases for years.

What is newer, is the commercialisation of a more dynamic way of managing and changing entries in that database, including by end-users.

Smart contracts are particularly innovate and could allow more responsive pricing and/or resale/refund/amends of tickets.

The technology would be particularly supportive of dematerialisation of paper ticketing. And the creation of ticket records that overcome issues with multiple TOCs and squabbles about whether a passenger has a ticket or not.

The fragmented nature of train ticketing in the UK makes it ideal for a single database (blockchain) to serve multiple TOCs and ticket retailers.

It's not as bleeding edge as it was - event ticketing is starting to use blockchain to reduce/eliminate fraud/ticket touts. There are a vast number of applications in financial services that are already operating as proof of concepts and going into production.

Anything still focussing on bitcoin is probably hype. Any critics dismissing blockchain through citing bitcoin probably lack the background of what has been happening. Blockchain has moved far beyond bitcoin with multiple systems that don't have the flaws of bitcoin.

 

[Bletchleyite]

I think there is something in this, but I think the cryptocurrency elements are probably totally unnecessary, just simple HTTPS types encryption - it would be adequate to have a single centralised database, but then to sync it with ticket issuing/verification devices in some form as data connections become available and unavailable. So it'd be a bit more like a Dropbox than bitcoin, simply because every device that could add or modify the "truth" is by definition trusted.

Where it could come in is the likes of the Far North Line. A ticket is sold for travel from Altnabreac with no signal, the ticket holder gets a printout of it with a generated reference number, but the ticket doesn't make it into the database until arrival at Inverness when the guard's machine picks up wifi (realistically it'd probably be before). This is no great problem because the passenger won't need to verify it until they get there.

You'd also have the potential to download the reference numbers of tickets known to be valid on a certain line onto the guard's verification device when wifi is available, so if someone buys a ticket at the TVM at Inverness it's already on the guard's machine before they leave.

As this issue primarily applies to rural areas, the number of tickets involved would be somewhat reduced from the rather scary problem of "all of them".

That said I still expect there will be 100% coverage of the UK within 10 years.

 

[takno]

You're right. In a lot of ways, blockchain isn't new. Companies have been operating large central databases for years.

What is newer, is the commercialisation of a more dynamic way of managing and changing entries in that database, including by end-users.

Smart contracts are particularly innovate and could allow more responsive pricing and/or resale/refund/amends of tickets.

The technology would be particularly supportive of dematerialisation of paper ticketing. And the creation of ticket records that overcome issues with multiple TOCs and squabbles about whether a passenger has a ticket or not.

The fragmented nature of train ticketing in the UK makes it ideal for a single database (blockchain) to serve multiple TOCs and ticket retailers.

It's not as bleeding edge as it was - event ticketing is starting to use blockchain to reduce/eliminate fraud/ticket touts. There are a vast number of applications in financial services that are already operating as proof of concepts and going into production.

Anything still focussing on bitcoin is probably hype. Any critics dismissing blockchain through citing bitcoin probably lack the background of what has been happening. Blockchain has moved far beyond bitcoin with multiple systems that don't have the flaws of bitcoin.

Smart contracts are a nice idea and have applications in certain fields, and ethereum is a much more interesting platform than bitcoin. They still haven't done anything to meaningfully solve the proof of work problem though, and until they do the ideas that underlie it are about as interesting to low-value transactions as nuclear fusion is to energy generation.

In the longer run, I think the appeal in train ticketing stands on a fundamental misunderstanding of the problem. We have a ticketing system which is too complex for anybody to understand, including staff. We don't have any particular issues with trust in the central point of authority, or any current issues with transactions not being honoured.

We would like to roll out self print tickets and smartcards, and in doing that we do have a few issues with access to central databases on trains etc, but there is nothing new in blockchain technologies that will help with that.

 

[linusnorton]

I thought I'd weigh in here because it's an area that I'm actively looking at. For reference my background is working for a TIS (ticket issuing machine) supplier so the design of my blockchain transport system is built around the trust issues of retailing and creating tickets.

The outline has been written up on my blog - http://ljn.io/posts/blockchain-transport/ - but I'll put a TLDR version here for discussion.

