Bitswap: move providing -> Exchange-layer, providerQueryManager -> ro…

…uting (#641) Bitswap: refactor content provider elements This PR performs a rather large and touchy refactor of things related to Content providing and Content discovery previously embedded into Bitswap. The motivations: * Make ProviderQueryManager options configurable * Align and separate coalesced layers: content routing must not be part of bitswap as in the future we will be using different exchanges (bitswap, http) for retrieval and content routing should be above exchange. * Align content routing interfaces with libp2p: to avoid crust, wrappers and user confusion, align Providers and Discovery types to libp2p.ContentRouting. * Reduce duplicated functionality: i.e. code that handles providing in multiple places and fails to take advantage of ProvideMany optimizations. As a result: * ProviderQueryManager is now part of the routing module * A new providing.Exchange has been created * Bitswap initialization params have changed and Bitswap Network doesn't provide anymore (see changelog for more details) Co-authored-by: Hector Sanjuan <[email protected]> Co-authored-by: Andrew Gillis <[email protected]>
ipfs · Nov 25, 2024 · 37756ce · 37756ce
1 parent c91cc1d
commit 37756ce
Show file tree

Hide file tree

Showing 35 changed files with 725 additions and 550 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -14,6 +14,52 @@ The following emojis are used to highlight certain changes:
 
 ## [Unreleased]
 
+- `bitswap`, `routing`, `exchange` ([#641](https://github.com/ipfs/boxo/pull/641)):
+  - ✨ Bitswap is no longer in charge of providing blocks to the newtork: providing functionality is now handled by a `exchange/providing.Exchange`, meant to be used with `provider.System` so that all provides follow the same rules (multiple parts of the code where handling provides) before.
+  - 🛠 `bitswap/client/internal/providerquerymanager` has been moved to `routing/providerquerymanager` where it belongs. In order to keep compatibility, Bitswap now receives a `routing.ContentDiscovery` parameter which implements `FindProvidersAsync(...)` and uses it to create a `providerquerymanager` with the default settings as before. Custom settings can be used by using a custom `providerquerymanager` to manually wrap a `ContentDiscovery` object and pass that in as `ContentDiscovery` on initialization while setting `bitswap.WithDefaultProviderQueryManager(false)` (to avoid re-wrapping it again).
+  - The renovated `providedQueryManager` will trigger lookups until it manages to connect to `MaxProviders`. Before it would lookup at most `MaxInProcessRequests*MaxProviders` and connection failures may have limited the actual number of providers found.
+  - 🛠 We have aligned our routing-related interfaces with the libp2p [`routing`](https://pkg.go.dev/github.com/libp2p/go-libp2p/core/routing#ContentRouting) ones, including in the `reprovider.System`.
+  - In order to obtain exactly the same behaviour as before (i.e. particularly ensuring that new blocks are still provided), what was done like:
+
+```go
+	bswapnet := network.NewFromIpfsHost(host, contentRouter)
+	bswap := bitswap.New(p.ctx, bswapnet, blockstore)
+	bserv = blockservice.New(blockstore, bswap)
+```
+  - becomes:
+
+```go
+	// Create network: no contentRouter anymore
+	bswapnet := network.NewFromIpfsHost(host)
+	// Create Bitswap: a new "discovery" parameter, usually the "contentRouter"
+	// which does both discovery and providing.
+	bswap := bitswap.New(p.ctx, bswapnet, discovery, blockstore)
+	// A provider system that handles concurrent provides etc. "contentProvider"
+	// is usually the "contentRouter" which does both discovery and providing.
+	// "contentProvider" could be used directly without wrapping, but it is recommended
+	// to do so to provide more efficiently.
+	provider := provider.New(datastore, provider.Online(contentProvider)
+	// A wrapped providing exchange using the previous exchange and the provider.
+	exch := providing.New(bswap, provider)
+
+	// Finally the blockservice
+	bserv := blockservice.New(blockstore, exch)
+	...
+```
+
+  - The above is only necessary if content routing is needed. Otherwise:
+
+```
+	// Create network: no contentRouter anymore
+	bswapnet := network.NewFromIpfsHost(host)
+	// Create Bitswap: a new "discovery" parameter set to nil (disable content discovery)
+	bswap := bitswap.New(p.ctx, bswapnet, nil, blockstore)
+	// Finally the blockservice
+	bserv := blockservice.New(blockstore, exch)
+```
+
+
+
 ### Added
 
