以太坊p2p原理与实现

区块链技术的去中心依赖于底层组网技术，以太坊的底层实现了p2pServer，大约可以分为这样三层。

• 底层路由表。封装了kad路由，节点的数据结构以及计算记录，节点搜索，验证等功能。
• 中层peer抽象，message开放发送接口，server对外提供peer检测，初始化，事件订阅，peer状态查询，启动，停止等功能
• 以太坊最上层peer，peerset再封装，通过协议的Run函数，在中层启动peer时，获取peer，最终通过一个循环截取稳定peer，包装在peerset中使用。

底层路由表

这里简化问题仅讨论Node Discovery Protocol。 这一层维护了一个buckets桶，总共有17个桶，每个桶有16个节点和10个替换节点。 Node放入时先要计算hash和localNode的距离。再按距离选择一个桶放进去，取的时候逐个计算target和每个桶中对象的举例，详细参考closest函数，后面会贴出来。

距离公式满足：f(x,y)=256-8*n-map(x[n+1]^y[n+1]) 注：n为相同节点数量 map为一个负相关的映射关系。

简单来说就是相似越多，值越小。细节参考Node.go的logdist函数。 这里需要了解算法Kademlia，

.
├── database.go         //封装node数据库相关操作
├── node.go             //节点数据结构
├── ntp.go              //同步时间
├── table.go            //路由表
├── udp.go              //网络相关操作

其中最重要的就是table对象，table公共方法有：

• newTable 实例创建
• Self local节点获取
• ReadRandomNodes 随机读取几个节点
• Close 关闭
• Resolve 在周边查找某个节点
• Lookup 查找某个节点的邻近节点

逐个来分析这些方法：

newTable

• 1：生成对象实例（获取数据库客户端，LocalNode etc）

    // If no node database was given, use an in-memory onedb, err := newNodeDB(nodeDBPath, Version, ourID)if err != nil {return nil, err}tab := &Table{net:        t,db:         db,self:       NewNode(ourID, ourAddr.IP, uint16(ourAddr.Port), uint16(ourAddr.Port)),bonding:    make(map[NodeID]*bondproc),bondslots:  make(chan struct{}, maxBondingPingPongs),refreshReq: make(chan chan struct{}),initDone:   make(chan struct{}),closeReq:   make(chan struct{}),closed:     make(chan struct{}),rand:       mrand.New(mrand.NewSource(0)),ips:        netutil.DistinctNetSet{Subnet: tableSubnet, Limit: tableIPLimit},}

• 2：载入引导节点，初始化k桶。

    if err := tab.setFallbackNodes(bootnodes); err != nil {return nil, err}for i := 0; i < cap(tab.bondslots); i++ {tab.bondslots <- struct{}{}}for i := range tab.buckets {tab.buckets[i] = &bucket{ips: netutil.DistinctNetSet{Subnet: bucketSubnet, Limit: bucketIPLimit},}}

• 3：将节点放入到桶里，生成一条协程用于刷新，验证节点。

    tab.seedRand()tab.loadSeedNodes(false)  //载入种子节点// Start the background expiration goroutine after loading seeds so that the search for// seed nodes also considers older nodes that would otherwise be removed by the// expiration.tab.db.ensureExpirer()go tab.loop()

载入种子节点

    func (tab *Table) loadSeedNodes(bond bool) {seeds := tab.db.querySeeds(seedCount, seedMaxAge)//数据库中的种子节点和引导节点合并seeds = append(seeds, tab.nursery...) if bond {seeds = tab.bondall(seeds)   //节点验证}for i := range seeds {seed := seeds[i]age := log.Lazy{Fn: func() interface{} { return time.Since(tab.db.bondTime(seed.ID)) }}log.Debug("Found seed node in database", "id", seed.ID, "addr", seed.addr(), "age", age)tab.add(seed)               //节点入桶}}

节点入桶，同时也要检查ip等限制。

    func (tab *Table) add(new *Node) {tab.mutex.Lock()defer tab.mutex.Unlock()b := tab.bucket(new.sha)   //获取当前节点对应的桶if !tab.bumpOrAdd(b, new) {// Node is not in table. Add it to the replacement list.tab.addReplacement(b, new)}}

