aboutsummaryrefslogtreecommitdiffstats
path: root/p2p/simulations/network.go
blob: a0e621b888994eede80af213d8f4c7023b5c4908 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
// Copyright 2017 The go-ethereum Authors
// This file is part of the go-ethereum library.
//
// The go-ethereum library is free software: you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// The go-ethereum library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.

package simulations

import (
    "bytes"
    "context"
    "encoding/json"
    "errors"
    "fmt"
    "math/rand"
    "sync"
    "time"

    "github.com/ethereum/go-ethereum/event"
    "github.com/ethereum/go-ethereum/log"
    "github.com/ethereum/go-ethereum/p2p"
    "github.com/ethereum/go-ethereum/p2p/enode"
    "github.com/ethereum/go-ethereum/p2p/simulations/adapters"
)

var DialBanTimeout = 200 * time.Millisecond

// NetworkConfig defines configuration options for starting a Network
type NetworkConfig struct {
    ID             string `json:"id"`
    DefaultService string `json:"default_service,omitempty"`
}

// Network models a p2p simulation network which consists of a collection of
// simulated nodes and the connections which exist between them.
//
// The Network has a single NodeAdapter which is responsible for actually
// starting nodes and connecting them together.
//
// The Network emits events when nodes are started and stopped, when they are
// connected and disconnected, and also when messages are sent between nodes.
type Network struct {
    NetworkConfig

    Nodes   []*Node `json:"nodes"`
    nodeMap map[enode.ID]int

    Conns   []*Conn `json:"conns"`
    connMap map[string]int

    nodeAdapter adapters.NodeAdapter
    events      event.Feed
    lock        sync.RWMutex
    quitc       chan struct{}
}

// NewNetwork returns a Network which uses the given NodeAdapter and NetworkConfig
func NewNetwork(nodeAdapter adapters.NodeAdapter, conf *NetworkConfig) *Network {
    return &Network{
        NetworkConfig: *conf,
        nodeAdapter:   nodeAdapter,
        nodeMap:       make(map[enode.ID]int),
        connMap:       make(map[string]int),
        quitc:         make(chan struct{}),
    }
}

// Events returns the output event feed of the Network.
func (net *Network) Events() *event.Feed {
    return &net.events
}

// NewNodeWithConfig adds a new node to the network with the given config,
// returning an error if a node with the same ID or name already exists
func (net *Network) NewNodeWithConfig(conf *adapters.NodeConfig) (*Node, error) {
    net.lock.Lock()
    defer net.lock.Unlock()

    if conf.Reachable == nil {
        conf.Reachable = func(otherID enode.ID) bool {
            _, err := net.InitConn(conf.ID, otherID)
            if err != nil && bytes.Compare(conf.ID.Bytes(), otherID.Bytes()) < 0 {
                return false
            }
            return true
        }
    }

    // check the node doesn't already exist
    if node := net.getNode(conf.ID); node != nil {
        return nil, fmt.Errorf("node with ID %q already exists", conf.ID)
    }
    if node := net.getNodeByName(conf.Name); node != nil {
        return nil, fmt.Errorf("node with name %q already exists", conf.Name)
    }

    // if no services are configured, use the default service
    if len(conf.Services) == 0 {
        conf.Services = []string{net.DefaultService}
    }

    // use the NodeAdapter to create the node
    adapterNode, err := net.nodeAdapter.NewNode(conf)
    if err != nil {
        return nil, err
    }
    node := &Node{
        Node:   adapterNode,
        Config: conf,
    }
    log.Trace("Node created", "id", conf.ID)
    net.nodeMap[conf.ID] = len(net.Nodes)
    net.Nodes = append(net.Nodes, node)

    // emit a "control" event
    net.events.Send(ControlEvent(node))

    return node, nil
}

// Config returns the network configuration
func (net *Network) Config() *NetworkConfig {
    return &net.NetworkConfig
}