 - `routing/http/server`: added built-in Prometheus instrumentation to http delegated `/routing/v1/` endpoints, with custom buckets for response size and duration to match real world data observed at [the `delegated-ipfs.dev` instance](https://docs.ipfs.tech/concepts/public-utilities/#delegated-routing). [#718](https://github.com/ipfs/boxo/pull/718) [#724](https://github.com/ipfs/boxo/pull/724)
@@ -117,6 +163,7 @@ The following emojis are used to highlight certain changes:
 - `bitswap/client` fix memory leak in BlockPresenceManager due to unlimited map growth. [#636](https://github.com/ipfs/boxo/pull/636)
 - `bitswap/network` fixed race condition when a timeout occurred before hole punching completed while establishing a first-time stream to a peer behind a NAT [#651](https://github.com/ipfs/boxo/pull/651)
 - `bitswap`: wantlist overflow handling now cancels existing entries to make room for newer entries. This fix prevents the wantlist from filling up with CIDs that the server does not have. [#629](https://github.com/ipfs/boxo/pull/629)
+- 🛠 `bitswap` & `bitswap/server` no longer provide to content routers, instead you can use the `provider` package because it uses a datastore queue and batches calls to ProvideMany.
 
 ## [v0.21.0]
 

diff --git a/bitswap/benchmarks_test.go b/bitswap/benchmarks_test.go
@@ -12,17 +12,16 @@ import (
 	"testing"
 	"time"
 
-	blocks "github.com/ipfs/go-block-format"
-	"github.com/ipfs/go-test/random"
-	protocol "github.com/libp2p/go-libp2p/core/protocol"
-
 	"github.com/ipfs/boxo/bitswap"
 	bsnet "github.com/ipfs/boxo/bitswap/network"
 	testinstance "github.com/ipfs/boxo/bitswap/testinstance"
 	tn "github.com/ipfs/boxo/bitswap/testnet"
 	mockrouting "github.com/ipfs/boxo/routing/mock"
+	blocks "github.com/ipfs/go-block-format"
 	cid "github.com/ipfs/go-cid"
 	delay "github.com/ipfs/go-ipfs-delay"
+	"github.com/ipfs/go-test/random"
+	protocol "github.com/libp2p/go-libp2p/core/protocol"
 )
 
 type fetchFunc func(b *testing.B, bs *bitswap.Bitswap, ks []cid.Cid)
@@ -135,24 +134,25 @@ func BenchmarkFetchFromOldBitswap(b *testing.B) {
 	benchmarkLog = nil
 	fixedDelay := delay.Fixed(10 * time.Millisecond)
 	bstoreLatency := time.Duration(0)
+	router := mockrouting.NewServer()
 
 	for _, bch := range mixedBenches {
 		b.Run(bch.name, func(b *testing.B) {
 			fetcherCount := bch.fetcherCount
 			oldSeedCount := bch.oldSeedCount
 			newSeedCount := bch.nodeCount - (fetcherCount + oldSeedCount)
 
-			net := tn.VirtualNetwork(mockrouting.NewServer(), fixedDelay)
+			net := tn.VirtualNetwork(fixedDelay)
 
 			// Simulate an older Bitswap node (old protocol ID) that doesn't
 			// send DONT_HAVE responses
 			oldProtocol := []protocol.ID{bsnet.ProtocolBitswapOneOne}
 			oldNetOpts := []bsnet.NetOpt{bsnet.SupportedProtocols(oldProtocol)}
 			oldBsOpts := []bitswap.Option{bitswap.SetSendDontHaves(false)}
-			oldNodeGenerator := testinstance.NewTestInstanceGenerator(net, oldNetOpts, oldBsOpts)
+			oldNodeGenerator := testinstance.NewTestInstanceGenerator(net, router, oldNetOpts, oldBsOpts)
 
