Home

Awesome

Zephyr

Copyright (c) 2023, Zephyr Protocol
Copyright (c) 2014-2023, The Monero Project
Portions Copyright (c) 2019-2021, Haven Protocol
Portions Copyright (c) 2012-2013 The Cryptonote developers.

Introduction

Zephyr is a revolutionary digital currency that combines the principles of privacy and stability. Grounded firmly on the academically peer-reviewed, formally verified and empirically battle-tested Minimal Djed protocol, Zephyr embodies the first over-collateralized, private stablecoin, seamlessly incorporating the privacy features of Monero into a stablecoin system. The research and development of the Djed protocol started as a collaboration between Ergo, Emurgo and IOHK in 2020, and has since had multiple implementations and deployments (cf. https://docs.djed.one).

In simple terms, Zephyr Protocol works as follows. Users can mint or redeem Zephyr Stable Dollar (ZSD) and Zephyr Reserve Share (ZRS) coins in exchange for the base coin (ZEPH). The price for stable coins is determined by a pricing oracle, the price for reserve coins is determined using a formula based on the current reserve data.

When minting new stable/reserve coins the base coins that are exchanged (+ fee) get added to the Zephyr reserve. This reserve is the collateral backing for the stablecoin holders and is ensured to be above 400% collateral at all times. In the event that collateralization falls below 400% due to price decrease, new stable coins are unable to be minted, however the reserve coin price will drop giving incentive for users to buy reserve coins at a reduced rate.

Reserve coin holders are entitled to redeem their reserve coins for the amount of equity in the reserve based on their proportion of reserve coin holdings. This means that as stablecoins are minted and redeemed, the accumulated fees will build up in the reserve as profit for holders of the reserve coin. There is a maximum reserve ratio of 800% to prevent dilution for the reserve coin holders.

Privacy: Zephyr uses a cryptographically sound system to allow you to send and receive funds without your transactions being easily revealed on the blockchain (the ledger of transactions that everyone has). This ensures that your purchases, receipts, and all transfers remain private by default.

Security: Using the power of a distributed peer-to-peer consensus network, every transaction on the network is cryptographically secured. Individual wallets have a 25-word mnemonic seed that is only displayed once and can be written down to backup the wallet. Wallet files should be encrypted with a strong passphrase to ensure they are useless if ever stolen.

Untraceability: By taking advantage of ring signatures, a special property of a certain type of cryptography, Zephyr is able to ensure that transactions are not only untraceable but have an optional measure of ambiguity that ensures that transactions cannot easily be tied back to an individual user or computer.

Decentralization: The utility of Zephyr depends on its decentralised peer-to-peer consensus network - anyone should be able to run the zephyr software, validate the integrity of the blockchain, and participate in all aspects of the zephyr network using consumer-grade commodity hardware. Decentralization of the zephyr network is maintained by software development that minimizes the costs of running the zephyr software and inhibits the proliferation of specialized, non-commodity hardware.

About this project

This is the core implementation of Zephyr. It is open source and completely free to use without restrictions, except for those specified in the license agreement below. There are no restrictions on anyone creating an alternative implementation of Zephyr that uses the protocol and network in a compatible manner.

As with many development projects, the repository on GitHub is considered to be the "staging" area for the latest changes. Before changes are merged into that branch on the main repository, they are tested by individual developers in their own branches, submitted as a pull request, and then subsequently tested by contributors who focus on testing and code reviews. That having been said, the repository should be carefully considered before using it in a production environment, unless there is a patch in the repository for a particular show-stopping issue you are experiencing. It is generally a better idea to use a tagged release for stability.

Anyone is welcome to contribute to Zephyr's codebase! If you have a fix or code change, feel free to submit it as a pull request directly to the "master" branch. In cases where the change is relatively small or does not affect other parts of the codebase, it may be merged in immediately by any one of the collaborators. On the other hand, if the change is particularly large or complex, it is expected that it will be discussed at length either well in advance of the pull request being submitted, or even directly on the pull request.