桶的选择

    func (tab *Table) bucket(sha common.Hash) *bucket {d := logdist(tab.self.sha, sha)  //计算hash举例if d <= bucketMinDistance {//这里按算法来看，只要hash前三位相等就会到第一个bucketsreturn tab.buckets[0]}return tab.buckets[d-bucketMinDistance-1]}

Resolve

根据Node的Id查找Node，先在当前的桶里面查找，查找一遍之后没找到就在周边的节点里面搜索一遍再找。

    // Resolve searches for a specific node with the given ID.// It returns nil if the node could not be found.func (tab *Table) Resolve(targetID NodeID) *Node {// If the node is present in the local table, no// network interaction is required.hash := crypto.Keccak256Hash(targetID[:])tab.mutex.Lock()//查找最近节点cl := tab.closest(hash, 1)tab.mutex.Unlock()if len(cl.entries) > 0 && cl.entries[0].ID == targetID {return cl.entries[0]}// Otherwise, do a network lookup.//不存在 搜索邻居节点result := tab.Lookup(targetID)for _, n := range result {if n.ID == targetID {return n}}return nil}

这里需要理解的函数是 closest，遍历所有桶的所有节点，查找最近的一个

    // closest returns the n nodes in the table that are closest to the// given id. The caller must hold tab.mutex.func (tab *Table) closest(target common.Hash, nresults int) *nodesByDistance {// This is a very wasteful way to find the closest nodes but// obviously correct. I believe that tree-based buckets would make// this easier to implement efficiently.close := &nodesByDistance{target: target}for _, b := range tab.buckets {for _, n := range b.entries {close.push(n, nresults)}}return close}func (h *nodesByDistance) push(n *Node, maxElems int) {ix := sort.Search(len(h.entries), func(i int) bool {return distcmp(h.target, h.entries[i].sha, n.sha) > 0})if len(h.entries) < maxElems {h.entries = append(h.entries, n)}if ix == len(h.entries) {// farther away than all nodes we already have.// if there was room for it, the node is now the last element.} else {// slide existing entries down to make room// this will overwrite the entry we just appended.//近的靠前边copy(h.entries[ix+1:], h.entries[ix:])h.entries[ix] = n}}

ReadRandomNodes

整体思路是先拷贝出来，再逐个桶的抽最上面的一个，剩下空桶移除，剩下的桶合并后，下一轮再抽桶的第一个节点，直到填满给定数据或者桶全部空掉。最后返回填到数组里面的数量。

    // ReadRandomNodes fills the given slice with random nodes from the// table. It will not write the same node more than once. The nodes in// the slice are copies and can be modified by the caller.func (tab *Table) ReadRandomNodes(buf []*Node) (n int) {if !tab.isInitDone() {return 0}tab.mutex.Lock()defer tab.mutex.Unlock()// Find all non-empty buckets and get a fresh slice of their entries.var buckets [][]*Node//拷贝节点for _, b := range tab.buckets {if len(b.entries) > 0 {buckets = append(buckets, b.entries[:])}}if len(buckets) == 0 {return 0}// Shuffle the buckets.for i := len(buckets) - 1; i > 0; i-- {j := tab.rand.Intn(len(buckets))buckets[i], buckets[j] = buckets[j], buckets[i]}// Move head of each bucket into buf, removing buckets that become empty.var i, j intfor ; i < len(buf); i, j = i+1, (j+1)%len(buckets) {b := buckets[j]buf[i] = &(*b[0])  //取第一个节点buckets[j] = b[1:] //移除第一个if len(b) == 1 {//空桶移除buckets = append(buckets[:j], buckets[j+1:]...)  }if len(buckets) == 0 {break          }}return i + 1}

