Create a gist now

Instantly share code, notes, and snippets.

Embed
addrv2 BIP proposal draft

  BIP: ??? (tbr after sending to mailing list)
  Layer: Peer Services
  Title: addrv2 message
  Author: Wladimir J. van der Laan <laanwj@gmail.com>
  Comments-Summary: No comments yet.
  Comments-URI: 
  Status: Draft
  Type: Standards Track
  Created: 2018-06-01
  License: BSD-2-Clause

Table of Contents

Introduction

Abstract

This document proposes a new P2P message to gossip longer node addresses over the P2P network. This is required to support new-generation Onion addresses, I2P, and potentially other networks that have longer endpoint addresses than fit in the 128 bits of the current addr message.

Copyright

This BIP is licensed under the 2-clause BSD license.

Motivation

Tor v3 hidden services are part of the stable release of Tor since version 0.3.2.9. They have various advantages compared to the old hidden services, among which better encryption and privacy [1]. These services have 256 bit addresses and thus do not fit in the existing addr message, which encapsulates onion addresses in OnionCat IPv6 addresses.

Other transport-layer protocols such as I2P have always used longer addresses. This change would make it possible to gossip such addresses over the P2P network, so that other peers can connect to them.

Specification

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119[2].

The addrv2 message is defined as a message where pchCommand == "addrv2". It is serialized in the standard encoding for P2P messages. Its format is similar to the current addr message format [3], with the difference that the fixed 16-byte IP address is replaced by a network ID and a variable-length address, and the time and services format has been changed to VARINT.

This means that the message contains a serialized std::vector of the following structure:

Type Name Description
VARINT (unsigned) time Time that this node was last seen as connected to the network. A time in Unix epoch time format, up to 64 bits wide.
VARINT (unsigned) services Service bits. A 64-wide bit field.
uint8_t networkID Network identifier. An 8-bit value that specifies which network is addressed.
std::vector<uint8_t> addr Network address. The interpretation depends on networkID.
uint16_t port Network port. If not relevant for the network this MUST be 0.

One message can contain up to 1,000 addresses. Clients SHOULD reject messages with more addresses.

Field addr has a variable length, with a maximum of 32 bytes (256 bits). Clients SHOULD reject longer addresses.

The list of reserved network IDs is as follows:

Network ID Enumeration Address length (bytes) Description
0x01 IPV4 4 IPv4 address (globally routed internet)
0x02 IPV6 16 IPv6 address (globally routed internet)
0x03 TORV2 10 Tor v2 hidden service address
0x04 TORV3 32 Tor v3 hidden service address
0x05 I2P 32 I2P overlay network address
0x06 CJDNS 16 Cjdns overlay network address

To allow for future extensibility, clients MUST ignore address types that they do not know about. Client MAY store and gossip address formats that they do not know about. Further network ID numbers MUST be reserved in a new BIP document.

Clients SHOULD reject addresses that have a different length than specified in this table for a specific address ID, as these are meaningless.

See the appendices for the address encodings to be used for the various networks.

Compatibility

Send addrv2 messages only, and exclusively, when the peer has a certain protocol version (or higher):

//! gossiping using `addrv2` messages starts with this version
static const int GOSSIP_ADDRV2_VERSION = 70016;
For older peers keep sending the legacy addr message, ignoring addresses with the newly introduced address types.

Reference implementation

The reference implementation is available at (to be done)

Considerations

(to be discussed)

  • Client MAY store and gossip address formats that they do not know about: does it ever make sense to gossip addresses outside a certain overlay network? Say, I2P addresses to Tor? I'm not sure. Especially for networks that have no exit nodes as there is no overlap with the globally routed internet at all.
  • Lower precision of time field? seconds precision seems overkill, and can even be harmful, there have been attacks that exploited high precision timestamps for mapping the current network topology.
    • (gmaxwell) If you care about space time field could be reduced to 16 bits easily. Turn it into a "time ago seen" quantized to 1 hour precision. (IIRC we quantize times to 2hrs regardless).
  • Rolling port into addr, or making the port optional, would make it possible to shave off two bytes for address types that don't have ports (however, all of the currently listed formats have a concept of port.). It could also be an optional data item (see below).
  • (gmaxwell) Optional (per-service) data could be useful for various things:
    • Node-flavors for striping (signalling which slice of the blocks the node has in selective pruning)
    • Payload for is alternative ports for other transports (e.g. UDP ports)
    • If we want optional flags. I guess the best thing would just be a byte to include the count of them, then a byte "type" for each one where the type also encodes if the payload is 0/8/16/32 bits. (using the two MSB of the type to encode the length). And then bound the count of them so that the total is still reasonably sized.

Rejected

(ideas that were considered but ae probably not the way to go)

  • Another option for backwards compatbility would be: Old clients will ignore the addrv2 message, so using addrv2 format for only the larger addresses (TORV3, I2P) would naturally preserve backward compatibility. However, there are some disadvantages as well:
    • The advantage of more compact representation of IPv4 addresses would be lost
    • There is no path toward deprecating legacy addr

Acknowledgements

- Jonas Schnelli: change services field to VARINT, to make the message more compact in the likely case instead of always using 8 bytes.

- Luke-Jr: change time field to VARINT, for post-2038 compatibility.

- Gregory Maxwell: various suggestions regarding extensibility

Appendix A: Tor v2 address encoding

The new message introduces a separate network ID for TORV2.

Clients MUST send Tor hidden service addresses with this network ID, with the 80-bit hidden service ID in the address field. This is the same as the representation in the legacy addr message, minus the 6 byte prefix of the OnionCat wrapping.

Clients SHOULD ignore OnionCat (fd87:d87e:eb43::/48) addresses on receive if they come with the IPV6 network ID.

Appendix B: Tor v3 address encoding

According to the spec [4], next-gen .onion addresses are encoded as follows:

onion_address = base32(PUBKEY | CHECKSUM | VERSION) + ".onion"
 CHECKSUM = H(".onion checksum" | PUBKEY | VERSION)[:2]

 where:
   - PUBKEY is the 32 bytes ed25519 master pubkey of the hidden service.
   - VERSION is an one byte version field (default value '\x03')
   - ".onion checksum" is a constant string
   - CHECKSUM is truncated to two bytes before inserting it in onion_address

Tor v3 addresses MUST be sent with the TORV3 network ID, with the 32-byte PUBKEY part in the address field. As VERSION will always be '\x03' in the case of v3 addresses, this is enough to reconstruct the onion address.

Appendix C: I2P address encoding

Like Tor, I2P naming uses a base32-encoded address format[5].

I2P uses 52 characters (256 bits) to represent the full SHA-256 hash, followed by .b32.i2p.

I2P addresses MUST be sent with the I2P network ID, with the decoded SHA-256 hash as address field.

Appendix D: Cjdns address encoding

Cjdns addresses are simply IPv6 addresses in the fc00::/8 range[6]. They MUST be sent with the CJDNS network ID.

References

  1. ^ Tor Rendezvous Specification - Version 3
  2. ^ RFC 2119
  3. ^ Bitcoin Developer Reference: addr message
  4. ^ Tor Rendezvous Specification - Version 3: Encoding onion addresses
  5. ^ I2P: Naming and address book
  6. ^ Cjdns whitepaper: Pulling It All Together

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment