Self-describing protocol identifiers
Go to file
decanus 9e718c50c3
updates
2020-02-28 00:24:16 +01:00
LICENSE Initial commit 2020-02-26 16:48:56 +01:00
README.md updates 2020-02-28 00:24:16 +01:00
multiprotocol.csv pushing rough 2020-02-27 18:43:10 +01:00

README.md

multiprotocol: multiformat inspired self-describing protocol identifiers

@TODO MOVE TABLE TO VAC SPECIFIC REPO

This specification describes a simple method for nodes to describe the set of their capabilities. The protocol is inspired by other multiformats, it provides both human and machine-readable representations.

The goal is to provide more granular identification for nodes beyond their connection information as provided by multiaddr.

Multiprotocol is generic in that any protocol can adapt the codes used for their own protocol. A namespace is used to differentiate between protocols, this number should be arbitrary enough to not cause overlap.

Protocol Definitions

Protocol values are defined using a csv table, current implementations support this standard. The CSV file MUST contain a header of the fields defined, these are code, size, name and comment. Their values should be as follows:

field description
code This field contains the code identifying the key.
size This field identifies the expected keys size, it can be any number or V, indicating that the value itself is length prefixed.
name The human readable name of the field.
comment Any developer related comments for the field.

Below is an example valid csv table, the values in it will be used further in the examples within this document.

code, size, name, comment
42,   0,    vac,  namespace
2,    V,    waku,
3,    V,    store,
4,    V,    relay, 

Specification

A multiprotocol, like a multiaddr is a recursive (TLV)+ (type-length-value repeating) encoding. It has two forms:

  • a human-readable version to be used when printing to the user (UTF-8)
  • a binary-packed version to be used in storage, transmissions on the wire, and as a primitive in other formats.

Human-readable

Below is a psuedo regex of the encoding itself.

/<namespace>/<protocol>/<version>(/<capability>/<version>|<capability>)+
  • namespace - the namespace represents the protocol namespace. In our case this would be vac.
  • protocol - the protocol represents the specific protocol we are identifying, in our case waku.
  • version - the version represents the global protocol version, this can be any integer.

Next we have our repeating fields:

  • capability - this represents a specific supported capability, for example store or relay.
  • version - this field is not required, if the size for a specific capability is 0, it represents the supported version, this can either be latest or earliest, we leave this to implementers to decide.

Now, let's look at some human-readable examples:

/vac/waku/2
/vac/waku/2/relay/2
/vac/waku/2/store/1

Machine-readable

@TODO

<namespace uvarint><protocol uvarint><version uvarint>(<protoCode uvarint><value []byte>)+

Examples:

0x2a 0x2 0x1 0x32 # /vac/waku/2
0x2a 0x2 0x1 0x32 0x4 0x1 0x32 # /vac/waku/2/relay/2
0x2a 0x2 0x1 0x32 0x3 0x1 0x32 # /vac/waku/2/store/1

With multiaddr:

/ip4/127.0.0.1/tcp/9000/vac/waku/0.2/relay/0.2

Implementations