Lookup

lookup会要求已知节点查找邻居节点，查找的邻居节点又递归的找它周边的节点

    for {// ask the alpha closest nodes that we haven't asked yetfor i := 0; i < len(result.entries) && pendingQueries < alpha; i++ {n := result.entries[i]if !asked[n.ID] {asked[n.ID] = truependingQueries++   go func() {// Find potential neighbors to bond withr, err := tab.net.findnode(n.ID, n.addr(), targetID)if err != nil {// Bump the failure counter to detect and evacuate non-bonded entriesfails := tab.db.findFails(n.ID) + 1tab.db.updateFindFails(n.ID, fails)log.Trace("Bumping findnode failure counter", "id", n.ID, "failcount", fails)if fails >= maxFindnodeFailures {log.Trace("Too many findnode failures, dropping", "id", n.ID, "failcount", fails)tab.delete(n)}}reply <- tab.bondall(r)}()}}if pendingQueries == 0 {// we have asked all closest nodes, stop the searchbreak}// wait for the next replyfor _, n := range <-reply {    //此处会阻塞请求if n != nil && !seen[n.ID] {seen[n.ID] = trueresult.push(n, bucketSize)}}pendingQueries--}

桶的维护

桶初始化完成后会进入一个循环逻辑，其中通过三个timer控制调整周期。

• 验证timer 间隔 10s左右
• 刷新timer 间隔 30 min
• 持久化timer 间隔 30s

    revalidate     = time.NewTimer(tab.nextRevalidateTime())refresh        = time.NewTicker(refreshInterval)copyNodes      = time.NewTicker(copyNodesInterval)

刷新逻辑:重新加载种子节点，查找周边节点，随机三个节点，并查找这三个节点的周围节点。

    func (tab *Table) doRefresh(done chan struct{}) {defer close(done)tab.loadSeedNodes(true)tab.lookup(tab.self.ID, false)for i := 0; i < 3; i++ {var target NodeIDcrand.Read(target[:])tab.lookup(target, false)}}

验证逻辑:验证每个桶的最末尾节点，如果该节点通过验证则放到队首（验证过程是本地节点向它发送ping请求，如果回应pong则通过）

    last, bi := tab.nodeToRevalidate()  //取最后一个节点if last == nil {// No non-empty bucket found.return}// Ping the selected node and wait for a pong.err := tab.ping(last.ID, last.addr())   //通信验证tab.mutex.Lock()defer tab.mutex.Unlock()b := tab.buckets[bi]if err == nil {// The node responded, move it to the front.log.Debug("Revalidated node", "b", bi, "id", last.ID)b.bump(last)    //提到队首return}

Peer/Server

相关文件

.
├── dial.go          //封装一个任务生成处理结构以及三种任务结构中（此处命名不太精确）
├── message.go       //定义一些数据的读写接口，以及对外的Send/SendItem函数
├── peer.go          //封装了Peer 包括消息读取
├── rlpx.go          //内部的握手协议
├── server.go        //初始化，维护Peer网络，还有一些对外的接口

这一层会不断的从路由中提取节点，提取出来的节点要经过身份验证，协议检查之后加入到peer里面，紧接着如果没有人使用这个peer，这个peer就会被删除，再重新选择一些节点出来继续这个流程，peer再其中是随生随销，这样做是为了平均的使用所有的节点，而不是仅仅依赖于特定的几个节点。因而这里从Server开始入手分析整个流程

    Peers()                             //peer对象PeerCount()                         //peer数量AddPeer(node *discover.Node)        //添加节点RemovePeer(node *discover.Node)     //删除节点SubscribeEvents(ch chan *PeerEvent) //订阅内部的事件（节点的增加，删除）//以上四个属于对外的接口，不影响内部逻辑Start()                             //server开始工作SetupConn(fd net.Conn, flags connFlag, dialDest *discover.Node)  //启动一个连接，经过两次验证之后，如果通过则加入到peer之中。

