bsips/bsip-0008.md
2015-12-28 10:07:51 +01:00

240 lines
12 KiB
Markdown

BSIP: 0008
Title: Privacy (STEALTH) Mode
Authors: Daniel Larimer <Dan@cryptonomex.com>
Fabian Schuh <Fabian@BitShares.eu>
Status: Draft
Type: Protocol
Created: 2015-12-16
Discussion: <https://github.com/cryptonomex/graphene/issues/452>
<https://bitsharestalk.org/index.php/topic,20104.0.html>
<https://bitsharestalk.org/index.php/topic,20499.0.html>
Worker: 1.14.18
# Abstract
Privacy Mode Transfers (a.k.a. Stealth Transfers) are used to maintain user
privacy. This feature helps set BitShares apart from most other crypto
currencies and offers tremendous value to the users who are most interested in
privacy, liberty, and freedom.
In practise they combine the techniques of *blinding signatures* to hide the
amount of a transfer and *stealth addresses* (similar to TITAN in BitShares 1)
to hide involved parties.
# Specifications
This proposal involves creating a new front-end feature on the account page in
the web wallet to allow users to enter the *privacy mode*. Here, users will be
able to create *private accounts* which are labeled private keys. They will
also be able to manage *private contacts* which are labeled public keys. Neither
private accounts nor private contacts are tracked on the blockchain since those
keys are not used directly. Instead, each transaction will derive a
transaction-specific address from the private contact's public key. A off-chain
memo helps the receiver identify the deposit and derive the corresponding
transaction-specific private key from the private account.
Users will be able to monitor their private balances and take the following
actions:
* Transfer from public account to their own private balance
* Transfer from one of their private accounts to one of their private contacts
* Transfer from one of their private accounts to any public account
* Register a new account using a private balance.
* Receive a private transfer from a 3rd party given a transfer receipt.
These features are already available on the blockchain level using *address
authorities* and *key authorities* and indicating that a transaction to that
account has to be made in private.
## GUI Mockup
The goal of this proposal is to have a front end developed (or integrated into an
existing web wallet) that allows ordinary people to use the privacy mode without
further knowledge about the underlying technology. Thus, a proposal for the user
interface has been drafted and is attached to this proposal as
[pdf](bsip-0008/gui-mockup.pdf).
[slide 1](bsip-0008/gui-mockup-0.png) [slide 2](bsip-0008/gui-mockup-1.png) [slide 3](bsip-0008/gui-mockup-2.png)
[slide 4](bsip-0008/gui-mockup-3.png) [slide 5](bsip-0008/gui-mockup-4.png) [slide 6](bsip-0008/gui-mockup-5.png)
[slide 7](bsip-0008/gui-mockup-6.png) [slide 8](bsip-0008/gui-mockup-7.png) [slide 9](bsip-0008/gui-mockup-8.png)
This draft has been designed with user experience in mind but may be improved if
it serves its purpose.
# Implementation Aspects
## Backup Considerations
Because private transfers are not recoverable from blockchain data alone,
backups of your wallet after receiving a new private transfer are *required*.
## Javascript Implementation
This proposal will require the use of a JavaScript library to perform the
necessary crypto operations in the web wallet (see [1]).
# Funding
BitSharestalk.org forum user `onceuponatime` has proposed to fund the
development and implementation of this feature in full as a private investor and
at zero cost to BitShares shareholders.
## Contract between `onceuponatime` and Cryptonomex:
The purpose of this contract is to develop a Privacy Mode feature, Privacy Mode
fee accumulation account, Maintenance Account, Initialization Package, and GUI
interface for BitShares scoped for a firm fixed price of $45K. The following
requirements apply:
1. The Privacy Mode feature shall be implemented as proposed in
https://github.com/cryptonomex/graphene/issues/452 (as amended).
2. It shall provide the following fee based services:
* Transfer from public account to their own private balance
* Transfer from one of their private accounts to one of their private contacts
* Transfer from one of their private accounts to any public account
* Register a new account using a private balance.
* Receive a private transfer from a 3rd party given a transfer receipt.
3. Each of these services shall charge a fee initially set at 3x the standard
transfer fee, but which may be adjusted from time to time by the owner(s) of
the Privacy Mode's management account (see [BSIP-0008](bsip-0008.md))
4. Fees shall be automatically distributed by the blockchain to the following
accounts:
* 20% to the BitShares network.
* 20% to a Maintenance Account.
* 60% to holder(s) of the Privacy Mode Fees accumulation account
5. The Maintenance Account shall be controlled by five specified manager
accounts in a 3 of 5 multisig configuration. These managers will control
the allocation of this fund to future maintenance and upgrade tasks.
6. The Initialization Package shall modify the blockchain to make the Privacy
Mode feature available to users.
7. The Initialization Package shall make provision for the creation of
generic Fee Based Assets (FBA) and set the fee for such
8. A GUI shall be provided in the Web and Light wallets to allow
ordinary users to easily use the Privacy Mode features.
9. Documentation of the Privacy Mode feature and Maintenance and Fee
Accumulation account shall be provided on the appropriate reference web
sites.