// StartAll starts all nodes in the network
func (net *Network) StartAll() error {
    for _, node := range net.Nodes {
        if node.Up {
            continue
        }
        if err := net.Start(node.ID()); err != nil {
            return err
        }
    }
    return nil
}

// StopAll stops all nodes in the network
func (net *Network) StopAll() error {
    for _, node := range net.Nodes {
        if !node.Up {
            continue
        }
        if err := net.Stop(node.ID()); err != nil {
            return err
        }
    }
    return nil
}

// Start starts the node with the given ID
func (net *Network) Start(id enode.ID) error {
    return net.startWithSnapshots(id, nil)
}

// startWithSnapshots starts the node with the given ID using the give
// snapshots
func (net *Network) startWithSnapshots(id enode.ID, snapshots map[string][]byte) error {
    net.lock.Lock()

    node := net.getNode(id)
    if node == nil {
        net.lock.Unlock()
        return fmt.Errorf("node %v does not exist", id)
    }
    if node.Up {
        net.lock.Unlock()
        return fmt.Errorf("node %v already up", id)
    }
    log.Trace("Starting node", "id", id, "adapter", net.nodeAdapter.Name())
    if err := node.Start(snapshots); err != nil {
        net.lock.Unlock()
        log.Warn("Node startup failed", "id", id, "err", err)
        return err
    }
    node.Up = true
    log.Info("Started node", "id", id)
    ev := NewEvent(node)
    net.lock.Unlock()

    net.events.Send(ev)

    // subscribe to peer events
    client, err := node.Client()
    if err != nil {
        return fmt.Errorf("error getting rpc client  for node %v: %s", id, err)
    }
    events := make(chan *p2p.PeerEvent)
    sub, err := client.Subscribe(context.Background(), "admin", events, "peerEvents")
    if err != nil {
        return fmt.Errorf("error getting peer events for node %v: %s", id, err)
    }
    go net.watchPeerEvents(id, events, sub)
    return nil
}

// watchPeerEvents reads peer events from the given channel and emits
// corresponding network events
func (net *Network) watchPeerEvents(id enode.ID, events chan *p2p.PeerEvent, sub event.Subscription) {
    defer func() {
        sub.Unsubscribe()

        // assume the node is now down
        net.lock.Lock()
        defer net.lock.Unlock()

        node := net.getNode(id)
        if node == nil {
            return
        }
        node.Up = false
        ev := NewEvent(node)
        net.events.Send(ev)
    }()
    for {
        select {
        case event, ok := <-events:
            if !ok {
                return
            }
            peer := event.Peer
            switch event.Type {

            case p2p.PeerEventTypeAdd:
                net.DidConnect(id, peer)

            case p2p.PeerEventTypeDrop:
                net.DidDisconnect(id, peer)

            case p2p.PeerEventTypeMsgSend:
                net.DidSend(id, peer, event.Protocol, *event.MsgCode)

            case p2p.PeerEventTypeMsgRecv:
                net.DidReceive(peer, id, event.Protocol, *event.MsgCode)

            }

        case err := <-sub.Err():
            if err != nil {
                log.Error("Error in peer event subscription", "id", id, "err", err)
            }
            return
        }
    }
}

// Stop stops the node with the given ID
func (net *Network) Stop(id enode.ID) error {
    net.lock.Lock()
    node := net.getNode(id)
    if node == nil {
        net.lock.Unlock()
        return fmt.Errorf("node %v does not exist", id)
    }
    if !node.Up {
        net.lock.Unlock()
        return fmt.Errorf("node %v already down", id)
    }
    node.Up = false
    net.lock.Unlock()

    err := node.Stop()
    if err != nil {
        net.lock.Lock()
        node.Up = true
        net.lock.Unlock()
        return err
    }
    log.Info("Stopped node", "id", id, "err", err)
    net.lock.Lock()
    ev := ControlEvent(node)
    net.lock.Unlock()
    net.events.Send(ev)
    return nil
}

