- Author(s): @xandkar, @Vagabond, Helium Systems, Inc. team
- Start Date: 2021-04-19
- Category: Technical
- Original HIP PR: #154
- Tracking Issue: #157
Hotspot owners, HNT holders, and other Helium participants, have different ways to organize assets. This HIP proposes a multi-signature key which can be used to authorize a transaction on the Helium blockchain. This key is a composite of the scalar keys which can be combined to generate a minimal subset of signatures required to make a valid transaction.
We have already seen that Hotspot owners are already engaging in interesting ways to own their Hotspots off-chain. This HIP proposes a way to allow for the following:
- Dividing up the responsibility of various assets (Hotspots, Validators, OUIs, HST, HNT, etc.) across multiple participants. This enables joint accounts and shared custody.
- Making it substantially more difficult for a compromise of an entire wallet due to a single key. This improves security via multi-owner verification.
- Creating a M of N backup where a loss of a single key does not cause a loss of the entire wallet.
Multi-signature keys are seen in many other blockchains and enable security applications not listed above.
This change should not affect any current asset owners as it is a purely additive change to existing key types.
We believe that this feature will enable new ownership organization of assets like Hotspots, Validators, OUIs, HNT, HST, etc.
Currently, a wallet address is derived from a single public key owned by a single private key. This private key is required to sign any transaction involving the address.
We propose:
- an address scheme composed of a multihash digest of
N
public keys and a specification of the number of signatures required to approve a transaction:M
such that0 < M <= N
; - a signature scheme composed of
N
public keys andM
corresponding signatures; - conditions for verification of the above composite signature, given the above composite public key;
- an implementation of all of the above.
Roughly, the schemes are:
- Multi-signature public key:
| net-type | key-type | m | n | multihash-digest-of-n-public-keys |
; - Multi-signature:
| public-key[0]..public-key[n-1] | isig[1]..isig[m] |
.
More precisely, given a sequence of n
public keys: public-key[0]..public-key[n-1]
, a
multihash-digest-of-n-public-keys
is its multihash digest and isig[1]..isig[m]
is a sequence of
triples: i | len | sig
where i
is the index into the preceding sequence of public keys, len
is
the length of the following signature (in bytes) and sig
is the signature.
net-type key-type m n multihash-digest-of-n-public-keys
+----------+----------+--------+--------+-----------------------------------+
| 4 bits | 4 bits | 8 bits | 8 bits | variable length |
+----------+----------+--------+--------+-----------------------------------+
pub keys isigs
+-----------------+------------------+
| variable length | variable length |
+-----------------+------------------+
Due to variable lengths of the components, parsing of the multi-signature requires knowledge of:
N
from the corresponding multi-pub-key;- length of each scalar key type (type itself is encoded, but the length corresponding to type is not, so must be known in advance).
After parsing the scalar components of the multi-pub-key and multi-signature, we consider the multi-signature to be valid if all of the following conditions are true:
- we have at least
M
scalar signatures within the composite multi-signature; - all of the provided scalar signature triples uniquely point to a public key within
0..N-1
range (i.e. any isig containing an-already-seen index - invalidates the multi-signature); - at least
M
of the scalar signatures pass verification.
We don't believe there are any drawbacks to this approach. In fact, it is an in-place upgrade to the existing supported key types.
An alternative that was considered, but rejected, was to have separate on-chain multisignature
transactions as we do for the vars_v1
transaction. This approach to multisignature keys makes that
approach unnecessary. We do not, at this time, intend to update vars_v1
but may do so at a later
date.
We don't believe there is any deployment impact as this approach is additive to existing key types.
This will be a successful feature if we see applications of multi-signature keys on chain.