License

See LICENSE.

Contributing

If you want to help out, see CONTRIBUTING for a set of guidelines.

Compiling Zephyr from source

Dependencies

The following table summarizes the tools and libraries required to build. A few of the libraries are also included in this repository (marked as "Vendored"). By default, the build uses the library installed on the system and ignores the vendored sources. However, if no library is found installed on the system, then the vendored source will be built and used. The vendored sources are also used for statically-linked builds because distribution packages often include only shared library binaries (.so) but not static library archives (.a).

DepMin. versionVendoredDebian/Ubuntu pkgArch pkgVoid pkgFedora pkgOptionalPurpose
GCC7NObuild-essentialbase-develbase-develgccNO
CMake3.5NOcmakecmakecmakecmakeNO
pkg-configanyNOpkg-configbase-develbase-develpkgconfNO
Boost1.58NOlibboost-all-devboostboost-develboost-develNOC++ libraries
OpenSSLbasically anyNOlibssl-devopensslopenssl-developenssl-develNOsha256 sum
libzmq4.2.0NOlibzmq3-devzeromqzeromq-develzeromq-develNOZeroMQ library
OpenPGM?NOlibpgm-devlibpgmopenpgm-develNOFor ZeroMQ
libnorm[2]?NOlibnorm-devYESFor ZeroMQ
libunbound1.4.16NOlibunbound-devunboundunbound-develunbound-develNODNS resolver
libsodium?NOlibsodium-devlibsodiumlibsodium-devellibsodium-develNOcryptography
libunwindanyNOlibunwind8-devlibunwindlibunwind-devellibunwind-develYESStack traces
liblzmaanyNOliblzma-devxzliblzma-develxz-develYESFor libunwind
libreadline6.3.0NOlibreadline6-devreadlinereadline-develreadline-develYESInput editing
expat1.1NOlibexpat1-devexpatexpat-develexpat-develYESXML parsing
GTest1.5YESlibgtest-dev[1]gtestgtest-develgtest-develYESTest suite
ccacheanyNOccacheccacheccacheccacheYESCompil. cache
DoxygenanyNOdoxygendoxygendoxygendoxygenYESDocumentation
GraphvizanyNOgraphvizgraphvizgraphvizgraphvizYESDocumentation
lrelease?NOqttools5-dev-toolsqt5-toolsqt5-toolsqt5-linguistYESTranslations
libhidapi?NOlibhidapi-devhidapihidapi-develhidapi-develYESHardware wallet
libusb?NOlibusb-1.0-0-devlibusblibusb-devellibusbx-develYESHardware wallet
libprotobuf?NOlibprotobuf-devprotobufprotobuf-develprotobuf-develYESHardware wallet
protoc?NOprotobuf-compilerprotobufprotobufprotobuf-compilerYESHardware wallet
libudev?NOlibudev-devsystemdeudev-libudev-develsystemd-develYESHardware wallet

[1] On Debian/Ubuntu libgtest-dev only includes sources and headers. You must build the library binary manually. This can be done with the following command sudo apt-get install libgtest-dev && cd /usr/src/gtest && sudo cmake . && sudo make then:

[2] libnorm-dev is needed if your zmq library was built with libnorm, and not needed otherwise

Install all dependencies at once on Debian/Ubuntu:

sudo apt update && sudo apt install build-essential cmake pkg-config libssl-dev libzmq3-dev libunbound-dev libsodium-dev libunwind8-dev liblzma-dev libreadline6-dev libexpat1-dev libpgm-dev qttools5-dev-tools libhidapi-dev libusb-1.0-0-dev libprotobuf-dev protobuf-compiler libudev-dev libboost-chrono-dev libboost-date-time-dev libboost-filesystem-dev libboost-locale-dev libboost-program-options-dev libboost-regex-dev libboost-serialization-dev libboost-system-dev libboost-thread-dev python3 ccache doxygen graphviz

