IBC-TLA+

We formalize main components of the IBC protocol with TLA+, a temporal logic specification language, and verify important requirements with model checking tool TLC. This document introduces the overall framework of the public code of our research, and explains the reports and feedbacks of the identified problems through our analysis.

Our research includes the connection, channel, and packet transmission of the core TAO layer of the IBC protocol, as well as fungible and non-fungible token transfer of the application layer. The modeling of each part consists of two modules: chain and environment. And we use the vscode extension of TLA+ (ref).

At the beginning of our research, the latest commitment of the IBC protocol was this, and subsequent commitments were not considered.

ICS-03 Connection

Chain

Description

This module represents the block chain, including the data structures and functions about connection handshake. See ICS03/Chain.tla.

Data Structure

  "ChainStore":{
    "client":{
      "clientID":"string"
      },
    "connectionEnds":[{
      "state":"string",
      "connectionID":"int",
      "clientID":"string",
      "counterpartyConnectionID":"int",
      "counterpartyClientID":"string",
      "versions":"Set(int)"
    }],
    "nextConnectionID":"int",
    "compatibleVersions":"Set(int)"
  }

Handle Functions

• HandleConnOpenInit
• HandleConnOpenTry
• HandleConnOpenAck
• HandleConnOpenConfirm

Auxiliary Functions

• Query Functions
  • getChainStore
  • queryClient
  • queryConnection
  • getCompatibleVersions
• Others
  • generateID
  • pickVersion
  • verifyConnectionState

Environment

Description

This module simulates the possible behavior of on-chian users and off-chain relayers by arbitrarily calling the Chain module. See ICS03/Environment.tla.

Key Function

HandleConnection(chainID) ==
    \E clientID, counterpartyClientID \in ClientIDs, connectionID, counterpartyConnectionID \in ConnectionIDs, version \in (SUBSET Versions \ {{}}): 
        LET 
            counterpartyChainID == getCounterpartyChainID(chainID)
            proof == proofConnectionState(counterpartyChainID, counterpartyConnectionID) 
        IN
        IF chainID = "chainA"
        THEN  
            /\ ChainA!HandleChannel(clientID, counterpartyClientID, connectionID, counterpartyConnectionID, version, proof)
            /\ UNCHANGED chainBvars
        ELSE  
            /\ ChainB!HandleChannel(clientID, counterpartyClientID, connectionID, counterpartyConnectionID, version, proof)
            /\ UNCHANGED chainAvars

ICS-04 Channel

Chain

Description

This module represents the block chain, including the data structures and functions about channel handshake. See ICS04/channel/Chain.tla.

Data Structure

  "ChainStore":{
    "client":{
      "clientID":"int"
      },
    "connectionEnds":[{
      "state":"string",
      "connectionID":"int",
      "clientID":"int",
      "counterpartyConnectionID":"int",
      "counterpartyClientID":"int",
      "versions":"Set(int)"
    }],
    "channelEnds":[{
        "portID":"int",
        "channelID":"int",
        "state":"string",
        "order":"string",
        "counterpartyPortID":"int",
        "counterpartyChannelID":"int",
        "connectionID":"int",
        "version":"int"
    }],
    "nextChannelID":"int"
  }

Handle Functions

• HandlechanOpenInit
• HandlechanOpenTry
• HandlechanOpenAck
• HandlechanOpenConfirm
• HandlechanCloseInit
• HandlechanCloseConfirm

Auxiliary Functions

• Query Functions
  • getChainStore
  • getChannelEnd
  • getConnection
• Others
  • generateID
  • verifyChannelState

Environment

Description

This module simulates the possible behavior of on-chian users and off-chain relayers by arbitrarily calling the Chain module. See ICS04/channel/Environment.tla.

Key Function

HandleChannel(chainID) ==
    \E order \in Orders, connectionID \in ConnectionIDs, 
        portID, counterpartyPortID \in PortIDs, channelID, counterpartyChannelID \in ChannelIDs, version, counterpartyVersion \in Versions: 
        LET 
            counterpartyChainID == getCounterpartyChainID(chainID)
            proof == proofChannelState(counterpartyChainID, counterpartyPortID, counterpartyChannelID) 
        IN
        IF chainID = "chainA"
        THEN  
            /\ ChainA!HandleChannel(order, connectionID, portID, channelID, counterpartyPortID, counterpartyChannelID, version, counterpartyVersion, proof)
            /\ UNCHANGED chainBvars
        ELSE  
            /\ ChainB!HandleChannel(order, connectionID, portID, channelID, counterpartyPortID, counterpartyChannelID, version, counterpartyVersion, proof)
            /\ UNCHANGED chainAvars

ICS-04 Packet

Chain

Description

This module represents the block chain, including the data structures and functions about channel handshake and packet delivery. See ICS04/packet/Chain.tla.

