libp2p-bitswap

Implementation of the bitswap protocol.

Efficiently syncing dags of blocks

Bitswap is a very simple protocol. It was adapted and simplified for ipfs-embed. The message format can be represented by the following enums.

pub enum BitswapRequest {
    Have(Cid),
    Block(Cid),
}

pub enum BitswapResponse {
    Have(bool),
    Block(Vec<u8>),
}

The mechanism for locating providers can be abstracted. A dht can be plugged in or a centralized db query. The bitswap api looks as follows:

#[derive(Debug)]
pub enum BitswapEvent {
    /// Received a block from a peer. Includes the number of known missing blocks for a
    /// sync query. When a block is received and missing blocks is not empty the counter
    /// is increased. If missing blocks is empty the counter is decremented.
    Progress(QueryId, usize),
    /// A get or sync query completed.
    Complete(QueryId, Result<()>),
}

pub trait BitswapStore: Send + Sync + 'static {
    /// The store params.
    type Params: StoreParams;
    /// A have query needs to know if the block store contains the block.
    fn contains(&mut self, cid: &Cid) -> Result<bool>;
    /// A block query needs to retrieve the block from the store.
    fn get(&mut self, cid: &Cid) -> Result<Option<Vec<u8>>>;
    /// A block response needs to insert the block into the store.
    fn insert(&mut self, block: &Block<Self::Params>) -> Result<()>;
    /// A sync query needs a list of missing blocks to make progress.
    fn missing_blocks(&mut self, cid: &Cid) -> Result<Vec<Cid>>;
}

pub struct BitswapConfig {
    /// Timeout of a request.
    pub request_timeout: Duration,
    /// Time a connection is kept alive.
    pub connection_keep_alive: Duration,
}

impl<P: StoreParams> Bitswap<P> {
    /// Creates a new `Bitswap` behaviour.
    pub fn new(config: BitswapConfig) -> Self;

    /// Adds an address for a peer.
    pub fn add_address(&mut self, peer_id: &PeerId, addr: Multiaddr);

    /// Removes an address for a peer.
    pub fn remove_address(&mut self, peer_id: &PeerId, addr: &Multiaddr);

    /// Starts a get query with an initial guess of providers.
    pub fn get(&mut self, cid: Cid, peers: impl Iterator<Item = PeerId>) -> QueryId;

    /// Starts a sync query with an the initial set of missing blocks.
    pub fn sync(&mut self, cid: Cid, peers: Vec<PeerId>, missing: impl Iterator<Item = Cid>) -> QueryId;

    /// Cancels an in progress query. Returns true if a query was cancelled.
    pub fn cancel(&mut self, id: QueryId) -> bool;

    /// Register bitswap stats in a prometheus registry.
    pub fn register_metrics(&self, registry: &Registry) -> Result<()>;
}

So what happens when you create a get request? First all the providers in the initial set are queried with the have request. As an optimization, in every batch of queries a block request is sent instead. If the get query finds a block it returns a query complete. If the block wasn't found in the initial set, a Providers event is emitted. This is where the bitswap consumer tries to locate providers by for example performing a dht lookup. After the locating of providers completes, it is signaled by calling inject_providers. The query manager then performs bitswap requests using the new provider set which results in the block being found or a BlockNotFound error.

Often we want to sync an entire dag of blocks. We can efficiently sync dags of blocks by adding a sync query that runs get queries in parallel for all the references of a block. The set of providers that had a block is used as the initial set in a reference query.

License

MIT OR Apache-2.0