10. Resulting software patch to the Graphene library shall have the same
license as the rest of Graphene subject to the condition that the results
of the Initialization package and fee distribution mechanisms are not
modified.
# Management Account
The STEALTH asset will be issued by the "management account" (see
[BSIP-0007](bsip-0007.md)) for this feature (as created by the initialization
package). `Onceuponatime` will be the initial owner of the issued asset (not the
issuer). This management account will have 3-of-5 multi-signature authority
assigned to the 5 *largest* STEALTH holders weighted proportional to stake and
will have the power to set the fee.
# Roadmap
* Feedback and discussion of this thread: *December 8 to December 10, 2015*
* Presentation of an amended Cryptonomex Worker Proposal: *Dec 11, 2015*
This worker proposal should include Milestones of what is intended to be
accomplished by the end of week 1, week 2, week 3, week 4 and week 5 so that
the Community can follow progress in the github.
* Voting for Worker Proposal: *Dec 11 to January 1, 2016*
* onceuponatime forwards $45,000 to Cryptonomex: *Jan.2, 2016*
* Cryptonomex does the development and testing of the feature: *(4 to 6 weeks)*
* Hard fork for implementation of the feature: *Monday Feb, 15th*
# Worker Proposal
The worker proposal `1.14.18` is dedicated to poll shareholders about their
willingness to implement the Privacy Mode feature in a hard fork. It serves the
private investor as an indicator whether shareholders would appreciate his
investment and agree to upgrade the protocol accordingly.
Since the implementation is funded by a private investor, the worker only pays
approx. $300.
# Discussion
For the best user experience this proposal is best combined with proposal for
*hosted wallets* (to be defined).
## No Participation in Referral Program
The referral program does not play with Privacy mode transfers (because it is
private, we don't know who the parties are or who referred them). This means
that even if the fee were the same as the basic transfer, on average the
lifetime member would be paying 5x more a Private Transfer than a Public
Transfer. If you charge 3x the basic transfer fee, then life time members will
pay 15x more for a Private transfer than a Public transfer.
Percentage based fees are not possible with Private Transfers either because the
amount being transferred is *private*!
## Worker as Poll
When the project is complete, the new software will be presented to the
witnesses as an opportunity to hard fork by switching to running the new
package/software. At that time witnesses will face a choice:
1. Honor the proposal at the time the decision to spend the private investor's money was made, or
2. Honor the proposal at the time the product is delivered (if it has changed in the mean time).
Acting against the wishes of Vote 2 could get them fired if enough people
disagree with them keeping the implied commitment of Vote 1.
Acting against the wishes of Vote 1 means that BitShares gets a reputation for
reneging on a prior vote upon which an entrepreneur has relied.
This means that even if some voters have changed their minds about this proposal
by installation time, they might *not* decide to fire otherwise faithful
witnesses for deciding to obey Vote 1 in order to preserve BitShares' reputation
in future deals.
Considering the participation rate in the worker poll: those who do not care
enough to vote are ignored in the decision making process as it is supposed to
be.
## Privacy Mode not Most Pressingly Needed
The argument was raised that the Privacy Mode as proposed here is not the
feature most pressingly needed for BitShares right now.
## Confidential Transfers already available in the CLI-Wallet
The privacy mode is already documented and available for use in the console-line
wallet client at almost the same fee as non-private transfers (December 2015).
However, since these transfers cannot participate in the Referral Program (see
above), all fees are paid to the BitShares network. This proposal, shall it be
implemented, splits the fee for private transactions 20%/80% (similar to regular
transfers with the referral program) and hands out 20% of the fee to the
BitShares network. The other 80% are given to the maintenance account and
holders of the FBA. As a result, for BitShares shareholders to keep their 20%
transaction fee, the other 80% need to be paid by the customer. However, the fee
for this type of transaction is only increased to 3x and not the required 5x.
The assumptions are that the increase in volume for this particular transaction
type and the additional customers compensate for this loss.
# Summary for Shareholders
The idea of this proposal is to increase the use of BitShares by adding an
option for improved privacy to the BitShares customers. This features comes at
almost no cost to the shareholders (see below) since it is funded by a private
investor. Customers using this feature will initially need to pay 3x the basic
transfer fee to the improved privacy. The revenue from this is split as follows:
20% go to the BitShares shareholders as burned BTS, 20% go to a maintenance
account that will fund future development and updates for this feature and is
controlled by several individuals. The other 80% go to holders of a special
asset that will be created during a protocol upgrade and issued to the private
investor to compensate for his investment.
A hidden cost for BitShares shareholders is that in order to keep 20% of the
transaction fees, the fee has to be 5x the basic transfer fee but is only
increased by 3x. Even though this fee is a variable defined by the committee and
thus can be changed, the authors of this BSIP see a good compromise between
attracting new users and generating a higher revenue stream by relatively low
fees.
# Copyright
This document is placed in the public domain.
# See Also
* [1] https://github.com/arhag/crypto-experiments/tree/emscripten/emscripten/libsecp256k1-demo