Start初始化

Start做了三件事，生成路由表于建立底层网络。生成DialState用于驱动维护本地peer的更新与死亡，监听本地接口用于信息应答。这里主要分析peer的维护过程。函数是run函数。

    func (srv *Server) Start() (err error) {//**************初始化代码省略if !srv.NoDiscovery && srv.DiscoveryV5 {unhandled = make(chan discover.ReadPacket, 100)sconn = &sharedUDPConn{conn, unhandled}}// node tableif !srv.NoDiscovery {//路由表生成cfg := discover.Config{PrivateKey:   srv.PrivateKey,AnnounceAddr: realaddr,NodeDBPath:   srv.NodeDatabase,NetRestrict:  srv.NetRestrict,Bootnodes:    srv.BootstrapNodes,Unhandled:    unhandled,}ntab, err := discover.ListenUDP(conn, cfg)if err != nil {return err}srv.ntab = ntab}if srv.DiscoveryV5 {//路由表生成var (ntab *discv5.Networkerr  error)if sconn != nil {ntab, err = discv5.ListenUDP(srv.PrivateKey, sconn, realaddr, "", srv.NetRestrict) //srv.NodeDatabase)} else {ntab, err = discv5.ListenUDP(srv.PrivateKey, conn, realaddr, "", srv.NetRestrict) //srv.NodeDatabase)}if err != nil {return err}if err := ntab.SetFallbackNodes(srv.BootstrapNodesV5); err != nil {return err}srv.DiscV5 = ntab}dynPeers := srv.maxDialedConns()//newDialState 对象生成，这个对象包含Peer的实际维护代码dialer := newDialState(srv.StaticNodes, srv.BootstrapNodes, srv.ntab, dynPeers, srv.NetRestrict)// handshake  协议加载srv.ourHandshake = &protoHandshake{Version: baseProtocolVersion, Name: srv.Name, ID: discover.PubkeyID(&srv.PrivateKey.PublicKey)}for _, p := range srv.Protocols {srv.ourHandshake.Caps = append(srv.ourHandshake.Caps, p.cap())}// listen/dial//监听本地端口if srv.ListenAddr != "" {if err := srv.startListening(); err != nil {return err}}if srv.NoDial && srv.ListenAddr == "" {srv.log.Warn("P2P server will be useless, neither dialing nor listening")}srv.loopWG.Add(1)//重要的一句，开个协程，在其中做peer的维护go srv.run(dialer)srv.running = truereturn nil}

run 开始peer的生成

该函数中定义了两个队列

    runningTasks []task //正在执行的任务queuedTasks  []task //尚未执行的任务

定义了三个匿名函数

    //从正在执行任务中删除任务delTask := func(t task) {for i := range runningTasks {if runningTasks[i] == t {runningTasks = append(runningTasks[:i], runningTasks[i+1:]...)break}}}//开始一批任务startTasks := func(ts []task) (rest []task) {i := 0for ; len(runningTasks) < maxActiveDialTasks && i < len(ts); i++ {t := ts[i]srv.log.Trace("New dial task", "task", t)go func() {t.Do(srv); taskdone <- t  }()runningTasks = append(runningTasks, t)}return ts[i:]}//启动开始一批任务再调用dialstate的newTasks函数生成一批任务，加载到任务队列里面scheduleTasks := func() {// Start from queue first.queuedTasks = append(queuedTasks[:0], startTasks(queuedTasks)...)// Query dialer for new tasks and start as many as possible now.if len(runningTasks) < maxActiveDialTasks {nt := dialstate.newTasks(len(runningTasks)+len(queuedTasks), peers, time.Now())queuedTasks = append(queuedTasks, startTasks(nt)...)}}