// Connect connects two nodes together by calling the "admin_addPeer" RPC
// method on the "one" node so that it connects to the "other" node
func (net *Network) Connect(oneID, otherID enode.ID) error {
    log.Debug("Connecting nodes with addPeer", "id", oneID, "other", otherID)
    conn, err := net.InitConn(oneID, otherID)
    if err != nil {
        return err
    }
    client, err := conn.one.Client()
    if err != nil {
        return err
    }
    net.events.Send(ControlEvent(conn))
    return client.Call(nil, "admin_addPeer", string(conn.other.Addr()))
}

// Disconnect disconnects two nodes by calling the "admin_removePeer" RPC
// method on the "one" node so that it disconnects from the "other" node
func (net *Network) Disconnect(oneID, otherID enode.ID) error {
    conn := net.GetConn(oneID, otherID)
    if conn == nil {
        return fmt.Errorf("connection between %v and %v does not exist", oneID, otherID)
    }
    if !conn.Up {
        return fmt.Errorf("%v and %v already disconnected", oneID, otherID)
    }
    client, err := conn.one.Client()
    if err != nil {
        return err
    }
    net.events.Send(ControlEvent(conn))
    return client.Call(nil, "admin_removePeer", string(conn.other.Addr()))
}

// DidConnect tracks the fact that the "one" node connected to the "other" node
func (net *Network) DidConnect(one, other enode.ID) error {
    net.lock.Lock()
    defer net.lock.Unlock()
    conn, err := net.getOrCreateConn(one, other)
    if err != nil {
        return fmt.Errorf("connection between %v and %v does not exist", one, other)
    }
    if conn.Up {
        return fmt.Errorf("%v and %v already connected", one, other)
    }
    conn.Up = true
    net.events.Send(NewEvent(conn))
    return nil
}

// DidDisconnect tracks the fact that the "one" node disconnected from the
// "other" node
func (net *Network) DidDisconnect(one, other enode.ID) error {
    net.lock.Lock()
    defer net.lock.Unlock()
    conn := net.getConn(one, other)
    if conn == nil {
        return fmt.Errorf("connection between %v and %v does not exist", one, other)
    }
    if !conn.Up {
        return fmt.Errorf("%v and %v already disconnected", one, other)
    }
    conn.Up = false
    conn.initiated = time.Now().Add(-DialBanTimeout)
    net.events.Send(NewEvent(conn))
    return nil
}

// DidSend tracks the fact that "sender" sent a message to "receiver"
func (net *Network) DidSend(sender, receiver enode.ID, proto string, code uint64) error {
    msg := &Msg{
        One:      sender,
        Other:    receiver,
        Protocol: proto,
        Code:     code,
        Received: false,
    }
    net.events.Send(NewEvent(msg))
    return nil
}

// DidReceive tracks the fact that "receiver" received a message from "sender"
func (net *Network) DidReceive(sender, receiver enode.ID, proto string, code uint64) error {
    msg := &Msg{
        One:      sender,
        Other:    receiver,
        Protocol: proto,
        Code:     code,
        Received: true,
    }
    net.events.Send(NewEvent(msg))
    return nil
}

// GetNode gets the node with the given ID, returning nil if the node does not
// exist
func (net *Network) GetNode(id enode.ID) *Node {
    net.lock.RLock()
    defer net.lock.RUnlock()
    return net.getNode(id)
}

// GetNode gets the node with the given name, returning nil if the node does
// not exist
func (net *Network) GetNodeByName(name string) *Node {
    net.lock.RLock()
    defer net.lock.RUnlock()
    return net.getNodeByName(name)
}

func (net *Network) getNodeByName(name string) *Node {
    for _, node := range net.Nodes {
        if node.Config.Name == name {
            return node
        }
    }
    return nil
}