Data Structure

  "ChainStore":{
    "client":{
      "clientID":"int"
      },
    "connectionEnds":[{
      "state":"string",
      "connectionID":"int",
      "clientID":"int",
      "counterpartyConnectionID":"int",
      "counterpartyClientID":"int",
      "versions":"Set(int)"
    }],
    "channelEnds":[{
        "portID":"int",
        "channelID":"int",
        "state":"string",
        "order":"string",
        "counterpartyPortID":"int",
        "counterpartyChannelID":"int",
        "connectionID":"int",
        "version":"int"
    }],
    "nextSequenceSend":["int"],
    "nextSequenceRecv":["int"],
    "nextSequenceAck":["int"],
    "commitments":["Set({'portID':'int','channelID':'int','sequence':'int'})"],
    "receipts":["Set({'portID':'int','channelID':'int','sequence':'int','state':'string'})"],
    "acks":["Set({'portID':'int','channelID':'int','sequence':'int'})"],
    "nextChannelID":"int"
  }

Handle Functions

• HandlechanOpenInit
• HandlechanOpenTry
• HandlechanOpenAck
• HandlechanOpenConfirm
• HandlechanCloseInit
• HandlechanCloseConfirm
• HandleSendPacket
• HandleRecvPacketandWriteAcknowledgement
• HandleAcknowledgePacket
• HandleTimeoutPacket
• HandleTimeoutOnClose

Auxiliary Functions

• Query Functions
  • getChainStore
  • getChannelEnd
  • getConnection
  • getNextSeqSend
  • getNextSeqRecv
  • getNextSeqAck
  • getPacketReceipt
  • getPacketAck
  • getPacketCommit
• Others
  • generateID
  • verifyChannelState
  • verifyPacketData
  • verifyPacketAck
  • verifyNextSeqRecv
  • verifyPacketReceiptAbsence
  • verifyPacketReceipt
  • isSendTimeout
  • isRecvTimeout

Environment

Description

This module simulates the possible behavior of on-chian users and off-chain relayers by arbitrarily calling the Chain module. See ICS04/packet/Environment.tla.

Key Function

HandleChannel(chainID) ==
    \E order \in Orders, connectionID \in ConnectionIDs, 
        portID, counterpartyPortID \in PortIDs, channelID, counterpartyChannelID \in ChannelIDs, version, counterpartyVersion \in Versions: 
        LET 
            counterpartyChainID == getCounterpartyChainID(chainID)
            proof == proofChannelState(counterpartyChainID, counterpartyPortID, counterpartyChannelID) 
        IN
        IF chainID = "chainA"
        THEN  
            /\ ChainA!HandleChannel(order, connectionID, portID, channelID, counterpartyPortID, counterpartyChannelID, version, counterpartyVersion, proof)
            /\ UNCHANGED chainBvars
        ELSE  
            /\ ChainB!HandleChannel(order, connectionID, portID, channelID, counterpartyPortID, counterpartyChannelID, version, counterpartyVersion, proof)
            /\ UNCHANGED chainAvars

HandlePacket(chainID) == 
    \E sourcePort, destPort \in PortIDs, sourceChannel, destChannel \in ChannelIDs, sequence, nextSeqRecv \in Seqs:
    LET 
        counterpartyChainID == getCounterpartyChainID(chainID)
        proofSource == proofPacketState(counterpartyChainID, sourcePort, sourceChannel, sequence)
        proofDest == proofPacketState(counterpartyChainID, destPort, destChannel, sequence)
        proofChannel == proofChannelState(counterpartyChainID, destPort, destChannel)
    IN 
    IF chainID = "chainA"
    THEN  
        /\ ChainA!HandlePacket(sourcePort, sourceChannel, destPort, destChannel, sequence, nextSeqRecv, proofChannel, proofSource, proofDest)
        /\ UNCHANGED chainBvars
    ELSE  
        /\ ChainB!HandlePacket(sourcePort, sourceChannel, destPort, destChannel, sequence, nextSeqRecv, proofChannel, proofSource, proofDest)
        /\ UNCHANGED chainAvars

ICS-20 Fungible Token Transfer

Chain

Description

See ICS20/Chain.tla.