 			// Regular new Bitswap node
-			newNodeGenerator := testinstance.NewTestInstanceGenerator(net, nil, nil)
+			newNodeGenerator := testinstance.NewTestInstanceGenerator(net, router, nil, nil)
 			var instances []testinstance.Instance
 
 			// Create new nodes (fetchers + seeds)
@@ -294,9 +294,10 @@ func BenchmarkDatacenterMultiLeechMultiSeed(b *testing.B) {
 		numblks := 1000
 
 		for i := 0; i < b.N; i++ {
-			net := tn.RateLimitedVirtualNetwork(mockrouting.NewServer(), d, rateLimitGenerator)
+			net := tn.RateLimitedVirtualNetwork(d, rateLimitGenerator)
 
-			ig := testinstance.NewTestInstanceGenerator(net, nil, nil)
+			router := mockrouting.NewServer()
+			ig := testinstance.NewTestInstanceGenerator(net, router, nil, nil)
 			defer ig.Close()
 
 			instances := ig.Instances(numnodes)
@@ -312,9 +313,9 @@ func BenchmarkDatacenterMultiLeechMultiSeed(b *testing.B) {
 
 func subtestDistributeAndFetch(b *testing.B, numnodes, numblks int, d delay.D, bstoreLatency time.Duration, df distFunc, ff fetchFunc) {
 	for i := 0; i < b.N; i++ {
-		net := tn.VirtualNetwork(mockrouting.NewServer(), d)
-
-		ig := testinstance.NewTestInstanceGenerator(net, nil, nil)
+		net := tn.VirtualNetwork(d)
+		router := mockrouting.NewServer()
+		ig := testinstance.NewTestInstanceGenerator(net, router, nil, nil)
 
 		instances := ig.Instances(numnodes)
 		rootBlock := random.BlocksOfSize(1, rootBlockSize)
@@ -327,9 +328,9 @@ func subtestDistributeAndFetch(b *testing.B, numnodes, numblks int, d delay.D, b
 
 func subtestDistributeAndFetchRateLimited(b *testing.B, numnodes, numblks int, d delay.D, rateLimitGenerator tn.RateLimitGenerator, blockSize int64, bstoreLatency time.Duration, df distFunc, ff fetchFunc) {
 	for i := 0; i < b.N; i++ {
-		net := tn.RateLimitedVirtualNetwork(mockrouting.NewServer(), d, rateLimitGenerator)
-
-		ig := testinstance.NewTestInstanceGenerator(net, nil, nil)
+		net := tn.RateLimitedVirtualNetwork(d, rateLimitGenerator)
+		router := mockrouting.NewServer()
+		ig := testinstance.NewTestInstanceGenerator(net, router, nil, nil)
 		defer ig.Close()
 
 		instances := ig.Instances(numnodes)
@@ -437,7 +438,7 @@ func runDistribution(b *testing.B, instances []testinstance.Instance, blocks []b
 
 func allToAll(b *testing.B, provs []testinstance.Instance, blocks []blocks.Block) {
 	for _, p := range provs {
-		if err := p.Blockstore().PutMany(context.Background(), blocks); err != nil {
+		if err := p.Blockstore.PutMany(context.Background(), blocks); err != nil {
 			b.Fatal(err)
 		}
 	}
@@ -452,10 +453,10 @@ func overlap1(b *testing.B, provs []testinstance.Instance, blks []blocks.Block)
 	bill := provs[0]
 	jeff := provs[1]
 