Install all dependencies at once on Arch:

sudo pacman -Syu --needed base-devel cmake boost openssl zeromq libpgm unbound libsodium libunwind xz readline expat gtest python3 ccache doxygen graphviz qt5-tools hidapi libusb protobuf systemd

Install all dependencies at once on Fedora:

sudo dnf install gcc gcc-c++ cmake pkgconf boost-devel openssl-devel zeromq-devel openpgm-devel unbound-devel libsodium-devel libunwind-devel xz-devel readline-devel expat-devel gtest-devel ccache doxygen graphviz qt5-linguist hidapi-devel libusbx-devel protobuf-devel protobuf-compiler systemd-devel

Install all dependencies at once on openSUSE:

sudo zypper ref && sudo zypper in cppzmq-devel libboost_chrono-devel libboost_date_time-devel libboost_filesystem-devel libboost_locale-devel libboost_program_options-devel libboost_regex-devel libboost_serialization-devel libboost_system-devel libboost_thread-devel libexpat-devel libminiupnpc-devel libsodium-devel libunwind-devel unbound-devel cmake doxygen ccache fdupes gcc-c++ libevent-devel libopenssl-devel pkgconf-pkg-config readline-devel xz-devel libqt5-qttools-devel patterns-devel-C-C++-devel_C_C++

Install all dependencies at once on macOS with the provided Brewfile:

brew update && brew bundle --file=contrib/brew/Brewfile

FreeBSD 12.1 one-liner required to build dependencies:

pkg install git gmake cmake pkgconf boost-libs libzmq4 libsodium unbound

Cloning the repository

Clone recursively to pull-in needed submodule(s):

git clone --recursive https://github.com/ZephyrProtocol/zephyr

If you already have a repo cloned, initialize and update:

cd zephyr && git submodule init && git submodule update

Note: If there are submodule differences between branches, you may need to use git submodule sync && git submodule update after changing branches to build successfully.

Build instructions

Zephyr uses the CMake build system and a top-level Makefile that invokes cmake commands as needed.

On Linux and macOS

Dependencies need to be built with -fPIC. Static libraries usually aren't, so you may have to build them yourself with -fPIC. Refer to their documentation for how to build them.

On the Raspberry Pi

Tested on a Raspberry Pi Zero with a clean install of minimal Raspbian Stretch (2017-09-07 or later) from https://www.raspberrypi.org/downloads/raspbian/. If you are using Raspian Jessie, please see note in the following section.

Note for Raspbian Jessie users:

If you are using the older Raspbian Jessie image, compiling Zephyr is a bit more complicated. The version of Boost available in the Debian Jessie repositories is too old to use with Zephyr, and thus you must compile a newer version yourself. The following explains the extra steps and has been tested on a Raspberry Pi 2 with a clean install of minimal Raspbian Jessie.

On Windows:

Binaries for Windows are built on Windows using the MinGW toolchain within MSYS2 environment. The MSYS2 environment emulates a POSIX system. The toolchain runs within the environment and cross-compiles binaries that can run outside of the environment as a regular Windows application.

Preparing the build environment

Cloning

Building

On FreeBSD:

The project can be built from scratch by following instructions for Linux above(but use gmake instead of make). If you are running Zephyr in a jail, you need to add sysvsem="new" to your jail configuration, otherwise lmdb will throw the error message: Failed to open lmdb environment: Function not implemented.

Zephyr is also available as a port or package as zephyr-cli.

On OpenBSD:

You will need to add a few packages to your system. pkg_add cmake gmake zeromq libiconv boost libunbound.

The doxygen and graphviz packages are optional and require the xbase set. Running the test suite also requires py3-requests package.