Data Structure

"ChainStore":{
    "connectionEnds":[{
      "state":string,
      "connectionID":int,
      "clientID":int,
      "counterpartyConnectionID":int,
      "counterpartyClientID":int,
      "versions":Set(int)
    }],
    "channelEnds":[{
        "portID":int,
        "channelID":int,
        "state":string,
        "order":string,
        "counterpartyPortID":int,
        "counterpartyChannelID":int,
        "connectionID":int,
        "version":int
    }],
    "nextSequenceSend":[int],
    "nextSequenceRecv":[int],
    "nextSequenceAck":[int],
    "commitments":[Set({'portID':int,'channelID':int,'sequence':int,'data':struct})],
    "receipts":[Set({'portID':int,'channelID':int,'sequence':int,'state':string})],
    "acks":[Set({'portID':int,'channelID':int,'sequence':int,'state':string})]
  }
"FungibleTokenAccounts":[(string, Seq(Set(string))) -> int]

Environment

Description

See ICS20/Enviroment

Key Function

Next == 
    \/ \E chainID \in ChainIDs, 
            sequence, nextSeqRecv \in Seqs,
            channel \in ChannelIDs,
            denomination \in {x[2] : x \in DOMAIN accountsA} \union {x[2]: x \in DOMAIN accountsB}, 
            amount \in 1..MaxAmount:
        LET 
            port == getPortID(channel)
            channelEnd == getChannelEnd(chainID, channel)
            counterpartyChannel == channelEnd.counterpartyChannelID
            counterpartyPort == channelEnd.counterpartyPortID
            counterpartyChainID == getCounterpartyChainID(chainID)
            ack == getPacketAck(counterpartyChainID, counterpartyPort, counterpartyChannel, sequence)
            proof == proofPacketState(counterpartyChainID, counterpartyPort, counterpartyChannel, sequence)
        IN 
        IF chainID = "chainA"
        THEN  
            /\ ChainA!Actions(port, channel, counterpartyPort, counterpartyChannel, sequence, nextSeqRecv, ack, proof, denomination, amount)
            /\ UNCHANGED chainVarsB
        ELSE  
            /\ ChainB!Actions(port, channel, counterpartyPort, counterpartyChannel, sequence, nextSeqRecv, ack, proof, denomination, amount)
            /\ UNCHANGED chainVarsA
    \/ UNCHANGED vars

ICS-721 Non-Fungible Token Transfer

Chain

Description

See ICS721/Chain.tla.

Data Structure

"ChainStore":{
    "connectionEnds":[{
      "state":string,
      "connectionID":int,
      "clientID":int,
      "counterpartyConnectionID":int,
      "counterpartyClientID":int,
      "versions":Set(int)
    }],
    "channelEnds":[{
        "portID":int,
        "channelID":int,
        "state":string,
        "order":string,
        "counterpartyPortID":int,
        "counterpartyChannelID":int,
        "connectionID":int,
        "version":int
    }],
    "nextSequenceSend":[int],
    "nextSequenceRecv":[int],
    "nextSequenceAck":[int],
    "commitments":[Set({'portID':int,'channelID':int,'sequence':int,'data':struct})],
    "receipts":[Set({'portID':int,'channelID':int,'sequence':int,'state':string})],
    "acks":[Set({'portID':int,'channelID':int,'sequence':int,'state':string})]
  }
"NonFungibleTokenAccounts":[(Seq(Set(string)), int) -> string]

Environment

Description

See ICS721/Enviroment

Key Function

Next == 
    \/ \E chainID \in ChainIDs, 
            sequence, nextSeqRecv \in Seqs,
            channel \in ChannelIDs,
            classId \in LimitClassIds,
            tokenIds \in ((SUBSET TokenIds) \ {{}}):
        LET 
            port == getPortID(channel)
            channelEnd == getChannelEnd(chainID, channel)
            counterpartyChannel == channelEnd.counterpartyChannelID
            counterpartyPort == channelEnd.counterpartyPortID
            counterpartyChainID == getCounterpartyChainID(chainID)
            ack == getPacketAck(counterpartyChainID, counterpartyPort, counterpartyChannel, sequence)
            proof == proofPacketState(counterpartyChainID, counterpartyPort, counterpartyChannel, sequence)
        IN 
        IF chainID = "chainA"
        THEN  
            /\ ChainA!Actions(port, channel, counterpartyPort, counterpartyChannel, sequence, nextSeqRecv, ack, proof, classId, tokenIds)
            /\ UNCHANGED chainVarsB
        ELSE  
            /\ ChainB!Actions(port, channel, counterpartyPort, counterpartyChannel, sequence, nextSeqRecv, ack, proof, classId, tokenIds)
            /\ UNCHANGED chainVarsA
    \/ UNCHANGED vars

Reports and Feedbacks

This section mainly discusses the reasons, triggering paths, repair suggestions of identified problems and community feedbacks to our reports.

