Overview

Unified Addresses specify multiple methods for payment to a Recipient's wallet. The Sender's wallet can then non-interactively select the method of payment.

Importantly, any wallet can support Unified Addresses, even when that wallet only supports a subset of payment methods for receiving and/or sending.

Despite having some similar characteristics, the Unified Address standard is orthogonal to Payment Request URIs 29 and similar schemes. Since Payment Requests encode addresses as alphanumeric strings, no change to ZIP 321 is required in order to use Unified Addresses in Payment Requests.

Concepts

Wallets follow a model Interaction Flow as follows:

1. A Producer generates an Address.
2. The Producer encodes the Address.
3. The Producer wallet or human user distributes this Address Encoding. This ZIP leaves distribution mechanisms out of scope.
4. A Consumer wallet or user imports the Address Encoding through any of a variety of mechanisms (QR code scanning, Payment URIs, cut-and-paste, or “in-band” protocols like Reply-To memos).
5. A Consumer wallet decodes the Address Encoding and performs validity checks.
6. (Perhaps later in time) if the Consumer wallet is a Sender, it can execute a transfer of ZEC (or other assets or protocol state changes) to the Address.

Encodings of the same Address may be distributed zero or more times through different means. Zero or more Consumers may import Addresses. Zero or more of those (that are Senders) may execute a Transfer. A single Sender may execute multiple Transfers over time from a single import.

Steps 1 to 5 inclusive also apply to Interaction Flows for Unified Full Viewing Keys and Unified Incoming Viewing Keys.

Addresses

A Unified Address (or UA for short) combines one or more Receivers.

When new Transport Protocols are introduced to the Zcash protocol after Unified Addresses are standardized, those should introduce new Receiver Types but not different Address types outside of the UA standard. There needs to be a compelling reason to deviate from the standard, since the benefits of UA come precisely from their applicability across all new protocol upgrades.

Receivers

Every wallet must properly parse encodings of a Unified Address or Unified Viewing Key containing unrecognised Items.

A wallet may process unrecognised Items by indicating to the user their presence or similar information for usability or diagnostic purposes.

Transport Encoding

The Unified String Encoding is “opaque” to human readers: it does not allow visual identification of which Receivers or Receiver Types are present.

The Unified String Encoding is resilient against typos, transcription errors, cut-and-paste errors, truncation, or other likely UX hazards.

There is a well-defined Unified QR Encoding of a Unified Address (or UFVK or UIVK) as a QR code, which produces QR codes that are reasonably compact and robust.

There is a well-defined transformation between the Unified QR Encoding and Unified String Encoding of a given UA/UVK in either direction.

The Unified String Encoding fits into ZIP-321 Payment URIs 29 and general URIs without introducing parse ambiguities.

The encoding must support sufficiently many Recipient Types to allow for reasonable future expansion.

The encoding must allow all wallets to safely and correctly parse out unrecognised Receiver Types well enough to ignore them.

Transfers

When executing a Transfer the Sender selects a Receiver via a Selection process.

Given a valid UA, Selection must treat any unrecognised Item as though it were absent.

• This property is crucial for forward compatibility to ensure users who upgrade to newer protocols / UAs don't lose the ability to smoothly interact with older wallets.
• This property is crucial for allowing Transparent-Only UA-Conformant wallets to interact with newer shielded wallets, removing a disincentive for adopting newer shielded wallets.
• This property also allows Transparent-Only wallets to upgrade to shielded support without re-acquiring counterparty UAs. If they are re-acquired, the user flow and usability will be minimally disrupted.

Experimental Usage

Unified Addresses and Unified Viewing Keys must be able to include Receivers and Viewing Keys of experimental types, possibly alongside non-experimental ones. These experimental Receivers or Viewing Keys must be used only by wallets whose users have explicitly opted into the corresponding experiment.

Viewing Keys

A Unified Full Viewing Key (resp. Unified Incoming Viewing Key) can be used in a similar way to a Full Viewing Key (resp. Incoming Viewing Key) as described in the Zcash Protocol Specification 2.

For a Transparent P2PKH Address that is derived according to BIP 32 30 and BIP 44 33, the nearest equivalent to a Full Viewing Key or Incoming Viewing Key for a given BIP 44 account is an extended public key, as defined in the section “Extended keys” of BIP 32. Therefore, UFVKs and UIVKs should be able to include such extended public keys.

A wallet should support deriving a UIVK from a UFVK, and a Unified Address from a UIVK.

Open Issues and Known Concerns

Privacy impacts of transparent or cross-pool transactions, and the associated UX issues, will be addressed in ZIP 315 (in preparation).

Revisions

• Revision 0: The initial version of this specification.

• Revision 1: This version adds support for MUST-understand Typecodes and Address Expiration Metadata. It also drops the restriction that a UA/UVK must contain at least one shielded Item.

A Revision 0 UA/UVK is distinguished from a Revision 1 UA/UVK by its Human-Readable Part, as follows.

Let prefix be: * “u”, if this is a UA/UVK prior to Revision 1; * “ur”, if this is a UA/UVK from Revision 1 onward.

The Human-Readable Parts (as defined in 36) of Unified Addresses are defined as: * prefix, for Unified Addresses on Mainnet; * prefix || “test”, for Unified Addresses on Testnet.

The Human-Readable Parts of Unified Viewing Keys are defined as:

• prefix || “ivk” for Unified Incoming Viewing Keys on Mainnet;
• prefix || “ivktest” for Unified Incoming Viewing Keys on Testnet;
• prefix || “view” for Unified Full Viewing Keys on Mainnet;
• prefix || “viewtest” for Unified Full Viewing Keys on Testnet.

Encoding of Unified Addresses

Rather than defining a Bech32 string encoding of Orchard Shielded Payment Addresses, we instead define a Unified Address format that is able to encode a set of Receivers of different types. This enables the Consumer of a Unified Address to choose the Receiver of the best type it supports, providing a better user experience as new Receiver Types are added in the future.

Assume that we are given a set of one or more Receiver Encodings for distinct types. That is, the set may optionally contain one Receiver of each of the Receiver Types in the following fixed Priority List:

• Typecode \(\mathtt{0x03}\) — an Orchard raw address as defined in 11;
• Typecode \(\mathtt{0x02}\) — a Sapling raw address as defined in 10;
• Typecode \(\mathtt{0x01}\) — a Transparent P2SH address, or Typecode \(\mathtt{0x00}\) — a Transparent P2PKH address.

If, and only if, the user of a Producer or Consumer wallet explicitly opts into an experiment as described in Experimental Usage, the specification of the experiment MAY include additions to the above Priority List (such additions SHOULD maintain the intent of preferring more recent shielded protocols).

We say that a Receiver Type is “preferred” over another when it appears earlier in this Priority List (as potentially modified by experiments).

The Sender of a payment to a Unified Address MUST use the Receiver of the most preferred Receiver Type that it supports from the set.

For example, consider a wallet that supports sending funds to Orchard Receivers, and does not support sending to any Receiver Type that is preferred over Orchard. If that wallet is given a UA that includes an Orchard Receiver and possibly other Receivers, it MUST send to the Orchard Receiver.

The raw encoding of a Unified Address is a concatenation of \((\mathtt{typecode}, \mathtt{length}, \mathtt{addr})\) encodings of the consituent Receivers, in ascending order of Typecode:

• \(\mathtt{typecode} : \mathtt{compactSize}\) — the Typecode from the above Priority List;
• \(\mathtt{length} : \mathtt{compactSize}\) — the length in bytes of \(\mathtt{addr};\)
• \(\mathtt{addr} : \mathtt{byte[length]}\) — the Receiver Encoding.

The values of the \(\mathtt{typecode}\) and \(\mathtt{length}\) fields MUST be less than or equal to \(\mathtt{0x2000000}.\) (The limitation on the total length of encodings described below imposes a smaller limit for \(\mathtt{length}\) in practice.)

A Receiver Encoding is the raw encoding of a Shielded Payment Address, or the \(160\!\) -bit script hash of a P2SH address 38, or the \(160\!\) -bit validating key hash of a P2PKH address 37.