- Retailing tickets requires trust from the operators because retailing tickets is tied to creating tickets and when you create a ticket you are creating a liability for the operator to transport the passenger
- Currently that trust is earnt via an accreditation process with RSP (part of RDG) because issuing is difficult (there are many arcane formats) and each ticket issue must be accurately reported to the settlement system (LENNON)
- In theory, it is possible to provide that trust by having smart contracts issue tickets onto a blockchain and do the financial settlement in a single transaction
- It would then be easier to add more retailers into the industry because of the reduced barrier to entry
- On top of that, it would be possible to integrate tickets from different industries (bus, plane) and geographical regions (France, Germany) and not have to understand their ticket formats
- Then it would be possible for retailers to sell tickets for global, multi-modal journeys in a single transaction

That's what I'm currently working on. One limitation I've set myself is that I cannot change any of the existing infrastructure (gates etc) because it is just too expensive to change, though as mentioned above, it would be nice if that infrastructure did eventually use the blockchain directly.

So far I haven't found a use case for introducing a token (like a cryptocurrency) and I think that's a good think. Any sort of exchange rate would scare people and stifle adoption. That said it is difficult to monetize the network - someone has to bare the cost of running the machines whether it's a proof of work, proof of stake or other consensus system.

Thoughts are welcome

 

[takno]

I thought I'd weigh in here because it's an area that I'm actively looking at. For reference my background is working for a TIS (ticket issuing machine) supplier so the design of my blockchain transport system is built around the trust issues of retailing and creating tickets.

The outline has been written up on my blog - http://ljn.io/posts/blockchain-transport/ - but I'll put a TLDR version here for discussion....
Thoughts are welcome

Thanks for sharing Linus. It's always interesting to see a concrete idea, and I'd be interested to see more development on it. I'd push back on a couple of things.

As far as I can see the accreditation and agent-level settlement system in UK rail currently is complex for two reasons:

• The product being sold is complex with the transport providers are on the hook for what is promised to the customer
• There is little control over how many products are created.

It's important in addressing the first problem to not simplify away the consumer benefits, or pass the liability for mistakes onto the consumer instead. There is a real danger with smart contracts that the contract is not coded to match the understanding of the feeble human flesh it is designed to transport. Hell, it may not even match the printed itinerary they were provided with at the time of purchase. If this happens do you throw the agent off the network, or hold the customer liable and depend on the bad press to stop people using the agent? The way I see it, none of the problems go away and you are just left with a fancy new digital mess instead of a tastefully-branded tangerine mess.

The second problem will ultimately be addressed by a database of tickets (or sectors if a ticket covers multiple modes), and checking of individual ticket numbers at the time of travel. In UK rail that is probably most cheaply done using a central authority since there are few operators, most operating under government concession, and all bought into RSP/ATOC/RDG. You don't need and probably won't benefit from using smart contracts in this scenario at all.

I'm interested in the idea of spreading the ticketing to more modes and handling cross-currency transactions, but I wonder how interested consumers are in actually doing that, and whether the benefits outweigh the likely very significant costs of getting set up in the first place. The airline industry in the mid-nineties had great through-ticketing facilities, and we are now at a situation where the majority of air travel is bought direct from a single supplier, and consumers seem fairly happy with it that way. You could argue that they like it that way because of the price, and that the unified system was expensive because of intermediaries. I think that quietly ignores the probably-higher cost of itermediation through blockchain, but more importantly it misses the reason why the intermediaries are costly in the sector.

Put simply, what's appealing to consumers about through ticketing is that you pay a pre-determined fee and everyone will work together to get you home. That means that you have to have appropriate agreements in place so that the customer doesn't have to worry when Barry's Brokedown Buses, which is driving you to the airport for a £1 cut of the ticket price breaks down, and Once-In-A-Lifetime-Air has to find them a first class seat on another day to qualify for its $10000 cut. What happens then? Does Barry forgo his year's profits to cover the loss, or do the airline eat it? Did somebody get insurance to cover it?

My take on the majority of uses for blockchain is that they are an extreme version of the insane belief amongst Randian idealists that the market will solve all mankind's problems. All you've really done by changing the intermediation layer is give the world's layer of chancers and frauds free reign until version two when you realise that all that bureaucracy was there for a reason

 

[Bletchleyite]

Do we actually *want* more retailers?

I'm not sure I wouldn't see it as more sensible if the *only* place of online sale was nationalrail.co.uk, with that tendered periodically like the TOCs themselves. The plethora of retailers, including the well-known one that not only charges you extra but also has some sheep-based advertising that somewhat stretches the truth, doesn't really help anyone - there's not a lot other than the UI to compete on, and if a single site's UI was good enough where's the gain?

(I know there's a preference for "mixing deck" planners on here, and indeed those are my preference, but there's no reason a single site couldn't have an "expert mode" or "travel agent mode" like that).

OK, fans of trainsplit etc won't appreciate this, but there's no reason sites couldn't offer advice for a cost just like an Independent Financial Adviser.

