TrustChainTutorial.md

Creating your first TrustChain application

In this tutorial we show how to build a simple application using TrustChain ledger as a tamper-proof storage. Specifically, we show how to create a proposal block, register a listener for incoming blocks of a certain type, and automatically create agreement blocks if the proposal block complies with our integrity rules.

This guide assumes that you have a high-level understanding of TrustChain architecture, in particular that you are familiar the idea of half-blocks, a basic data structure of a pair-wise ledger. You can read more in the TrustChain protocol specification.

Load TrustChain overlay

First, we initialize IPv8 in a similar way as described in the overlay tutorial.

TrustChainCommunity.Factory requires two compulsory arguments: TrustChainSettings defining protocol parameters, and the persistence storage implementing TrustChainStore interface. The following snippet can be used to configure the overlay with the default settings and the SQLite store implementation:

val settings = TrustChainSettings()
val driver = AndroidSqliteDriver(Database.Schema, this, "trustchain.db")
val store = TrustChainSQLiteStore(Database(driver))
val randomWalk = RandomWalk.Factory()
val trustChainCommunity = OverlayConfiguration(
    TrustChainCommunity.Factory(settings, store),
    listOf(randomWalk)
)

As previously, the OverlayConfiguration should be passed to IPv8Configuration, which is subsequently used to initialize IPv8.

Create a proposal block

In TrustChain, a half-block pair represents exactly one transaction. A half-block pair consists of a proposal block and an agreement block. When the transaction is performed between two parties, the first party creates a proposal block and sends it to the counterparty. The counterparty validates the transaction and signs an agreement blocks, and sends it back. Only at that point, the transaction is considered valid. For this reason, the protocol needs to be interactive and cooperation of both parties is required.

We get a reference to the TrustChainCommunity:

val trustchain = IPV8Android.getInstance().getOverlay<TrustChainCommunity>()!!

We start by creating a proposal block. We use TrustChainCommunity.createProposalBlock method that takes three parameters:

• The block type is used to distinguish blocks which are created by our application. We will later register a listener that will be triggered when a blocks of a specific type is received.
• The transaction can be represented by a map including arbitrary values that are serializable by TransactionEncoding class. Most of the Kotlin primitive types (Int, BigInteger, Long, String, Float, Boolean, null), and basic data structures (List, Set, Map), are supported.
• The public key indicates which counterparty is supposed to sign the agreement block. If we have a reference to the Peer object, we can get their serialized public key as Peer.publicKey.keyToBin().

The method creates a TrustChainBlock, stores it in the local database, and broadcasts it to all peers.

val transaction = mapOf("message" to message)
val publicKey = peer.publicKey.keyToBin()
trustchain.createProposalBlock("demo_block", transaction, publicKey)

Transaction validation

Optionally, we can register TransactionValidator for our block type to enforce integrity requirements. In the validate method, we also get access to TrustChainStore if there we need to define any context-sensitive rules depending on the previous blocks in the chain. Invalid blocks are discarded and not stored in the database.

In our case, we require proposal blocks to include a message field. We accept any transaction for agreement blocks as we are only concerned with the signature.

trustchain.registerTransactionValidator("demo_block", object : TransactionValidator {
    override fun validate(
        block: TrustChainBlock,
        database: TrustChainStore
    ): Boolean {
        return block.transaction["message"] != null || block.isAgreement
    }
})

Create agreement blocks

We now register BlockSigner that will be notified about incoming proposal blocks that we should create agreement blocks for. At the point when the onSignatureRequest method is called, the last few blocks (the exact count can be configured in TrustChainSettings.validationRange) have already been validated.

To create the agreement block, TrustChainCommunity.createAgreementBlock method should be called. The method accepts the proposal block and a transaction for the agreement block. This can be either the copy of the original transaction, an empty map, or any other content depending on the use case.

In our demo, we are just going to sign all valid blocks and include an empty transaction in the agreement block. In case your application requires user confirmation to sign agreement blocks, the block signer is not needed at all and the agreement blocks can be created in response to user interaction.

trustchain.registerBlockSigner("demo_block", object : BlockSigner {
    override fun onSignatureRequest(block: TrustChainBlock) {
        trustchain.createAgreementBlock(block, mapOf<Any?, Any?>())
    }
})

Get notified about incoming blocks

We can register a block listener to be notified about all incoming blocks. When the listener methods get called, the block has already been validated using the registered TransactionValidator, so we can just assume all received blocks are valid. TrustChainCommunity.addListener takes two parameters: the type of blocks we want to be notified about, and an object implementing BlockListener interface. onBlockReceived gets called for every incoming block of the specified type. We can perform any application-specific logic there if needed. For now, we just log the block ID and the transaction.

There can be multiple listeners registered for a single block type and a listener can be removed with TrustChainCommunity.removeListener method.

trustchain.addListener("demo_block", object : BlockListener {
    override fun onBlockReceived(block: TrustChainBlock) {
        Log.d("TrustChainDemo", "onBlockReceived: ${block.blockId} ${block.transaction}")
    }
})

Request a chain crawl

By default, blocks are only received passively when other peers broadcast them. However, it is also possible to request an active crawl of the chain from the specific peer using the TrustChainCommunity.crawlChain(Peer) method. It keeps sending crawl requests until we have all known blocks stored in our database. The crawlChain function is a suspending function that will resume once the whole chain is crawled, or a timeout is reached.