Let padding be the Human-Readable Part of the Unified Address in US-ASCII, padded to 16 bytes with zero bytes. We append padding to the concatenated encodings, and then apply the \(\mathsf{F4Jumble}\) algorithm as described in Jumbling. (In order for the limitation on the \(\mathsf{F4Jumble}\) input size to be met, the total length of encodings MUST be at most \(\ell^\mathsf{MAX}_M - 16\) bytes, where \(\ell^\mathsf{MAX}_M\) is defined in Jumbling.) The output is then encoded with Bech32m 36, ignoring any length restrictions. This is chosen over Bech32 in order to better handle variable-length inputs.

To decode a Unified Address Encoding, a Consumer MUST use the following procedure:

• Decode using Bech32m, rejecting any address with an incorrect checksum.
• Apply \(\mathsf{F4Jumble}^{-1}\) (this can also reject if the input is not in the correct range of lengths).
• Let padding be the Human-Readable Part, padded to 16 bytes as for encoding. If the result ends in padding, remove these 16 bytes; otherwise reject.
• Parse the result as a raw encoding as described above, rejecting the entire Unified Address if it does not parse correctly.

The Human-Readable Part is as specified for a Unified Address in Revisions.

A wallet MAY allow its user(s) to configure which Receiver Types it can send to. It MUST NOT allow the user(s) to change the order of the Priority List used to choose the Receiver Type, except by opting into experiments.

Encoding of Unified Full/Incoming Viewing Keys

Unified Full or Incoming Viewing Keys are encoded and decoded analogously to Unified Addresses. A Consumer MUST use the decoding procedure from the previous section. For Viewing Keys, a Consumer will normally take the union of information provided by all contained Receivers, and therefore the Priority List defined in the previous section is not used.

For each FVK Type or IVK Type currently defined in this specification, the same Typecode is used as for the corresponding Receiver Type in a Unified Address. Additional FVK Types and IVK Types MAY be defined in future, and these will not necessarily use the same Typecode as the corresponding Unified Address.

The following FVK or IVK Encodings are used in place of the \(\mathtt{addr}\) field:

• An Orchard FVK or IVK Encoding, with Typecode \(\mathtt{0x03},\) is is the raw encoding of the Orchard Full Viewing Key or Orchard Incoming Viewing Key respectively.
• A Sapling FVK Encoding, with Typecode \(\mathtt{0x02},\) is the encoding of \((\mathsf{ak}, \mathsf{nk}, \mathsf{ovk}, \mathsf{dk})\) given by \(\mathsf{EncodeExtFVKParts}(\mathsf{ak}, \mathsf{nk}, \mathsf{ovk}, \mathsf{dk})\) , where \(\mathsf{EncodeExtFVKParts}\) is defined in 15. This SHOULD be derived from the Extended Full Viewing Key at the Account level of the ZIP 32 hierarchy.
• A Sapling IVK Encoding, also with Typecode \(\mathtt{0x02},\) is an encoding of \((\mathsf{dk}, \mathsf{ivk})\) given by \(\mathsf{dk}\,||\,\mathsf{I2LEOSP}_{256}(\mathsf{ivk}).\)
• There is no defined way to represent a Viewing Key for a Transparent P2SH Address in a UFVK or UIVK (because P2SH Addresses cannot be diversified in an unlinkable way). The Typecode \(\mathtt{0x01}\) MUST NOT be included in a UFVK or UIVK by Producers, and MUST be treated as unrecognised by Consumers.
• For Transparent P2PKH Addresses that are derived according to BIP 32 30 and BIP 44 33, the FVK and IVK Encodings have Typecode \(\mathtt{0x00}.\) Both of these are encodings of the chain code and public key \((\mathsf{c}, \mathsf{pk})\) given by \(\mathsf{c}\,||\,\mathsf{ser_P}(\mathsf{pk})\) . (This is the same as the last 65 bytes of the extended public key format defined in section “Serialization format” of BIP 32 31.) However, the FVK uses the key at the Account level, i.e. at path \(m / 44' / coin\_type' / account'\) , while the IVK uses the external (non-change) child key at the Change level, i.e. at path \(m / 44' / coin\_type' / account' / 0\) .