Uncompleted Handshake

Problem 1

Reasons and Triggering Paths

function connOpenInit(...) {
    ...
    identifier = generateIdentifier()
    ...
    connection = ConnectionEnd{...}
    provableStore.set(connectionPath(identifier), connection)
    ...
}

function connOpenTry(...) {
    ...
    identifier = generateIdentifier()
    ...
    connection = ConnectionEnd{...}
    provableStore.set(connectionPath(identifier), connection)
    ...
}

As above, both connOpenInit and connOpenTry need to create a new connection end with a new identifier and the allocated connection and and identifiers will not be reallocated. It is acceptable that limited identifiers can only be used to create limited connections. But if connOpenInit runs out of identifiers, connOpenTry cannot be executed, and there is no connection that can be established, which is a possible DoS attack.

For example, both Chain A and Chain B can only allocate one identifier. If both Chain A and Chain B want to establish a connection with the other party by executing the connOpenInit, neither party can execute connOpenTry to respond to the connection initiated by the other party, resulting in no connection being established.

Recommendation and Reports

The direct method is to allow the use of allocated connection ends to complete connection handshake. For example, if both Chain A and Chain B want to initiate a connection with each other, they should directly use two INIT state connection ends that have already been created to complete the handshake, instead of creating two more TRYOPEN state connection ends to complete two handshakes. We do not report this problem because similar solutions have been deprecated due to their complex and error prone processing logic, which can refer to this. Considering the high transaction costs and data representation range on the blockchain, identifiers or other resources may be considered sufficient.

Problem 2

Reasons and Triggering Paths

A INIT state connection end of the counterparty chain can be used to create multiple TRYOPEN state connection ends, because connOpenTry only verifies the counterparty connection end's state and does not check if it already has a corresponding connection end. However, only one of these TRYOPEN state connection ends will match the INIT state connection end in the connOpenAck because the ```connOpenAck`` will change connection end's state. The remaining connection ends will freeze in their previous states. These frozen ends will interfere with the normal use of users, such as mistakenly executing transactions through these unopened connections/channels, resulting in transaction errors.

function connOpenAck(...) {
    ...
    abortTransactionUnless((connection.state === INIT && connection.versions.indexOf(version) !== -1))
    ...
    connection.state = OPEN
    ...
    connection.counterpartyConnectionIdentifier = counterpartyIdentifier
    provableStore.set(connectionPath(identifier), connection)
}

Recommendation and Reports

There are two ways to solve this problem. One is to determine whether there is a corresponding connection end when executing the connOpenTry, and the other is to allow the closure of connection ends that cannot complete the handshake. However, the former requires traversing existing connection ends, while the latter requires the introduction of new judgment and permission mechanisms. Similar situations have occurred in the actual use of IBC, and some have proposed mechanisms similar to the latter, but they have not been adopted.

Incorrect design of ordered_allow_timeout channels

Problem 3

Reasons and Triggering Paths

Ordered_allow_timeout channel is a special channel that allows for timeout situations in ordered transmission. However, there are some errors in determining the timeout condition for this channel type.

According to

function recvPacket(...): {
  ...
      switch channel.order {
  ...
        case ORDERED_ALLOW_TIMEOUT:
          if (getConsensusHeight() >= packet.timeoutHeight && packet.timeoutHeight != 0) || (currentTimestamp() >= packet.timeoutTimestamp && packet.timeoutTimestamp != 0) {
            nextSequenceRecv = nextSequenceRecv + 1
            provableStore.set(nextSequenceRecvPath(packet.destPort, packet.destChannel), nextSequenceRecv)
            provableStore.set(
            packetReceiptPath(packet.destPort, packet.destChannel, packet.sequence),
              TIMEOUT_RECEIPT
            )
          }
          return;
  ...
      }
  ...
      switch channel.order {
  ...
        case ORDERED_ALLOW_TIMEOUT:
          nextSequenceRecv = nextSequenceRecv + 1
          provableStore.set(nextSequenceRecvPath(packet.destPort, packet.destChannel), nextSequenceRecv)
          break;
  ...
  }
}

A module can claim a packet is timeout if

nextSequenceRecv  <= packet.sequence

means function recvPacket has not been called. or

(nextSequenceRecv  == packet.sequence + 1) &&  (verifyPacketReceipt(...packet.sequence, TIMEOUT_RECEIPT)) == True)

means function recvPacket has been called and timeout height/timestamp has passed. But in

function timeoutPacket(...): {

...
    switch channel.order {
...
      case ORDERED_ALLOW_TIMEOUT:
        abortTransactionUnless(nextSequenceRecv == packet.sequence)
        abortTransactionUnless(connection.verifyPacketReceipt(
          proofHeight,
          proof,
          packet.destPort,
          packet.destChannel,
          packet.sequence
          TIMEOUT_RECEIPT,
        ))
        break;
...
    } 
...
}

function timeoutOnClose(...): {
  ...

    if channel.order === ORDERED || channel.order == ORDERED_ALLOW_TIMEOUT {
      abortTransactionUnless(connection.verifyNextSequenceRecv(
        proofHeight,
        proof,
        packet.destPort,
        packet.destChannel,
        nextSequenceRecv
      ))
       abortTransactionUnless(nextSequenceRecv <= packet.sequence)
  ...
  ...
  }
}

The right one should be

abortTransactionUnless(nextSequenceRecv  <= packet.sequence) || (verifyPacketReceipt(...packet.sequence, TIMEOUT_RECEIPT) == True)

Recommendation and Reports

We've reported this problem (ref) and it's being fixed (ref).

Problem 4

Reasons and Triggering Paths

For ORDERED_ALLOW_TIMEOUT channel, timeoutPacket will increase nextSequenceAck.

function timeoutPacket(...): {
    ...
    if channel.order === ORDERED_ALLOW_TIMEOUT {
      nextSequenceAck = nextSequenceAck + 1
      provableStore.set(nextSequenceAckPath(packet.srcPort, packet.srcChannel), nextSequenceAck)
    }
    ...
  }

Assume that module A sent two packets (e.g. sequences are 1,2 and the init nextSequenceAck == 1) and module B received No. 1 packet successfully. But No. 2 packet was timeout and module A called timeoutPacket for it, which caused the nextSequenceAck == 2. And now module A can't call acknowledgePacket for No.1 packet, because acknowledgePacket requires packet.sequence === nextSequenceAck.

function acknowledgePacket(...): {
    ...
    if (channel.order === ORDERED || channel.order == ORDERED_ALLOW_TIMEOUT) {
      nextSequenceAck = provableStore.get(nextSequenceAckPath(packet.sourcePort, packet.sourceChannel))
      abortTransactionUnless(packet.sequence === nextSequenceAck)
      ...
    }
    ...
  }

Recommendation and Reports

There are different solutions to solve this problem, but if data packets are required to be received and answered in the order they were sent, conditions should be added to ensure that all datagram operations are strictly executed in order.

function timeoutPacket(...): {
    ...
    if channel.order === ORDERED_ALLOW_TIMEOUT {
      abortTransactionUnless(nextSequenceAck === packet.sequence)
      nextSequenceAck = nextSequenceAck + 1
      provableStore.set(nextSequenceAckPath(packet.srcPort, packet.srcChannel), nextSequenceAck)
    }
    ...
  }

We reported this problem, but the developer think it can be handled by the application (ref). Of course, applications can address this issue in a targeted manner, but application developers may not be aware of the need to specifically address this issue. And this problem is actually due to the protocol not fully implementing its semantics: for ordered channels, packets should be processed strictly in order, and the loose conditions of timeoutPacket break this strict order.

Inappropriate handling of abnormal channel states

Problem 5

Reasons and Triggering Paths

Assume that module a on chain A sends a packet to module b on chain B (by calling sendPacket and assume that the channel is already open in both chains), and module b receives the packet successfully (by calling recvPacket), then module b closes the channel (by calling chanCloseInit). Now module a has two datagrams on the way and two function acknowledgePacket and chanCloseConfirm to call. If chanCloseConfirm is called first due to the concurrency between relayers or incorrect behavior of relay (the relay is not trusted), acknowledgePacket can not be called bacause the channel is closed.

function acknowledgePacket(...): {
    ...
    channel = provableStore.get(channelPath(packet.sourcePort, packet.sourceChannel))
    abortTransactionUnless(channel !== null)
    abortTransactionUnless(channel.state === OPEN)
    ...
  }

And the packet can't be ack or timeout (since it has been received successfully).

Recommendation and Reports

At the protocol level, it is possible to solve this problem by allowing packets to be acknowledged after the channel is closed, but developers believe that such issues should be addressed at the application level (ref).

Problem 6

Reasons and Triggering Paths

As ICS04, A module can call sendPacket before a channel open, which is called optimistic sending.

function sendPacket(...): {
    ...
    abortTransactionUnless(channel !== null)
    abortTransactionUnless(channel.state !== CLOSED)
    ...
  }

Assume the module successively calls chanOpenInit, sendPacket and chanCloseInit, which results that there is a in-flight packet and the closed channel.

The module can't call timeoutOnClose since there is no counterparty channel.

function timeoutOnClose(
  packet: Packet,
  proof: CommitmentProof,
  proofClosed: CommitmentProof,
  proofHeight: Height,
  nextSequenceRecv: Maybe<uint64>,
  relayer: string): Packet {
    ...
    expected = ChannelEnd{CLOSED, channel.order, channel.portIdentifier,
                          channel.channelIdentifier, channel.connectionHops.reverse(), channel.version}
    abortTransactionUnless(connection.verifyChannelState(
      proofHeight,
      proofClosed,
      channel.counterpartyPortIdentifier,
      channel.counterpartyChannelIdentifier,
      expected
    ))
    ...    
  }

It does not match what the timeoutOnClose is designed for and

Any in-flight packets can be timed-out as soon as a channel is closed

Recommendation and Reports

There are different ways to solve this problem. We have reported this issue to the developers, who believe that although there is an issue at the protocol level, optimistic sending has been disabled in practical implementation (ref). In fact, optimistic sending is disabled because similar issues have occurred in practical use (ref). The root cause of these problems is that optimistic sending allows packets to be sent on channels that are not yet open, which may result in the channel being in a state of no counterparty, neither open nor closed, closed, or open. But the two functions timeoutOnClose and timeoutPacket fail to cover all situations.

Abnormal Refund of Transferred Token

See report.

Code Redundancy

Reasons and Triggering Paths

With protocol updates, some code may no longer be in effect. It may be difficult for manual analysis to discover and verify these redundant codes. But through formal analysis, we can prove that these codes are no longer effective.

For example, the condition in pendingDatagrams in ICS18 is localEnd.state === INIT && (remoteEnd === null || remoteEnd.state === INIT). It gives the impression that the remoteEnd corresponding to localEnd is determined, which is misleading. Actually, counterpartyconnectionIdentifier in localEnd is empty, which means remoteEnd must be null.

function pendingDatagrams(...):{
...
  connections = chain.getConnectionsUsingClient(counterparty)
  for (const localEnd of connections) {
    remoteEnd = counterparty.getConnection(localEnd.counterpartyIdentifier)
    if (localEnd.state === INIT && (remoteEnd === null || remoteEnd.state === INIT))
      counterpartyDatagrams.push(ConnOpenTry{...})
  }
...
}

Recommendation and Reports

We have reported this kind of problems and it has been fixed (ref 1, 2).

Inconsistent in Documents

Reasons and Triggering Paths

Due to the continuous maintenance of protocol documents by multiple individuals, it is inevitable that some inconsistencies or errors may occur. Due to the need for careful review of documents during the formal analysis, it is easier to discover these documents.

For exampleConnOpenInit called by handleConnOpenInit in ICS26 should correspond to connOpenInit in ICS03. In ICS03, connOpenInit takes counterpartyPrefix, clientIdentifier, counterpartyClientIdentifier, version as parameters.

function connOpenInit(
  counterpartyPrefix: CommitmentPrefix,
  clientIdentifier: Identifier,
  counterpartyClientIdentifier: Identifier,
  version: string,
  delayPeriodTime: uint64,
  delayPeriodBlocks: uint64) 

But in ICS26, connOpenInit takes identifier, desiredCounterpartyIdentifier, clientIdentifier, counterpartyClientIdentifier, version as parameters.

function handleConnOpenInit(datagram: ConnOpenInit) {
    handler.connOpenInit(
      datagram.identifier,
      datagram.desiredCounterpartyIdentifier,
      datagram.clientIdentifier,
      datagram.counterpartyClientIdentifier,
      datagram.version
    )
}

Recommendation and Reports

We have reported this kind of problems and it has been fixed (ref 1, 2).