Persistence of important data on the Tangle

Question

To this date, the IOTA network truncates all stored transactions after a snapshot, so that we basically only have an image of addresses with funds.

Now lets assume following scenario: Company A stores their car's system metrics such as mileage and service count on the Tangle with a (right now non existent) standardized schema. This means that zero value transactions are issued by the car or a central node of the company, which include the data as the payload inside the transactions. Lets assume a buyer B wants to verify, that in fact that second hand car of Company A he/she buys, really has the mileage written on the tag of the seller. Now buyer B could verify that in fact the mileage is correct by querying the IOTA ledger for transaction of the particular car. I imagine that this would work by scanning a QR code from the car's windshield where a service provider (company S) provides an app which lists the last X amount of transaction of the car, pulled from the Tangle and displays them in a pretty way.

There are several questions which arise from this scenario:

• If a snapshot happens, the transactions of the car containing the relevant data is lost. There's no incentive for anybody but the car company A to permanently store the transactions of the cars they manufactured after a snapshot. Obviously we would have X amount of other companies storing data on the ledger but really nobody but the companies themselves would really care about the their own produced data. Now the data is only held by company A, transferred via company S: how would you verify that the data wasn't tempered with after it got from the companies permanodes to your smartphone's display? Your only source of the data is company A because nobody else holds it, so you can't check it against other sources. Even if the payload of the transaction held a signing key instead of the actual data, there's no way to ensure that the signing key wasn't forged at a later time or during transmission of the data from company A->S->your smartphone.
• With the Tangle you can only really verify that a particular transaction was issued by a specific party but the data inside the payload must not really be true at all: while this works for the value transfer inside the network by using the native IOTA token, it doesn't when the data is generated somewhere else in the first place. In that sense, the data of the mileage could be faked by the car issuing the transaction in the first place. How does someone verify data truthfulness if the data is generated by non-standardized proprietary sensors before they even hit the Tangle in the first place?

For question 1 I could imagine that a standardized permanode of all car manufactures could be used, where everyone submits their metrics data and it is stored indefinitely and is legitimized by all governments using the specific technology which allows to query that database. However, what would be the incentive to use the Tangle for this particular application in the first place, when a simple standardized and approved API by multiple governments would be much faster and optimization would be easier to implement by the working group who created it?

Answer 1

With the Tangle you can only really verify that a particular transaction was issued by a specific party but the data inside the payload must not really be true at all.

This is of course true but it is not a problem that Iota is attempting to solve. If forgery was happening on a large scale then the company doing it would risk a lot and would be easily exposed. If it happens on individual cases it is something for the sensor/chip manufacturers to guard against. Iota is concerned only with data integrity once it is sent to the tangle. However I could imagine their might be some system with firmwares being hashed and sent to the tangle as an audit trail.

Now to the larger question of the data's persistent availability.

If a snapshot happens, the transactions of the car containing the relevant data is lost. There's no incentive for anybody but the car company A to permanently store the transactions of the cars they manufactured after a snapshot.

It is true that car company A is incentivized but we will see why other parties will also store that data.

Someone who queries the tangle for data has to rely on two things to be sure that they have the true data.

1. The user needs to see that the data they request is part of the tangles unbroken history of branching transactions. Each transaction is always hashed with the parent transaction hashes so that nobody can insert something into history. So a company that just keeps only their own transactions is not proving anything. They need to have the entire history, including everyone else's transactions.

2. The user needs to be sure that they are not being fooled by an entirely made up tangle. So they must cross-check the data they request with one or more other copies of the tangle. I imagine that an eventual system of querying the tangle history will have to include some kind of random cross check with other perma-nodes.

Because of these factors it would seem that anybody who has an interest in storing historic data is going to have to store a lot of it but clearly not all companies will have the scale to be able to do that. This leads to an outcome which will also form a part of the tangles eventual infinite scalabilty: Economic clustering.

Essentially data will not be stored universally across all nodes but be separated by clusters of nodes. So for instance, all the data for all cars in Germany may be on one cluster. And each car manufacturer that operates in that territory would therefore run a node in that cluster and not have to store all the data that doesn't interest them.

Clusters would be able to share data/transactions with other clusters that they need to. So maybe all car clusters in Germany will connect with all Parking clusters in Germany.