The Human-Readable Part is as specified for a Unified Viewing Key in Revisions.

Requirements for both Unified Addresses and Unified Viewing Keys

• A Revision 0 Unified Address or Unified Viewing Key MUST contain at least one shielded Item (Typecodes \(\mathtt{0x02}\) and \(\mathtt{0x03}\) ). This requirement is dropped for Revision 1 UA/UVKs.
• The \(\mathtt{typecode}\) and \(\mathtt{length}\) fields are encoded as \(\mathtt{compactSize}.\) 39 (Although existing Receiver Encodings and Viewing Key Encodings are all less than 256 bytes and so could use a one-byte length field, encodings for experimental types may be longer.)
• Within a single UA or UVK, all HD-derived Receivers, FVKs, and IVKs SHOULD represent an Address or Viewing Key for the same account (as used in the ZIP 32 or BIP 44 Account level).
• For Transparent Addresses, the Receiver Encoding does not include the first two bytes of a raw encoding.
• There is intentionally no Typecode defined for a Sprout Shielded Payment Address or Sprout Incoming Viewing Key. Since it is no longer possible (since activation of ZIP 211 in the Canopy network upgrade 25) to send funds into the Sprout chain value pool, this would not be generally useful.
• With the exception of MUST-understand Metadata Items, Consumers MUST ignore constituent Items with Typecodes they do not recognise.
• Consumers MUST reject Unified Addresses/Viewing Keys in which the same Typecode appears more than once, or that include both P2SH and P2PKH Transparent Addresses, or that contain only a Transparent Address.
• Consumers MUST reject Unified Addresses/Viewing Keys in which any constituent Item does not meet the validation requirements of its encoding, as specified in this ZIP and the Zcash Protocol Specification 2.
• Consumers MUST reject Unified Addresses/Viewing Keys in which the constituent Items are not ordered in ascending Typecode order. Note that this is different to priority order, and does not affect which Receiver in a Unified Address should be used by a Sender.
• There MUST NOT be additional bytes at the end of the raw encoding that cannot be interpreted as specified above.
• If the encoding of a Unified Address/Viewing Key is shown to a user in an abridged form due to lack of space, at least the first 20 characters MUST be included.

Adding new types

It is intended that new Receiver Types and Viewing Key Types SHOULD be introduced either by a modification to this ZIP or by a new ZIP, in accordance with the ZIP Process 14.

For experimentation prior to proposing a ZIP, experimental types MAY be added using the reserved Typecodes \(\mathtt{0xFFFA}\) to \(\mathtt{0xFFFF}\) inclusive. This provides for six simultaneous experiments, which can be referred to as experiments A to F. This should be sufficient because experiments are expected to be reasonably short-term, and should otherwise be either standardized in a ZIP (and allocated a Typecode outside this reserved range) or discontinued.

New types SHOULD maintain the same distinction between FVK and IVK authority as existing types, i.e. an FVK is intended to give access to view all transactions to and from the address, while an IVK is intended to give access only to view incoming payments (as opposed to change).

Metadata Items

Typecodes \(\mathtt{0xC0}\) to \(\mathtt{0xFC}\) inclusive are reserved to indicate Metadata Items other than Receivers or Viewing Keys. These Items MAY affect the overall interpretation of the UA/UVK (for example, by specifying an expiration date).

As of Revision 1 of this ZIP, the subset of Metadata Typecodes in the range \(\mathtt{0xE0}\) to \(\mathtt{0xFC}\) inclusive are designated as "MUST-understand": if a Consumer is unable to recognise the meaning of a Metadata Item with a Typecode in this range, then it MUST regard the entire UA/UVK as unsupported and not process it further.

A Revision 0 UA/UVK (determined by its HRP as specified in Revisions) MUST NOT include any Metadata Items with a MUST-understand Typecode; a Consumer MUST reject as invalid any UA/UVK that violates this requirement.

Since Metadata Items are not Receivers, they MUST NOT be selected by a Sender when choosing a Receiver to send to, and since they are not Viewing Keys, they MUST NOT provide additional authority to view information about transactions.

New Metadata Types SHOULD be introduced either by a modification to this ZIP or by a new ZIP, in accordance with the ZIP Process 14.