 

[linusnorton]

It's important in addressing the first problem to not simplify away the consumer benefits, or pass the liability for mistakes onto the consumer instead. There is a real danger with smart contracts that the contract is not coded to match the understanding of the feeble human flesh it is designed to transport. Hell, it may not even match the printed itinerary they were provided with at the time of purchase. If this happens do you throw the agent off the network, or hold the customer liable and depend on the bad press to stop people using the agent? The way I see it, none of the problems go away and you are just left with a fancy new digital mess instead of a tastefully-branded tangerine mess.

I agree that there is a disconnect between what websites say you can do and what the rules actually are. Even the data that is made available online often contradicts itself (DTD fares restriction data vs NRE restrictions info for example) and there are many cases where there is only vague guidance given (refund rules). Retailers are always expected to handle refunds of tickets they are sold and I don't see that adding a blockchain/smart contracts would change that - hopefully it would provide a single more authority view on the rules by having an "official" refund contract provided, rather than each retailers own interpretation of the rules.

The second problem will ultimately be addressed by a database of tickets (or sectors if a ticket covers multiple modes), and checking of individual ticket numbers at the time of travel. In UK rail that is probably most cheaply done using a central authority since there are few operators, most operating under government concession, and all bought into RSP/ATOC/RDG. You don't need and probably won't benefit from using smart contracts in this scenario at all.

Yield management / quota control would be quite easy to build into the blockchain contracts. However, the quota information is only worthwhile if there is a single source of truth. Unless RDG decided they wanted it on a blockchain it would remain in NRS - though I don't necessarily think that a reservation system is the place for quota controlled products. Smart contract execution should be deterministic and NRS is decidedly not so I am still working through the best way to integrate quota controlled products.

Put simply, what's appealing to consumers about through ticketing is that you pay a pre-determined fee and everyone will work together to get you home. That means that you have to have appropriate agreements in place so that the customer doesn't have to worry when Barry's Brokedown Buses, which is driving you to the airport for a £1 cut of the ticket price breaks down, and Once-In-A-Lifetime-Air has to find them a first class seat on another day to qualify for its $10000 cut. What happens then? Does Barry forgo his year's profits to cover the loss, or do the airline eat it? Did somebody get insurance to cover it?

Initially my implementation would be individual tickets put together into a single transaction so in your scenario the consumer would lose out. That's no worse than if the consumer had arranged each leg with different operators themselves using multiple websites. My hope is that once the system becomes established that true through tickets are created with direct operator negotiation (this already happens in Europe). I assume details like missed journeys would be sorted out as part of that negotiation.

My take on the majority of uses for blockchain is that they are an extreme version of the insane belief amongst Randian idealists that the market will solve all mankind's problems. All you've really done by changing the intermediation layer is give the world's layer of chancers and frauds free reign until version two when you realise that all that bureaucracy was there for a reason

If you were talking about cryptocurrencies I'd probably agree but for the blockchain I think there are many legitimate use cases. Unfortunately most of them are not very sexy when compared to the insane hype of cryptocurrencies and don't attract much media attention.

Do we actually *want* more retailers?

I'm not sure I wouldn't see it as more sensible if the *only* place of online sale was nationalrail.co.uk, with that tendered periodically like the TOCs themselves. The plethora of retailers, including the well-known one that not only charges you extra but also has some sheep-based advertising that somewhat stretches the truth, doesn't really help anyone - there's not a lot other than the UI to compete on, and if a single site's UI was good enough where's the gain?

(I know there's a preference for "mixing deck" planners on here, and indeed those are my preference, but there's no reason a single site couldn't have an "expert mode" or "travel agent mode" like that).

OK, fans of trainsplit etc won't appreciate this, but there's no reason sites couldn't offer advice for a cost just like an Independent Financial Adviser.

This idea does touch some fairly sensitive subjects. In this case it's privatization. For better or worse we have a privatized railway and it looks like other countries are trying to follow suite (though not necessarily in the same way). If we have privatized railways then we should have an open market to retail tickets. Would split ticket sites ever have happened without an open market? I think they would probably have continued to be swept under the rug.

I'd also suggest that's about more than just the UK. There are UK retailers that would like to sell European / worldwide tickets and there are retailers around the world that would like to sell UK tickets. Right now they are restricted to a very limited number of approved retailers that they could partner with.

I won't even mention the lack of competition the trainline have.

 

[takno]

If you were talking about cryptocurrencies I'd probably agree but for the blockchain I think there are many legitimate use cases. Unfortunately most of them are not very sexy when compared to the insane hype of cryptocurrencies and don't attract much media attention.