定义了一个循环，分不同的chanel执行对应的逻辑

   for {//调度开始找生成任务scheduleTasks()select {case <-srv.quit://退出break runningcase n := <-srv.addstatic: //增加一个节点  该节点最终会生成一个dialTask //并在newTasks的时候加入到读列srv.log.Debug("Adding static node", "node", n)dialstate.addStatic(n)case n := <-srv.removestatic://直接删除该节点 节点不再参与维护，很快就会死掉了dialstate.removeStatic(n)if p, ok := peers[n.ID]; ok {p.Disconnect(DiscRequested)}case op := <-srv.peerOp://  Peers 和 PeerCount 两个外部接口，只是读取peer信息op(peers)srv.peerOpDone <- struct{}{}case t := <-taskdone://task完成后会根据不同的任务类型进行相应的处理srv.log.Trace("Dial task done", "task", t)dialstate.taskDone(t, time.Now())delTask(t)case c := <-srv.posthandshake://身份验证通过 if trusted[c.id] {// Ensure that the trusted flag is set before checking against MaxPeers.c.flags |= trustedConn}select {case c.cont <- srv.encHandshakeChecks(peers, inboundCount, c):case <-srv.quit:break running}case c := <-srv.addpeer://身份协议验证通过 加入队列err := srv.protoHandshakeChecks(peers, inboundCount, c)if err == nil {// The handshakes are done and it passed all checks.p := newPeer(c, srv.Protocols)// If message events are enabled, pass the peerFeed// to the peerif srv.EnableMsgEvents {p.events = &srv.peerFeed}name := truncateName(c.name)srv.log.Debug("Adding p2p peer", "name", name, "addr", c.fd.RemoteAddr(), "peers", len(peers)+1)go srv.runPeer(p)   //触发事件 此处是最上层截取peer的位置，如果此物没有外部影响，那么这个peer很快就被销毁了peerAdd++fmt.Printf("--count %d--- add %d-- del %d--\n",len(peers),peerAdd,peerDel)peers[c.id] = pif p.Inbound() {inboundCount++}}// The dialer logic relies on the assumption that// dial tasks complete after the peer has been added or// discarded. Unblock the task last.select {case c.cont <- err:case <-srv.quit:break running}case pd := <-srv.delpeer://移除peerd := common.PrettyDuration(mclock.Now() - pd.created)pd.log.Debug("Removing p2p peer", "duration", d, "peers", len(peers)-1, "req", pd.requested, "err", pd.err)delete(peers, pd.ID())peerDel++fmt.Printf("--count %d--- add %d-- del %d--\n",len(peers),peerAdd,peerDel)if pd.Inbound() {inboundCount--}}}

记住上面的代码，再来逐个的看：

scheduleTasks

scheduleTasks调度生成任务，生成的任务中有一种dialTask的任务，该任务结构如下

    type dialTask struct {flags        connFlagdest         *discover.NodelastResolved time.TimeresolveDelay time.Duration}func (t *dialTask) Do(srv *Server) {if t.dest.Incomplete() {if !t.resolve(srv) {return}}err := t.dial(srv, t.dest)  //此处会调用到setupConn函数if err != nil {log.Trace("Dial error", "task", t, "err", err)// Try resolving the ID of static nodes if dialing failed.if _, ok := err.(*dialError); ok && t.flags&staticDialedConn != 0 {if t.resolve(srv) {t.dial(srv, t.dest)}}}}

dial最终回调用到setupConn函数，函数只保留重点的几句，篇幅有点长了

    func (srv *Server) setupConn(c *conn, flags connFlag, dialDest *discover.Node) error {//身份验证码 获取设备，标识等信息if c.id, err = c.doEncHandshake(srv.PrivateKey, dialDest); err != //此处会往chanel中添加连接对象，最终触发循环中的posthandshake分支err = srv.checkpoint(c, srv.posthandshake)  //协议验证phs, err := c.doProtoHandshake(srv.ourHandshake)c.caps, c.name = phs.Caps, phs.Name//此处会往chanel中添加连接对象 最终触发循环中的addpeer分支err = srv.checkpoint(c, srv.addpeer)}