Address Expiration Metadata

As of Revision 1, Typecodes \(\mathtt{0xE0}\) and \(\mathtt{0xE1}\) are reserved for optional address expiry metadata. A producer MAY choose to generate Unified Addresses containing either or both of the following Metadata Item Types, or none.

The value of a \(\mathtt{0xE0}\) item MUST be an unsigned 32-bit integer in little-endian order specifying the Address Expiry Height, a block height of the Zcash chain associated with the UA/UVK. A Unified Address containing metadata Typecode \(\mathtt{0xE0}\) MUST be considered expired when the height of the Zcash chain is greater than this value.

The value of a \(\mathtt{0xE1}\) item MUST be an unsigned 64-bit integer in little-endian order specifying a Unix Epoch Time, hereafter referred to as the Address Expiry Time. A Unified Address containing Metadata Typecode \(\mathtt{0xE1}\) MUST be considered expired when the current time is after the Address Expiry Time.

A Sender that supports Revision 1 of this specification MUST set a non-zero nExpiryHeight field in transactions it creates that are sent to a Unified Address that defines an Address Expiry Height. If the nExpiryHeight normally constructed by the Sender would be greater than the Address Expiry Height, then the transaction MUST NOT be sent. If only an Address Expiry Time is specified, then the Sender SHOULD choose a value for nExpiryHeight such that the transaction will expire no more than 24 hours after the current time. If both \(\mathtt{0xE0}\) and \(\mathtt{0xE1}\) Metadata Items are present, then both restrictions apply.

If a Sender sends to multiple Unified Addresses in the same transaction, then all of the Address Expiry constraints imposed by the individual addresses apply.

If a wallet user attempts to send to an expired address, the error presented to the user by the wallet SHOULD include a suggestion that the user should attempt to obtain a currently-valid address for the intended recipient. A wallet MUST NOT send to an address that it knows to have expired.

Address expiration imposes no constraints on the Producer of an address. A Producer MAY generate multiple Unified Addresses with the same Receivers but different expiration metadata and/or any number of distinct Diversified Unified Addresses with the same or different expiry metadata, in any combination. Note that although changes to metadata will result in a visually distinct address, such updated addresses will be directly linkable to the original addresses because they share the same Receivers.

When deriving a UIVK from a UFVK containing Typecodes \(\mathtt{0xE0}\) and/or \(\mathtt{0xE1}\) , these Metadata Items MUST be retained unmodified in the derived UIVK.

When deriving a Unified Address from a UFVK or UIVK containing a Metadata Item having Typecode \(\mathtt{0xE0}\) , the derived Unified Address MUST contain a Metadata Item having Typecode \(\mathtt{0xE0}\) such that the Address Expiry Height of the resulting address is less than or equal to the Expiry Height of the viewing key.

When deriving a Unified Address from a UFVK or UIVK containing a Metadata Item having Typecode \(\mathtt{0xE1}\) , the derived Unified Address MUST contain a Metadata Item having Typecode \(\mathtt{0xE1}\) such that the Address Expiry Time of the resulting address is less than or equal to the Expiry Time of the viewing key.

Producers of Diversified Unified Addresses should be aware that the expiration metadata could potentially be used to link addresses from the same source. Normally, if Diversified Unified Addresses derived from the same UIVK contain only Sapling and/or Orchard Receivers and no Metadata Items, they will be unlinkable as described in 8; this property does not hold when Metadata Items are present. It is RECOMMENDED that when deriving Unified Addresses from a UFVK or UIVK containing expiry metadata that the Expiry Height and Expiry Time of each distinct address vary from one another, so as to reduce the likelihood that addresses may be linked via their expiry metadata.

Deriving Internal Keys

In addition to external addresses suitable for giving out to Senders, a wallet typically requires addresses for internal operations such as change and auto-shielding.

We desire the following properties for viewing authority of both shielded and transparent key trees:

• A holder of an FVK can derive external and internal IVKs, and external and internal \(\mathsf{ovk}\) components.
• A holder of the external IVK cannot derive the internal IVK, or any of the \(\mathsf{ovk}\) components.
• A holder of the external \(\mathsf{ovk}\) component cannot derive the internal \(\mathsf{ovk}\) component, or any of the IVKs.