I agree there are a couple of interesting technologies there, and sadly the use cases they are likely to drive are mostly very boring indeed to most people. That's what make the hype more annoying - they are fairly specialist technologies which actually have an impact similar to something like column-based datastores, which normal people have quite rightly never heard of. Instead of being able to ignore them, or learn about the stuff that's relevant to projects I'm doing, I'm faced with customers and random people who expect me to have an opinion on them, and want me to justify not including them in completely unrelated projects. Standard hype-cycle issues I guess...

Your platform definitely sounds interesting though, and taking it beyond the UK there could well be something there that improves peoples lives. And even if the blockchain element ends up being a minor cog in a bigger platform, hype cycles can be a great way to drum up interest and funding for an idea!

 

[Bletchleyite]

I won't even mention the lack of competition the trainline have.

As I mentioned I would prefer the competition to occur at the point of tender/contract than "on rail". However, Trainline have plenty of competition; the players in the market were listed on another post recently and there are at least 5 of them, including one (ATOS[1]) who are a really big player in railway IT generally but also a major worldwide company.

[1] The "mixing deck" is their product, and RedSpottedHanky their equivalent non-TOC branded site.

 

[35B]

Do we actually *want* more retailers?

I'm not sure I wouldn't see it as more sensible if the *only* place of online sale was nationalrail.co.uk, with that tendered periodically like the TOCs themselves. The plethora of retailers, including the well-known one that not only charges you extra but also has some sheep-based advertising that somewhat stretches the truth, doesn't really help anyone - there's not a lot other than the UI to compete on, and if a single site's UI was good enough where's the gain?

(I know there's a preference for "mixing deck" planners on here, and indeed those are my preference, but there's no reason a single site couldn't have an "expert mode" or "travel agent mode" like that).

OK, fans of trainsplit etc won't appreciate this, but there's no reason sites couldn't offer advice for a cost just like an Independent Financial Adviser.

I disagree. That presumes that the tendering process works effectively in the consumer interest, and that it can predict how retail will develop during the contract term. Even a brief glance at franchising should demonstrate the weakness of that hypothesis.

The benefit of multiple retailers, even in a constrained market like U.K. rail, is that there is scope for ideas to develop, because there is no single controlling mind that can presume to know how customers wish to purchase their tickets.

 

[Abpj17]

Smart contracts are a nice idea and have applications in certain fields, and ethereum is a much more interesting platform than bitcoin. They still haven't done anything to meaningfully solve the proof of work problem though, and until they do the ideas that underlie it are about as interesting to low-value transactions as nuclear fusion is to energy generation.

In the longer run, I think the appeal in train ticketing stands on a fundamental misunderstanding of the problem. We have a ticketing system which is too complex for anybody to understand, including staff. We don't have any particular issues with trust in the central point of authority, or any current issues with transactions not being honoured.

We would like to roll out self print tickets and smartcards, and in doing that we do have a few issues with access to central databases on trains etc, but there is nothing new in blockchain technologies that will help with that.

"proof of work" doesn't need solving if you're not going to use it. Not all blockchain implementations require it.

There are loads of issues with transactions not being honoured - the multitude of ticket disputes is evidence of that. In practice, the big benefit comes from paperless ticketing. But blockchain is a potentially effective way to implement that across the UK relative to other options.

 

[takno]

"proof of work" doesn't need solving if you're not going to use it. Not all blockchain implementations require it.

There are loads of issues with transactions not being honoured - the multitude of ticket disputes is evidence of that. In practice, the big benefit comes from paperless ticketing. But blockchain is a potentially effective way to implement that across the UK relative to other options.

Blockchain really isn't an effective way to implement paperless ticketing for the UK, for reasons I went into at length.

There isn't anything particularly innovative about distributed ledgers except for the mechanism for establishing trust in them, and I haven't seen anything apart from proof of work that actually does that. Life's too short to spend time on these technical fashion parades

 

[Abpj17]

Blockchain really isn't an effective way to implement paperless ticketing for the UK, for reasons I went into at length.

There isn't anything particularly innovative about distributed ledgers except for the mechanism for establishing trust in them, and I haven't seen anything apart from proof of work that actually does that. Life's too short to spend time on these technical fashion parades

You seem rather wedded to one particular variant of blockchain rather than the wide range of options that actually exist in practice? If trust is already established (a closed network like ToCs) then you don't need proof of work.

As above, I'm not claiming there is anything radically new. But blockchain is a very efficient and effective implementation compared to some of the predecessors.

It's potentially effective for paperless ticketing. And the concept of smart contracts could potentially eliminate ticket disputes and arguments about validity, while adding extra flexibility for customers.