// GetNodes returns the existing nodes
func (net *Network) GetNodes() (nodes []*Node) {
    net.lock.RLock()
    defer net.lock.RUnlock()

    nodes = append(nodes, net.Nodes...)
    return nodes
}

func (net *Network) getNode(id enode.ID) *Node {
    i, found := net.nodeMap[id]
    if !found {
        return nil
    }
    return net.Nodes[i]
}

// GetRandomUpNode returns a random node on the network, which is running.
func (net *Network) GetRandomUpNode(excludeIDs ...enode.ID) *Node {
    net.lock.RLock()
    defer net.lock.RUnlock()
    return net.getRandomNode(net.getUpNodeIDs(), excludeIDs)
}

func (net *Network) getUpNodeIDs() (ids []enode.ID) {
    for _, node := range net.Nodes {
        if node.Up {
            ids = append(ids, node.ID())
        }
    }
    return ids
}

// GetRandomDownNode returns a random node on the network, which is stopped.
func (net *Network) GetRandomDownNode(excludeIDs ...enode.ID) *Node {
    net.lock.RLock()
    defer net.lock.RUnlock()
    return net.getRandomNode(net.getDownNodeIDs(), excludeIDs)
}

func (net *Network) getDownNodeIDs() (ids []enode.ID) {
    for _, node := range net.GetNodes() {
        if !node.Up {
            ids = append(ids, node.ID())
        }
    }
    return ids
}

func (net *Network) getRandomNode(ids []enode.ID, excludeIDs []enode.ID) *Node {
    filtered := filterIDs(ids, excludeIDs)

    l := len(filtered)
    if l == 0 {
        return nil
    }
    return net.getNode(filtered[rand.Intn(l)])
}

func filterIDs(ids []enode.ID, excludeIDs []enode.ID) []enode.ID {
    exclude := make(map[enode.ID]bool)
    for _, id := range excludeIDs {
        exclude[id] = true
    }
    var filtered []enode.ID
    for _, id := range ids {
        if _, found := exclude[id]; !found {
            filtered = append(filtered, id)
        }
    }
    return filtered
}

// GetConn returns the connection which exists between "one" and "other"
// regardless of which node initiated the connection
func (net *Network) GetConn(oneID, otherID enode.ID) *Conn {
    net.lock.RLock()
    defer net.lock.RUnlock()
    return net.getConn(oneID, otherID)
}

// GetOrCreateConn is like GetConn but creates the connection if it doesn't
// already exist
func (net *Network) GetOrCreateConn(oneID, otherID enode.ID) (*Conn, error) {
    net.lock.Lock()
    defer net.lock.Unlock()
    return net.getOrCreateConn(oneID, otherID)
}

func (net *Network) getOrCreateConn(oneID, otherID enode.ID) (*Conn, error) {
    if conn := net.getConn(oneID, otherID); conn != nil {
        return conn, nil
    }

    one := net.getNode(oneID)
    if one == nil {
        return nil, fmt.Errorf("node %v does not exist", oneID)
    }
    other := net.getNode(otherID)
    if other == nil {
        return nil, fmt.Errorf("node %v does not exist", otherID)
    }
    conn := &Conn{
        One:   oneID,
        Other: otherID,
        one:   one,
        other: other,
    }
    label := ConnLabel(oneID, otherID)
    net.connMap[label] = len(net.Conns)
    net.Conns = append(net.Conns, conn)
    return conn, nil
}

func (net *Network) getConn(oneID, otherID enode.ID) *Conn {
    label := ConnLabel(oneID, otherID)
    i, found := net.connMap[label]
    if !found {
        return nil
    }
    return net.Conns[i]
}