For shielded keys, these properties are achieved by the one-wayness of \(\mathsf{PRF^{expand}}\) and of \(\mathsf{CRH^{ivk}}\) or \(\mathsf{Commit^{ivk}}\) (for Sapling and Orchard respectively). Derivation of an internal shielded FVK from an external shielded FVK is specified in the "Sapling internal key derivation" 18 and "Orchard internal key derivation" 20 sections of ZIP 32.

To satisfy the above properties for transparent (P2PKH) keys, we derive the external and internal \(\mathsf{ovk}\) components from the transparent FVK \((\mathsf{c}, \mathsf{pk})\) (described in Encoding of Unified Full/Incoming Viewing Keys) as follows:

• Let \(I_\mathsf{ovk} = \mathsf{PRF^{expand}}_{\mathsf{LEOS2BSP}_{256}(\mathsf{c})}\big([\mathtt{0xd0}] \,||\, \mathsf{ser_P}(\mathsf{pk})\big)\) where \(\mathsf{ser_P}(pk)\) is \(33\) bytes, as specified in 31.
• Let \(\mathsf{ovk_{external}}\) be the first \(32\) bytes of \(I_\mathsf{ovk}\) and let \(\mathsf{ovk_{internal}}\) be the remaining \(32\) bytes of \(I_\mathsf{ovk}\) .

Since an external P2PKH FVK encodes the chain code and public key at the Account level, we can derive both external and internal child keys from it, as described in BIP 44 34. It is possible to derive an internal P2PKH FVK from the external P2PKH FVK (i.e. its parent) without having the external spending key, because child derivation at the Change level is non-hardened.

Deriving a UIVK from a UFVK

As a consequence of the specification in MUST-understand Typecodes, as of Revision 1 a Consumer of a UFVK MUST understand the meaning of all "MUST-understand" Metadata Item Typecodes present in the UFVK in order to derive a UIVK from it.

If the source UFVK is Revision 1 then the derived UIVK SHOULD be Revision 1; if the source UFVK includes any Metadata Items then the derived UIVK MUST be Revision 1.

For Metadata Items recognised by the Consumer, the processing of the Item when deriving a UIVK is specified in the section or ZIP describing that Item.

The following derivations are applied to each component FVK:

• For a Sapling FVK, the corresponding Sapling IVK is obtained as specified in 4.
• For an Orchard FVK, the corresponding Orchard IVK is obtained as specified in 5.
• For a Transparent P2PKH FVK, the corresponding Transparent P2PKH IVK is obtained by deriving the child key with non-hardened index \(0\) as described in 32.

In each case, the Typecode remains the same as in the FVK.

Items (including Metadata Items that are not "MUST-understand") that are unrecognised by a given Consumer, or that are specified in experiments that the user has not opted into (see Experimental Usage), MUST be dropped when the Consumer derives a UIVK from a UFVK.

Deriving a Unified Address from a UIVK

As a consequence of the specification in MUST-understand Typecodes, as of Revision 1 a Consumer of a UIVK MUST understand the meaning of all "MUST-understand" Metadata Item Typecodes present in the UIVK in order to derive a Unified Address from it.

If the source UIVK is Revision 1 then the derived Unified Address SHOULD be Revision 1; if the source UIVK includes any Metadata Items then the derived Unified Address MUST be Revision 1.

For Metadata Items recognised by the Consumer, the processing of the Item when deriving a Unified Address is specified in the section or ZIP describing that Item.

To derive a Unified Address from a UIVK we need to choose a diversifier index, which MUST be valid for all of the Viewing Key Types in the UIVK. That is,

• A Sapling diversifier index MUST generate a valid diversifier as defined in ZIP 32 section “Sapling diversifier derivation” 17.
• A Transparent diversifier index MUST be in the range \(0\) to \(2^{31} - 1\) inclusive.
• There are no additional constraints on an Orchard diversifier index.

The following derivations are applied to each component IVK using the diversifier index:

• For a Sapling IVK, the corresponding Sapling Receiver is obtained as specified in 4.
• For an Orchard IVK, the corresponding Orchard Receiver is obtained as specified in 5.
• For a Transparent P2PKH IVK, the diversifier index is used as a BIP 44 child key index at the Index level 35 to derive the corresponding Transparent P2PKH Receiver. As is usual for derivations below the BIP 44 Account level, non-hardened (public) derivation 32 MUST be used. The IVK is assumed to correspond to the extended public key for the external (non-change) element of the path. That is, if the UIVK was constructed correctly then the BIP 44 path of the Transparent P2PKH Receiver will be \(m / 44' / \mathit{coin\_type\kern0.05em'} / \mathit{account\kern0.1em'} / 0 / \mathit{diversifier\_index}.\)

In each case, the Typecode remains the same as in the IVK.

Items (including Metadata Items that are not "MUST-understand") that are unrecognised by a given Consumer, or that are specified in experiments that the user has not opted into (see Experimental Usage), MUST be dropped when the Consumer derives a Unified Address from a UIVK.

See Address Expiration Metadata for discussion of potential linking of Diversified Unified Addresses via their metadata.

Usage of Outgoing Viewing Keys

When a Sender constructs a transaction that creates Sapling or Orchard notes, it uses an outgoing viewing key, as described in 6 and 7, to encrypt an outgoing ciphertext. Decryption with the outgoing viewing key allows recovering the sent note plaintext, including destination address, amount, and memo. The intention is that this outgoing viewing key should be associated with the source of the funds.

However, the specification of which outgoing viewing key should be used is left somewhat open in 6 and 7; in particular, it was unclear whether transfers should be considered as being sent from an address, or from a ZIP 32 account 21. The adoption of multiple shielded protocols that support outgoing viewing keys (i.e. Sapling and Orchard) further complicates this question, since from NU5 activation, nothing at the consensus level prevents a wallet from spending both Sapling and Orchard notes in the same transaction. (Recommendations about wallet usage of multiple pools will be given in ZIP 315 27.)

Here we refine the protocol specification in order to allow more precise determination of viewing authority for UFVKs.

A Sender will attempt to determine a "sending Account" for each transfer. The preferred approach is for the API used to perform a transfer to directly specify a sending Account. Otherwise, if the Sender can ascertain that all funds used in the transfer are from addresses associated with some Account, then it SHOULD treat that as the sending Account. If not, then the sending Account is undetermined.

The Sender also determines a "preferred sending protocol" —one of "transparent", "Sapling", or "Orchard"— corresponding to the most preferred Receiver Type (as given in Encoding of Unified Addresses) of any funds sent in the transaction.

If the sending Account has been determined, then the Sender SHOULD use the external or internal \(\mathsf{ovk}\) (according to the type of transfer), as specified by the preferred sending protocol, of the full viewing key for that Account (i.e. at the ZIP 32 Account level).

If the sending Account is undetermined, then the Sender SHOULD choose one of the addresses, restricted to addresses for the preferred sending protocol, from which funds are being sent (for example, the first one for that protocol), and then use the external or internal \(\mathsf{ovk}\) (according to the type of transfer) of the full viewing key for that address.

Jumbling

Security goal (near second preimage resistance):

• An adversary is given \(q\) Unified Addresses/Viewing Keys, generated honestly.
• The attack goal is to produce a “partially colliding” valid Unified Address/Viewing Key that:
  1. has a string encoding matching that of one of the input Addresses/Viewing Keys on some subset of characters (for concreteness, consider the first \(n\) and last \(m\) characters, up to some bound on \(n+m\) );
  2. is controlled by the adversary (for concreteness, the adversary knows at least one of the private keys of the constituent Addresses).

Security goal (nonmalleability):

• In this variant, part b) above is replaced by the meaning of the new Address/Viewing Key being “usefully” different than the one it is based on, even though the adversary does not know any of the private keys. For example, if it were possible to delete a shielded constituent Address from a UA leaving only a Transparent Address, that would be a significant malleability attack.

Solution

We use an unkeyed 4-round Feistel construction to approximate a random permutation. (As explained below, 3 rounds would not be sufficient.)