Build zephyr: gmake

Note: you may encounter the following error when compiling the latest version of Zephyr as a normal user:

LLVM ERROR: out of memory
c++: error: unable to execute command: Abort trap (core dumped)

Then you need to increase the data ulimit size to 2GB and try again: ulimit -d 2000000

On NetBSD:

Check that the dependencies are present: pkg_info -c libexecinfo boost-headers boost-libs protobuf readline libusb1 zeromq git-base pkgconf gmake cmake | more, and install any that are reported missing, using pkg_add or from your pkgsrc tree. Readline is optional but worth having.

Third-party dependencies are usually under /usr/pkg/, but if you have a custom setup, adjust the "/usr/pkg" (below) accordingly.

Clone the zephyr repository recursively and checkout the most recent release as described above. Then build zephyr: gmake BOOST_ROOT=/usr/pkg LDFLAGS="-Wl,-R/usr/pkg/lib" release. The resulting executables can be found in build/NetBSD/[Release version]/Release/bin/.

On Solaris:

The default Solaris linker can't be used, you have to install GNU ld, then run cmake manually with the path to your copy of GNU ld:

mkdir -p build/release
cd build/release
cmake -DCMAKE_LINKER=/path/to/ld -D CMAKE_BUILD_TYPE=Release ../..
cd ../..

Then you can run make as usual.

Building portable statically linked binaries

By default, in either dynamically or statically linked builds, binaries target the specific host processor on which the build happens and are not portable to other processors. Portable binaries can be built using the following targets:

Cross Compiling

You can also cross-compile static binaries on Linux for Windows and macOS with the depends system.

The required packages are the names for each toolchain on apt. Depending on your distro, they may have different names. The depends system has been tested on Ubuntu 18.04 and 20.04.

Using depends might also be easier to compile Zephyr on Windows than using MSYS. Activate Windows Subsystem for Linux (WSL) with a distro (for example Ubuntu), install the apt build-essentials and follow the depends steps as depicted above.

The produced binaries still link libc dynamically. If the binary is compiled on a current distribution, it might not run on an older distribution with an older installation of libc. Passing -DBACKCOMPAT=ON to cmake will make sure that the binary will run on systems having at least libc version 2.17.

Trezor hardware wallet support

If you have an issue with building Monero with Trezor support, you can disable it by setting USE_DEVICE_TREZOR=OFF, e.g.,

USE_DEVICE_TREZOR=OFF make release

For more information, please check out Trezor src/device_trezor/README.md.

Gitian builds

See contrib/gitian/README.md.

Running zephyrd

The build places the binary in bin/ sub-directory within the build directory from which cmake was invoked (repository root by default). To run in the foreground:

./bin/zephyrd

To list all available options, run ./bin/zephyrd --help. Options can be specified either on the command line or in a configuration file passed by the --config-file argument. To specify an option in the configuration file, add a line with the syntax argumentname=value, where argumentname is the name of the argument without the leading dashes, for example, log-level=1.

To run in background:

./bin/zephyrd --log-file zephyrd.log --detach

To run as a systemd service, copy zephyrd.service to /etc/systemd/system/ and zephyrd.conf to /etc/. The example service assumes that the user zephyr exists and its home is the data directory specified in the example config.

If you're on Mac, you may need to add the --max-concurrency 1 option to zephyr-wallet-cli, and possibly zephyrd, if you get crashes refreshing.

Internationalization

See README.i18n.md.

Using Tor

There is a new, still experimental, integration with Tor. The feature allows connecting over IPv4 and Tor simultaneously - IPv4 is used for relaying blocks and relaying transactions received by peers whereas Tor is used solely for relaying transactions received over local RPC. This provides privacy and better protection against surrounding node (sybil) attacks.

While Zephyr isn't made to integrate with Tor, it can be used wrapped with torsocks, by setting the following configuration parameters and environment variables:

Example command line to start zephyrd through Tor:

DNS_PUBLIC=tcp torsocks zephyrd --p2p-bind-ip 127.0.0.1 --no-igd

A helper script is in contrib/tor/zephyr-over-tor.sh. It assumes Tor is installed already, and runs Tor and Zephyr with the right configuration.

