Skip to content

identify spec #97

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
May 1, 2019
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
84 changes: 84 additions & 0 deletions identify/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,84 @@
# Identify v1.0.0

The identify protocol is used to exchange basic information with other peers
in the network, including addresses, public keys, and capabilities.

There are two variations of the identify protocol, `identify` and `identify/push`.

### `identify`

The `identify` protocol has the protocol id `/ipfs/id/1.0.0`, and it is used
to query remote peers for their information.

The protocol works by opening a stream to the remote peer you want to query, using
`/ipfs/id/1.0.0` as the protocol id string. The peer being identified responds by returning
an `Identify` message and closes the stream.

### `identify/push`

The `identify/push` protocol has the protocol id `/ipfs/id/push/1.0.0`, and it is used
to inform known peers about changes that occur at runtime.

When a peer's basic information changes, for example, because they've obtained a new
public listen address, they can use `identify/push` to inform others about the new
information.

The push variant works by opening a stream to each remote peer you want to update, using
`/ipfs/id/push/1.0.0` as the protocol id string. When the remote peer accepts the stream,
the local peer will send an `Identify` message and close the stream.

Upon recieving the pushed `Identify` message, the remote peer should update their local
metadata repository with the information from the message. Note that missing fields
should be ignored, as peers may choose to send partial updates containing only the fields
whose values have changed.

## The Identify Message
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should probably note that any missing fields should be ignored (so we can do a partial identify push).

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@yusefnapora want to add it?


```protobuf
message Identify {
optional string protocolVersion = 5;
optional string agentVersion = 6;
optional bytes publicKey = 1;
repeated bytes listenAddrs = 2;
optional bytes observedAddr = 4;
repeated string protocols = 3;
}
```

### protocolVersion

The protocol version identifies the family of protocols used by the peer.
The current protocol version is `ipfs/0.1.0`; if the protocol major or minor
version does not match the protocol used by the initiating peer, then the connection
is considered unusable and the peer must close the connection.

### agentVersion

This is a free-form string, identifying the implementation of the peer.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can we encourage a particular format, even if non-normative? the default is go-libp2p/<semver>, and IPFS changes this to "go-ipfs/" + version.CurrentVersionNumber + "/" + version.CurrentCommit.

https://github.com/ipfs/go-ipfs/blob/master/core/core.go#L99

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nah, let's let people call their agents whatever they want!

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

although an example might be useful.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

example sounds good

The usual format is `agent-name/version`, where `agent-name` is
the name of the program or library and `version` is its semantic version.

### publicKey

This is the public key of the peer, marshalled in binary form as specicfied
in [peer-ids](../peer-ids).


### listenAddrs

These are the addresses on which the peer is listening as multi-addresses.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does the multiaddress also include /ipfs/Qm...? If so, is /p2p also supported?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It doesn't include it. Well, it shouldn't.

### observedAddr
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There are two possibilities here: this can be either the stack of protocols the listener used in order to reach back the dialer, or the stack of protocols we think the dialer used to reach us.
For the sake of forward compatibility, I think it'd be a nice idea to define this more precisely, and I'd be in favour of the former.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added a source disambiguator, per discussion in meatspace.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Per discussion I'm not satisfied with the wording, because I'd like to disambiguate what is returned when p2p-circuit is used, but I can't find any appropriate wording.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We can add an example for circuit addresses specifically, as they seem to be the contentious issue.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added some examples, hopefully it's clearer now.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

maybe something like: "This address describes the dialed peer's observed route to the dialing peer."

so wordy tho

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yeah, and it's not really any better than the current wording.


This is the connection source address of the stream initiating peer as observed by the peer
being identified; it is a multi-address. The initiator can use this address to infer
the existence of a NAT and its public address.

For example, in the case of a TCP/IP transport the observed addresses will be of the form
`/ip4/x.x.x.x/tcp/xx`. In the case of a circuit relay connection, the observed address will
be of the form `/p2p/QmRelay/p2p-circuit`. In the case of onion transport, there is no
observable source address.

### protocols

This is a list of protocols supported by the peer.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not ordered in any specific manner.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

right, but do we need to say that?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A list implies order. If order is not important, then we can say it's an unordered set. We shouldn't be sending duplicates anyway, right?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

it's a list on the wire, there is no set datatype.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

might be worth clarifying that this are multistream protocols (or protocol strings)?