// InitConn(one, other) retrieves the connection model for the connection between
// peers one and other, or creates a new one if it does not exist
// the order of nodes does not matter, i.e., Conn(i,j) == Conn(j, i)
// it checks if the connection is already up, and if the nodes are running
// NOTE:
// it also checks whether there has been recent attempt to connect the peers
// this is cheating as the simulation is used as an oracle and know about
// remote peers attempt to connect to a node which will then not initiate the connection
func (net *Network) InitConn(oneID, otherID enode.ID) (*Conn, error) {
    net.lock.Lock()
    defer net.lock.Unlock()
    if oneID == otherID {
        return nil, fmt.Errorf("refusing to connect to self %v", oneID)
    }
    conn, err := net.getOrCreateConn(oneID, otherID)
    if err != nil {
        return nil, err
    }
    if conn.Up {
        return nil, fmt.Errorf("%v and %v already connected", oneID, otherID)
    }
    if time.Since(conn.initiated) < DialBanTimeout {
        return nil, fmt.Errorf("connection between %v and %v recently attempted", oneID, otherID)
    }

    err = conn.nodesUp()
    if err != nil {
        log.Trace("Nodes not up", "err", err)
        return nil, fmt.Errorf("nodes not up: %v", err)
    }
    log.Debug("Connection initiated", "id", oneID, "other", otherID)
    conn.initiated = time.Now()
    return conn, nil
}

// Shutdown stops all nodes in the network and closes the quit channel
func (net *Network) Shutdown() {
    for _, node := range net.Nodes {
        log.Debug("Stopping node", "id", node.ID())
        if err := node.Stop(); err != nil {
            log.Warn("Can't stop node", "id", node.ID(), "err", err)
        }
    }
    close(net.quitc)
}

// Reset resets all network properties:
// empties the nodes and the connection list
func (net *Network) Reset() {
    net.lock.Lock()
    defer net.lock.Unlock()

    //re-initialize the maps
    net.connMap = make(map[string]int)
    net.nodeMap = make(map[enode.ID]int)

    net.Nodes = nil
    net.Conns = nil
}

// Node is a wrapper around adapters.Node which is used to track the status
// of a node in the network
type Node struct {
    adapters.Node `json:"-"`

    // Config if the config used to created the node
    Config *adapters.NodeConfig `json:"config"`

    // Up tracks whether or not the node is running
    Up bool `json:"up"`
}

// ID returns the ID of the node
func (n *Node) ID() enode.ID {
    return n.Config.ID
}

// String returns a log-friendly string
func (n *Node) String() string {
    return fmt.Sprintf("Node %v", n.ID().TerminalString())
}

// NodeInfo returns information about the node
func (n *Node) NodeInfo() *p2p.NodeInfo {
    // avoid a panic if the node is not started yet
    if n.Node == nil {
        return nil
    }
    info := n.Node.NodeInfo()
    info.Name = n.Config.Name
    return info
}

// MarshalJSON implements the json.Marshaler interface so that the encoded
// JSON includes the NodeInfo
func (n *Node) MarshalJSON() ([]byte, error) {
    return json.Marshal(struct {
        Info   *p2p.NodeInfo        `json:"info,omitempty"`
        Config *adapters.NodeConfig `json:"config,omitempty"`
        Up     bool                 `json:"up"`
    }{
        Info:   n.NodeInfo(),
        Config: n.Config,
        Up:     n.Up,
    })
}

// Conn represents a connection between two nodes in the network
type Conn struct {
    // One is the node which initiated the connection
    One enode.ID `json:"one"`

    // Other is the node which the connection was made to
    Other enode.ID `json:"other"`

    // Up tracks whether or not the connection is active
    Up bool `json:"up"`
    // Registers when the connection was grabbed to dial
    initiated time.Time

    one   *Node
    other *Node
}

// nodesUp returns whether both nodes are currently up
func (c *Conn) nodesUp() error {
    if !c.one.Up {
        return fmt.Errorf("one %v is not up", c.One)
    }
    if !c.other.Up {
        return fmt.Errorf("other %v is not up", c.Other)
    }
    return nil
}

// String returns a log-friendly string
func (c *Conn) String() string {
    return fmt.Sprintf("Conn %v->%v", c.One.TerminalString(), c.Other.TerminalString())
}

// Msg represents a p2p message sent between two nodes in the network
type Msg struct {
    One      enode.ID `json:"one"`
    Other    enode.ID `json:"other"`
    Protocol string   `json:"protocol"`
    Code     uint64   `json:"code"`
    Received bool     `json:"received"`
}

// String returns a log-friendly string
func (m *Msg) String() string {
    return fmt.Sprintf("Msg(%d) %v->%v", m.Code, m.One.TerminalString(), m.Other.TerminalString())
}

// ConnLabel generates a deterministic string which represents a connection
// between two nodes, used to compare if two connections are between the same
// nodes
func ConnLabel(source, target enode.ID) string {
    var first, second enode.ID
    if bytes.Compare(source.Bytes(), target.Bytes()) > 0 {
        first = target
        second = source
    } else {
        first = source
        second = target
    }
    return fmt.Sprintf("%v-%v", first, second)
}

// Snapshot represents the state of a network at a single point in time and can
// be used to restore the state of a network
type Snapshot struct {
    Nodes []NodeSnapshot `json:"nodes,omitempty"`
    Conns []Conn         `json:"conns,omitempty"`
}

// NodeSnapshot represents the state of a node in the network
type NodeSnapshot struct {
    Node Node `json:"node,omitempty"`

    // Snapshots is arbitrary data gathered from calling node.Snapshots()
    Snapshots map[string][]byte `json:"snapshots,omitempty"`
}

// Snapshot creates a network snapshot
func (net *Network) Snapshot() (*Snapshot, error) {
    return net.snapshot(nil, nil)
}

func (net *Network) SnapshotWithServices(addServices []string, removeServices []string) (*Snapshot, error) {
    return net.snapshot(addServices, removeServices)
}

func (net *Network) snapshot(addServices []string, removeServices []string) (*Snapshot, error) {
    net.lock.Lock()
    defer net.lock.Unlock()
    snap := &Snapshot{
        Nodes: make([]NodeSnapshot, len(net.Nodes)),
    }
    for i, node := range net.Nodes {
        snap.Nodes[i] = NodeSnapshot{Node: *node}
        if !node.Up {
            continue
        }
        snapshots, err := node.Snapshots()
        if err != nil {
            return nil, err
        }
        snap.Nodes[i].Snapshots = snapshots
        for _, addSvc := range addServices {
            haveSvc := false
            for _, svc := range snap.Nodes[i].Node.Config.Services {
                if svc == addSvc {
                    haveSvc = true
                    break
                }
            }
            if !haveSvc {
                snap.Nodes[i].Node.Config.Services = append(snap.Nodes[i].Node.Config.Services, addSvc)
            }
        }
        if len(removeServices) > 0 {
            var cleanedServices []string
            for _, svc := range snap.Nodes[i].Node.Config.Services {
                haveSvc := false
                for _, rmSvc := range removeServices {
                    if rmSvc == svc {
                        haveSvc = true
                        break
                    }
                }
                if !haveSvc {
                    cleanedServices = append(cleanedServices, svc)
                }

            }
            snap.Nodes[i].Node.Config.Services = cleanedServices
        }
    }
    for _, conn := range net.Conns {
        if conn.Up {
            snap.Conns = append(snap.Conns, *conn)
        }
    }
    return snap, nil
}

var snapshotLoadTimeout = 120 * time.Second

// Load loads a network snapshot
func (net *Network) Load(snap *Snapshot) error {
    // Start nodes.
    for _, n := range snap.Nodes {
        if _, err := net.NewNodeWithConfig(n.Node.Config); err != nil {
            return err
        }
        if !n.Node.Up {
            continue
        }
        if err := net.startWithSnapshots(n.Node.Config.ID, n.Snapshots); err != nil {
            return err
        }
    }