Using Tor on Tails

TAILS ships with a very restrictive set of firewall rules. Therefore, you need to add a rule to allow this connection too, in addition to telling torsocks to allow inbound connections. Full example:

sudo iptables -I OUTPUT 2 -p tcp -d 127.0.0.1 -m tcp --dport 18081 -j ACCEPT
DNS_PUBLIC=tcp torsocks ./zephyrd --p2p-bind-ip 127.0.0.1 --no-igd --rpc-bind-ip 127.0.0.1 \
    --data-dir /home/amnesia/Persistent/your/directory/to/the/blockchain

Pruning

One can store a pruned blockchain in order to reduce needed hard disk space. A pruned blockchain can only serve part of the historical chain data to other peers, but is otherwise identical in functionality to the full blockchain. To use a pruned blockchain, it is best to start the initial sync with --prune-blockchain. However, it is also possible to prune an existing blockchain using the zephyr-blockchain-prune tool or using the --prune-blockchain zephyrd option with an existing chain. If an existing chain exists, pruning will temporarily require disk space to store both the full and pruned blockchains.

For more detailed information see the 'Pruning' entry in the Moneropedia

Debugging

This section contains general instructions for debugging failed installs or problems encountered with Zephyr. First, ensure you are running the latest version built from the GitHub repo.

Obtaining stack traces and core dumps on Unix systems

We generally use the tool gdb (GNU debugger) to provide stack trace functionality, and ulimit to provide core dumps in builds which crash or segfault.

Run the build.

Once it stalls, enter the following command:

gdb /path/to/zephyrd `pidof zephyrd`

Type thread apply all bt within gdb in order to obtain the stack trace

Enter ulimit -c unlimited on the command line to enable unlimited filesizes for core dumps

Enter echo core | sudo tee /proc/sys/kernel/core_pattern to stop cores from being hijacked by other tools

Run the build.

When it terminates with an output along the lines of "Segmentation fault (core dumped)", there should be a core dump file in the same directory as zephyrd. It may be named just core, or core.xxxx with numbers appended.

You can now analyse this core dump with gdb as follows:

gdb /path/to/zephyrd /path/to/dumpfile`

Print the stack trace with bt

coredumpctl -1 gdb

To run Zephyr within gdb:

Type gdb /path/to/zephyrd

Pass command-line options with --args followed by the relevant arguments

Type run to run zephyrd

Analysing memory corruption

There are two tools available:

ASAN

Configure Zephyr with the -D SANITIZE=ON cmake flag, eg:

cd build/debug && cmake -D SANITIZE=ON -D CMAKE_BUILD_TYPE=Debug ../..

You can then run the zephyr tools normally. Performance will typically halve.

valgrind

Install valgrind and run as valgrind /path/to/zephyrd. It will be very slow.

LMDB

Instructions for debugging suspected blockchain corruption as per @HYC

There is an mdb_stat command in the LMDB source that can print statistics about the database but it's not routinely built. This can be built with the following command:

cd ~/zephyr/external/db_drivers/liblmdb && make

The output of mdb_stat -ea <path to blockchain dir> will indicate inconsistencies in the blocks, block_heights and block_info table.

The output of mdb_dump -s blocks <path to blockchain dir> and mdb_dump -s block_info <path to blockchain dir> is useful for indicating whether blocks and block_info contain the same keys.

These records are dumped as hex data, where the first line is the key and the second line is the data.

Known Issues

Protocols

Socket-based

Because of the nature of the socket-based protocols that drive zephyr, certain protocol weaknesses are somewhat unavoidable at this time. While these weaknesses can theoretically be fully mitigated, the effort required (the means) may not justify the ends. As such, please consider taking the following precautions if you are a zephyr node operator:

Blockchain-based

Certain blockchain "features" can be considered "bugs" if misused correctly. Consequently, please consider the following: