The ledger is the sequenced, tamper-resistant record of all state transitions in the fabric. State transitions are a result of chaincode invocations (‘transactions’) submitted by participating parties. Each transaction results in a set of asset key-value pairs that are committed to the ledger as creates, updates, or deletes.
The ledger is comprised of a blockchain (‘chain’) to store the immutable, sequenced record in blocks, as well as a state database to maintain current fabric state. There is one ledger per channel. Each peer maintains a copy of the ledger for each channel of which they are a member.
The chain is a transaction log, structured as hash-linked blocks, where each block contains a sequence of N transactions. The block header includes a hash of the block’s transactions, as well as a hash of the prior block’s header. In this way, all transactions on the ledger are sequenced and cryptographically linked together. In other words, it is not possible to tamper with the ledger data, without breaking the hash links. The hash of the latest block represents every transaction that has come before, making it possible to ensure that all peers are in a consistent and trusted state.
The chain is stored on the peer file system (either local or attached storage), efficiently supporting the append-only nature of the blockchain workload.
The ledger’s current state data represents the latest values for all keys ever included in the chain transaction log. Since current state represents all latest key values known to the channel, it is sometimes referred to as World State.
Chaincode invocations execute transactions against the current state data. To make these chaincode interactions extremely efficient, the latest values of all keys are stored in a state database. The state database is simply an indexed view into the chain’s transaction log, it can therefore be regenerated from the chain at any time. The state database will automatically get recovered (or generated if needed) upon peer startup, before transactions are accepted.
At a high level, the transaction flow consists of a transaction proposal sent by an application client to specific endorsing peers. The endorsing peers verify the client signature, and execute a chaincode function to simulate the transaction. The output is the chaincode results, a set of key/value versions that were read in the chaincode (read set), and the set of keys/values that were written in chaincode (write set). The proposal response gets sent back to the client along with an endorsement signature.
The client assembles the endorsements into a transaction payload and broadcasts it to an ordering service. The ordering service delivers ordered transactions as blocks to all peers on a channel.
Before committal, peers will validate the transactions. First, they will check the endorsement policy to ensure that the correct allotment of the specified peers have signed the results, and they will authenticate the signatures against the transaction payload.
Secondly, peers will perform a versioning check against the transaction read set, to ensure data integrity and protect against threats such as double-spending. The fabric has concurrency control whereby transactions execute in parallel (by endorsers) to increase throughput, and upon commit (by all peers) each transaction is verified to ensure that no other transaction has modified data it has read. In other words, it ensures that the data that was read during chaincode execution has not changed since execution (endorsement) time, and therefore the execution results are still valid and can be committed to the ledger state database. If the data that was read has been changed by another transaction, then the transaction in the block is marked as invalid and is not applied to the ledger state database. The client application is alerted, and can handle the error or retry as appropriate.
State database options include LevelDB and CouchDB (beta). LevelDB is the default key/value state database embedded in the peer process. CouchDB is an optional alternative external state database. Like the LevelDB key/value store, CouchDB can store any binary data that is modeled in chaincode (CouchDB attachment functionality is used internally for non-JSON binary data). But as a JSON document store, CouchDB additionally enables rich query against the chaincode data, when chaincode values (e.g. assets) are modeled as JSON data.
Both LevelDB and CouchDB support core chaincode operations such as getting and setting a key (asset), and querying based on keys. Keys can be queried by range, and composite keys can be modeled to enable equivalence queries against multiple parameters. For example a composite key of (owner,asset_id) can be used to query all assets owned by a certain entity. These key-based queries can be used for read-only queries against the ledger, as well as in transactions that update the ledger.
If you model assets as JSON and use CouchDB, you can also perform complex rich queries against the chaincode data values, using the CouchDB JSON query language within chaincode. These types of queries are excellent for understanding what is on the ledger. Proposal responses for these types of queries are typically useful to the client application, but are not typically submitted as transactions to the ordering service. In fact the fabric does not guarantee the result set is stable between chaincode execution and commit time for rich queries, and therefore rich queries are not appropriate for use in update transactions, unless your application can guarantee the result set is stable between chaincode execution time and commit time, or can handle potential changes in subsequent transactions. For example, if you perform a rich query for all assets owned by Alice and transfer them to Bob, a new asset may be assigned to Alice by another transaction between chaincode execution time and commit time, and you would miss this ‘phantom’ item.
CouchDB runs as a separate database process alongside the peer, therefore there are additional considerations in terms of setup, management, and operations. You may consider starting with the default embedded LevelDB, and move to CouchDB if you require the additional complex rich queries. It is a good practice to model chaincode asset data as JSON, so that you have the option to perform complex rich queries if needed in the future.
To enable CouchDB as the state database, configure the /fabric/sampleconfig/core.yaml stateDatabase section.