    // Prepare connection events counter.
    allConnected := make(chan struct{}) // closed when all connections are established
    done := make(chan struct{})         // ensures that the event loop goroutine is terminated
    defer close(done)

    // Subscribe to event channel.
    // It needs to be done outside of the event loop goroutine (created below)
    // to ensure that the event channel is blocking before connect calls are made.
    events := make(chan *Event)
    sub := net.Events().Subscribe(events)
    defer sub.Unsubscribe()

    go func() {
        // Expected number of connections.
        total := len(snap.Conns)
        // Set of all established connections from the snapshot, not other connections.
        // Key array element 0 is the connection One field value, and element 1 connection Other field.
        connections := make(map[[2]enode.ID]struct{}, total)

        for {
            select {
            case e := <-events:
                // Ignore control events as they do not represent
                // connect or disconnect (Up) state change.
                if e.Control {
                    continue
                }
                // Detect only connection events.
                if e.Type != EventTypeConn {
                    continue
                }
                connection := [2]enode.ID{e.Conn.One, e.Conn.Other}
                // Nodes are still not connected or have been disconnected.
                if !e.Conn.Up {
                    // Delete the connection from the set of established connections.
                    // This will prevent false positive in case disconnections happen.
                    delete(connections, connection)
                    log.Warn("load snapshot: unexpected disconnection", "one", e.Conn.One, "other", e.Conn.Other)
                    continue
                }
                // Check that the connection is from the snapshot.
                for _, conn := range snap.Conns {
                    if conn.One == e.Conn.One && conn.Other == e.Conn.Other {
                        // Add the connection to the set of established connections.
                        connections[connection] = struct{}{}
                        if len(connections) == total {
                            // Signal that all nodes are connected.
                            close(allConnected)
                            return
                        }

                        break
                    }
                }
            case <-done:
                // Load function returned, terminate this goroutine.
                return
            }
        }
    }()

    // Start connecting.
    for _, conn := range snap.Conns {

        if !net.GetNode(conn.One).Up || !net.GetNode(conn.Other).Up {
            //in this case, at least one of the nodes of a connection is not up,
            //so it would result in the snapshot `Load` to fail
            continue
        }
        if err := net.Connect(conn.One, conn.Other); err != nil {
            return err
        }
    }

    select {
    // Wait until all connections from the snapshot are established.
    case <-allConnected:
    // Make sure that we do not wait forever.
    case <-time.After(snapshotLoadTimeout):
        return errors.New("snapshot connections not established")
    }
    return nil
}

// Subscribe reads control events from a channel and executes them
func (net *Network) Subscribe(events chan *Event) {
    for {
        select {
        case event, ok := <-events:
            if !ok {
                return
            }
            if event.Control {
                net.executeControlEvent(event)
            }
        case <-net.quitc:
            return
        }
    }
}

func (net *Network) executeControlEvent(event *Event) {
    log.Trace("Executing control event", "type", event.Type, "event", event)
    switch event.Type {
    case EventTypeNode:
        if err := net.executeNodeEvent(event); err != nil {
            log.Error("Error executing node event", "event", event, "err", err)
        }
    case EventTypeConn:
        if err := net.executeConnEvent(event); err != nil {
            log.Error("Error executing conn event", "event", event, "err", err)
        }
    case EventTypeMsg:
        log.Warn("Ignoring control msg event")
    }
}

func (net *Network) executeNodeEvent(e *Event) error {
    if !e.Node.Up {
        return net.Stop(e.Node.ID())
    }

    if _, err := net.NewNodeWithConfig(e.Node.Config); err != nil {
        return err
    }
    return net.Start(e.Node.ID())
}

func (net *Network) executeConnEvent(e *Event) error {
    if e.Conn.Up {
        return net.Connect(e.Conn.One, e.Conn.Other)
    } else {
        return net.Disconnect(e.Conn.One, e.Conn.Other)
    }
}