Let \(H_i\) be a hash personalized by \(i,\) with maximum output length \(\ell_H\) bytes. Let \(G_i\) be a XOF (a hash function with extendable output length) based on \(H,\) personalized by \(i.\)

Define \(\ell^\mathsf{MAX}_M = (2^{16} + 1) \cdot \ell_H.\) For the instantiation using BLAKE2b defined below, \(\ell^\mathsf{MAX}_M = 4194368.\)

Given input \(M\) of length \(\ell_M\) bytes such that \(48 \leq \ell_M \leq \ell^\mathsf{MAX}_M,\) define \(\mathsf{F4Jumble}(M)\) by:

• let \(\ell_L = \mathsf{min}(\ell_H, \mathsf{floor}(\ell_M/2))\)
• let \(\ell_R = \ell_M - \ell_L\)
• split \(M\) into \(a\) of length \(\ell_L\) bytes and \(b\) of length \(\ell_R\) bytes
• let \(x = b \oplus G_0(a)\)
• let \(y = a \oplus H_0(x)\)
• let \(d = x \oplus G_1(y)\)
• let \(c = y \oplus H_1(d)\)
• return \(c \,||\, d.\)

The inverse function \(\mathsf{F4Jumble}^{-1}\) is obtained in the usual way for a Feistel construction, by observing that \(r = p \oplus q\) implies \(p = r \oplus q.\)

The first argument to BLAKE2b below is the personalization.

We instantiate \(H_i(u)\) by \(\mathsf{BLAKE2b‐}(8\ell_L)(\texttt{“UA_F4Jumble_H”} \,||\,\) \([i, 0, 0], u),\) with \(\ell_H = 64.\)

We instantiate \(G_i(u)\) as the first \(\ell_R\) bytes of the concatenation of \([\mathsf{BLAKE2b‐}512(\texttt{“UA_F4Jumble_G”} \,||\, [i] \,||\,\) \(\mathsf{I2LEOSP}_{16}(j), u) \text{ for } j \text{ from}\) \(0 \text{ up to } \mathsf{ceiling}(\ell_R/\ell_H)-1].\)

(In practice the lengths \(\ell_L\) and \(\ell_R\) will be roughly the same until \(\ell_M\) is larger than \(128\) bytes.)

Heuristic analysis

A 3-round unkeyed Feistel, as shown, is not sufficient:

Suppose that an adversary has a target input/output pair \((a \,||\, b, c \,||\, d),\) and that the input to \(H_0\) is \(x.\) By fixing \(x,\) we can obtain another pair \(((a \oplus t) \,||\, b', (c \oplus t) \,||\, d')\) such that \(a \oplus t\) is close to \(a\) and \(c \oplus t\) is close to \(c.\) ( \(b'\) and \(d'\) will not be close to \(b\) and \(d,\) but that isn't necessarily required for a valid attack.)

A 4-round Feistel thwarts this and similar attacks. Defining \(x\) and \(y\) as the intermediate values in the first diagram above:

• if \((x', y')\) are fixed to the same values as \((x, y),\) then \((a', b', c', d') = (a, b, c, d);\)
• if \(x' = x\) but \(y' \neq y,\) then the adversary is able to introduce a controlled \(\oplus\!\) -difference \(a \oplus a' = y \oplus y',\) but the other three pieces \((b, c, d)\) are all randomized, which is sufficient;
• if \(y' = y\) but \(x' \neq x,\) then the adversary is able to introduce a controlled \(\oplus\!\) -difference \(d \oplus d' = x \oplus x',\) but the other three pieces \((a, b, c)\) are all randomized, which is sufficient;
• if \(x' \neq x\) and \(y' \neq y,\) all four pieces are randomized.

Note that the size of each piece is at least 24 bytes.

It would be possible to make an attack more expensive by making the work done by a Producer more expensive. (This wouldn't necessarily have to increase the work done by the Consumer.) However, given that Unified Addresses may need to be produced on constrained computing platforms, this was not considered to be beneficial overall.

The padding contains the HRP so that the HRP has the same protection against malleation as the rest of the address. This may help against cross-network attacks, or attacks that confuse addresses with viewing keys.