posthandshake 分支仅仅做了验证，addpeer做的事情就比较多了，重要的就是执行runPeer函数

    func (srv *Server) runPeer(p *Peer) {// 广播 peer addsrv.peerFeed.Send(&PeerEvent{Type: PeerEventTypeAdd,Peer: p.ID(),})// run the protocolremoteRequested, err := p.run() //// 广播 peer dropsrv.peerFeed.Send(&PeerEvent{Type:  PeerEventTypeDrop,Peer:  p.ID(),Error: err.Error(),})//移除peersrv.delpeer <- peerDrop{p, err, remoteRequested}}func (p *Peer) run() (remoteRequested bool, err error) {//*************writeStart <- struct{}{}p.startProtocols(writeStart, writeErr)//*************//这一句阻塞性确保了peer的存活p.wg.Wait()  }func (p *Peer) startProtocols(writeStart <-chan struct{}, writeErr chan<- error) {p.wg.Add(len(p.running))for _, proto := range p.running {proto := protoproto.closed = p.closedproto.wstart = writeStartproto.werr = writeErrvar rw MsgReadWriter = protoif p.events != nil {rw = newMsgEventer(rw, p.events, p.ID(), proto.Name)}p.log.Trace(fmt.Sprintf("Starting protocol %s/%d", proto.Name, proto.Version))go func() {//其他的都是为这一句做准备的，在以太坊中p2p就是靠这一句对上层暴露peer对象err := proto.Run(p, rw)if err == nil {p.log.Trace(fmt.Sprintf("Protocol %s/%d returned", proto.Name, proto.Version))err = errProtocolReturned} else if err != io.EOF {p.log.Trace(fmt.Sprintf("Protocol %s/%d failed", proto.Name, proto.Version), "err", err)}p.protoErr <- errp.wg.Done()  }()}}

这样就可以可理出一条思路 scheduleTasks执行生成dialTask任务 dialTask任务执行过程中逐个填充posthandshake，addPeer这两个chanel。 addPeer执行时对上层暴露了Peer对象，完成后填充了delpeer，最后删除了Peer。

任务的生成

具体看代码中的注释

    func (s *dialstate) newTasks(nRunning int, peers map[discover.NodeID]*Peer, now time.Time) []task {if s.start.IsZero() {s.start = now}var newtasks []task//这里声明了一个添加任务的函数  addDial := func(flag connFlag, n *discover.Node) bool {if err := s.checkDial(n, peers); err != nil {log.Trace("Skipping dial candidate", "id", n.ID, "addr", &net.TCPAddr{IP: n.IP, Port: int(n.TCP)}, "err", err)return false}s.dialing[n.ID] = flag  //排除掉已经再测试的newtasks = append(newtasks, &dialTask{flags: flag, dest: n})return true}// Compute number of dynamic dials necessary at this point.needDynDials := s.maxDynDials     //当前系统中最大连接数目for _, p := range peers {        //扣除已建立链接的peerif p.rw.is(dynDialedConn) {needDynDials--}}for _, flag := range s.dialing {  //扣除已建立链接的peerif flag&dynDialedConn != 0 {needDynDials--}}//外部命令添加的节点 这种节点不占用needDynDials数目，//是为了保证手动加的节点能够起效for id, t := range s.static {err := s.checkDial(t.dest, peers)switch err {case errNotWhitelisted, errSelf:log.Warn("Removing static dial candidate", "id", t.dest.ID, "addr", &net.TCPAddr{IP: t.dest.IP, Port: int(t.dest.TCP)}, "err", err)delete(s.static, t.dest.ID)case nil:s.dialing[id] = t.flagsnewtasks = append(newtasks, t)}}// If we don't have any peers whatsoever, try to dial a random bootnode. This// scenario is useful for the testnet (and private networks) where the discovery// table might be full of mostly bad peers, making it hard to find good ones.if len(peers) == 0 && len(s.bootnodes) > 0 && needDynDials > 0 && //检查引导节点  因为引导节点比搜索到的节点更大概率靠谱 因而比较靠前now.Sub(s.start) > fallbackInterval {bootnode := s.bootnodes[0]s.bootnodes = append(s.bootnodes[:0], s.bootnodes[1:]...)s.bootnodes = append(s.bootnodes, bootnode)if addDial(dynDialedConn, bootnode) {needDynDials--}}//随机的从路由中抽取最大节点的二分之一randomCandidates := needDynDials / 2if randomCandidates > 0 {n := s.ntab.ReadRandomNodes(s.randomNodes)for i := 0; i < randomCandidates && i < n; i++ {if addDial(dynDialedConn, s.randomNodes[i]) {needDynDials--}}}// 从lookupbuf中抽取i := 0for ; i < len(s.lookupBuf) && needDynDials > 0; i++ {if addDial(dynDialedConn, s.lookupBuf[i]) {needDynDials--}}s.lookupBuf = s.lookupBuf[:copy(s.lookupBuf, s.lookupBuf[i:])]// 如果还是不够，路由再去搜索节点if len(s.lookupBuf) < needDynDials && !s.lookupRunning {s.lookupRunning = truenewtasks = append(newtasks, &discoverTask{})}// waitif nRunning == 0 && len(newtasks) == 0 && s.hist.Len() > 0 {t := &waitExpireTask{s.hist.min().exp.Sub(now)}newtasks = append(newtasks, t)}return newtasks}