-	if err := bill.Blockstore().PutMany(context.Background(), blks[:75]); err != nil {
+	if err := bill.Blockstore.PutMany(context.Background(), blks[:75]); err != nil {
 		b.Fatal(err)
 	}
-	if err := jeff.Blockstore().PutMany(context.Background(), blks[25:]); err != nil {
+	if err := jeff.Blockstore.PutMany(context.Background(), blks[25:]); err != nil {
 		b.Fatal(err)
 	}
 }
@@ -473,12 +474,12 @@ func overlap2(b *testing.B, provs []testinstance.Instance, blks []blocks.Block)
 		even := i%2 == 0
 		third := i%3 == 0
 		if third || even {
-			if err := bill.Blockstore().Put(context.Background(), blk); err != nil {
+			if err := bill.Blockstore.Put(context.Background(), blk); err != nil {
 				b.Fatal(err)
 			}
 		}
 		if third || !even {
-			if err := jeff.Blockstore().Put(context.Background(), blk); err != nil {
+			if err := jeff.Blockstore.Put(context.Background(), blk); err != nil {
 				b.Fatal(err)
 			}
 		}
@@ -490,7 +491,7 @@ func overlap2(b *testing.B, provs []testinstance.Instance, blks []blocks.Block)
 // but we're mostly just testing performance of the sync algorithm
 func onePeerPerBlock(b *testing.B, provs []testinstance.Instance, blks []blocks.Block) {
 	for _, blk := range blks {
-		err := provs[rand.Intn(len(provs))].Blockstore().Put(context.Background(), blk)
+		err := provs[rand.Intn(len(provs))].Blockstore.Put(context.Background(), blk)
 		if err != nil {
 			b.Fatal(err)
 		}

diff --git a/bitswap/bitswap.go b/bitswap/bitswap.go
@@ -5,7 +5,6 @@ import (
 	"fmt"
 
 	"github.com/ipfs/boxo/bitswap/client"
-	"github.com/ipfs/boxo/bitswap/internal/defaults"
 	"github.com/ipfs/boxo/bitswap/message"
 	"github.com/ipfs/boxo/bitswap/network"
 	"github.com/ipfs/boxo/bitswap/server"
@@ -45,9 +44,8 @@ type bitswap interface {
 }
 
 var (
-	_                  exchange.SessionExchange = (*Bitswap)(nil)
-	_                  bitswap                  = (*Bitswap)(nil)
-	HasBlockBufferSize                          = defaults.HasBlockBufferSize
+	_ exchange.SessionExchange = (*Bitswap)(nil)
+	_ bitswap                  = (*Bitswap)(nil)
 )
 
 type Bitswap struct {
@@ -58,7 +56,7 @@ type Bitswap struct {
 	net    network.BitSwapNetwork
 }
 
-func New(ctx context.Context, net network.BitSwapNetwork, bstore blockstore.Blockstore, options ...Option) *Bitswap {
+func New(ctx context.Context, net network.BitSwapNetwork, providerFinder client.ProviderFinder, bstore blockstore.Blockstore, options ...Option) *Bitswap {
 	bs := &Bitswap{
 		net: net,
 	}
@@ -85,14 +83,10 @@ func New(ctx context.Context, net network.BitSwapNetwork, bstore blockstore.Bloc
 		serverOptions = append(serverOptions, server.WithTracer(tracer))
 	}
 
-	if HasBlockBufferSize != defaults.HasBlockBufferSize {
-		serverOptions = append(serverOptions, server.HasBlockBufferSize(HasBlockBufferSize))
-	}
-
 	ctx = metrics.CtxSubScope(ctx, "bitswap")
 
 	bs.Server = server.New(ctx, net, bstore, serverOptions...)
-	bs.Client = client.New(ctx, net, bstore, append(clientOptions, client.WithBlockReceivedNotifier(bs.Server))...)
+	bs.Client = client.New(ctx, net, providerFinder, bstore, append(clientOptions, client.WithBlockReceivedNotifier(bs.Server))...)
 	net.Start(bs) // use the polyfill receiver to log received errors and trace messages only once
 
 	return bs
@@ -115,7 +109,6 @@ type Stat struct {
 	MessagesReceived uint64
 	BlocksSent       uint64
 	DataSent         uint64
-	ProvideBufLen    int
 }
 
 func (bs *Bitswap) Stat() (*Stat, error) {
@@ -138,7 +131,6 @@ func (bs *Bitswap) Stat() (*Stat, error) {
 		Peers:            ss.Peers,
 		BlocksSent:       ss.BlocksSent,
 		DataSent:         ss.DataSent,
-		ProvideBufLen:    ss.ProvideBufLen,
 	}, nil
 }