消息发送

另一个是message中的Send，SendItem函数 实现了MsgWriter的对象都可以调用这个函数写入，觉得这里没什么必要，完全可以封装到peer里面去，不过它上层做广播的时候确实是调用的这两个函数。

    func Send(w MsgWriter, msgcode uint64, data interface{}) error {size, r, err := rlp.EncodeToReader(data)if err != nil {return err}return w.WriteMsg(Msg{Code: msgcode, Size: uint32(size), Payload: r})}func SendItems(w MsgWriter, msgcode uint64, elems ...interface{}) error {return Send(w, msgcode, elems)}

以太坊上层调用

Peer/PeerSet

文件：go-ethereum/eth/peer.go

定义了两个struct，Peer和PeerSet。Peer封装了底层的p2p.Peer,集成了一些和业务相关的方法，比如SendTransactions，SendNewBlock等。PeerSet是Peer的集合

    type peer struct {id string*p2p.Peerrw p2p.MsgReadWriterversion  int         // Protocol version negotiatedforkDrop *time.Timer // Timed connection dropper if forks aren't validated in timehead common.Hashtd   *big.Intlock sync.RWMutexknownTxs    *set.Set // Set of transaction hashes known to be known by this peerknownBlocks *set.Set // Set of block hashes known to be known by this peer}type peerSet struct {peers  map[string]*peerlock   sync.RWMutexclosed bool}

Peer注册/注销

文件：go-ethereum/eth/handler.go manager.handle在检查了peer后会把这个peer注册到peerset中，表示此peer可用，发生错误后peerset注销该peer，返回错误，最后再Server中销毁。

    manager.SubProtocols = make([]p2p.Protocol, 0, len(ProtocolVersions))for i, version := range ProtocolVersions {// Skip protocol version if incompatible with the mode of operationif mode == downloader.FastSync && version < eth63 {continue}// Compatible; initialise the sub-protocolversion := version // Closure for the runmanager.SubProtocols = append(manager.SubProtocols, p2p.Protocol{Name:    ProtocolName,Version: version,Length:  ProtocolLengths[i],Run: func(p *p2p.Peer, rw p2p.MsgReadWriter) error {peer := manager.newPeer(int(version), p, rw)select {case manager.newPeerCh <- peer:manager.wg.Add(1)defer manager.wg.Done()//此处如果顺利会进入for循环 如果失败返回错误我会销毁掉这个peerreturn manager.handle(peer)  case <-manager.quitSync:return p2p.DiscQuitting}},NodeInfo: func() interface{} {return manager.NodeInfo()},PeerInfo: func(id discover.NodeID) interface{} {if p := manager.peers.Peer(fmt.Sprintf("%x", id[:8])); p != nil {return p.Info()}return nil},})}

参考

源码：https://github.com/ethereum/go-ethereum/tree/master/p2p

Kademlia算法：https://en.wikipedia.org/wiki/Kademlia

转自：（魂祭心） https://my.oschina.net/hunjixin/blog/1803029

如果你希望高效的学习以太坊DApp开发，可以访问汇智网提供的最热门在线互动教程：

1.适合区块链新手的以太坊DApp实战入门教程
2.区块链+IPFS+Node.js+MongoDB+Express去中心化以太坊电商应用开发实战

其他更多内容也可以访问这个以太坊博客