This project assumes the following:
- Go 1.22
- Linux operating system
The Autonity Oracle Server offers exchange rate data for currency pairs from various data providers,
consolidating these data points into a standardized report. This information is then pushed to the
oracle contract deployed on the Autonity L1 network via a transaction. The oracle contract aggregates those supplied data points to supply a reference exchange rate data through a voting mechanism using a commit-reveal scheme.
DApps deployed on the Autonity L1 network can access these data points via the oracle contract interface.
- The Oracle Server is operated and maintained by an Autonity validator node operator.
- The validator must participate in the Oracle protocol if selected to be part of the consensus committee.
- A proof of ownership for the oracle account is required when submitting a registration transaction to the Autonity Contract.
- Oracle transactions are refunded if successful. The balance of the oracle account is required to be funded to cover at least one voting transaction.
The Autonity oracle client provides a unified interface between itself and plugins that adapt data from
different data providers, for example Binance and Coingecko, etc. Any party can build a plugin implementing
this unified interface and so provide an adaptor for any data source on demand. The oracle client will scan and load
plugins from the plugin directory during runtime. Detection of new or changed plugins is dynamic;
no shutdown of the oracle client is required to detect and apply the change.
To coordinate data sampling in the oracle network, the L1 oracle contract issues a round event on every vote period (30 ~ 60 blocks). The round event carries a tuple (RoundID, SampleTS, Height, VotePeriod), which tells the oracle servers that on round with ID RoundID, a data sample with timestamp SampleTS is required for the data submission. The Height stands for the start height of the new round, while the VotePeriod stands for the round length of the new round. Thus the oracle server can estimate and manage data pre-samplings for the new round and then pick up the nearest sample referring to the required SampleTS.
To mitigate data deviation caused by the distributed system environment, a data pre-sampling mechanism is employed parameterised by SampleTS and Height log data from the round event. When approaching the next round's start boundary Height, the oracle server initiates data pre-sampling approximately 15 seconds in advance. During this pre-sampling window, the server samples data per second and selects the sample closest to the required SampleTS for data aggregation. The oracle server will then submit that sample to the L1 oracle contract as its price vote for the next oracle voting round.
In a production network, node operators should obtain real-time data from high-quality data sources. However, most commercial data providers price their services based on quality of service (QoS) and rate limits. To address this, a configuration parameter "refresh" has been introduced for each data plugin. This parameter represents the interval in seconds between data fetches after the last successful data sampling. A buffered sample is used before the next data fetch. Node operators should configure an appropriate "refresh" interval by estimating the data fetching rate and the QoS subscribed from the data provider. The default value of "refresh" is 30 seconds, indicating that the plugin will query the data from the data source once every 30 seconds, even during the data pre-sampling window. If the data source does not limit the rate, it's recommended to set "refresh" to 1, allowing the pre-sampling to fetch data every 1 second to obtain real-time data. If the default "refresh" of 30 seconds is kept, then the oracle server will be sampling data up to 30 seconds old rather than in real-time.
The configuration file:
# Oracle Server Configuration
logLevel: 3 # Logging verbosity: 0: NoLevel, 1: Trace, 2: Debug, 3: Info, 4: Warn, 5: Error
gasTipCap: 1 # Set the gas priority fee cap to issue the oracle data report transactions.
#Set the buffering time window in blocks to continue vote after the last penalty event. Default value is 86400 (1 day).
#With such time buffer, the node operator can check and repair the local infra without being slashed due to the voting.
#This is important for node operator to prevent node from getting slashed again.
voteBuffer: 86400 # Buffer time in seconds (3600 * 24)
#Set oracle server key file.
keyFile: "./UTC--2023-02-27T09-10-19.592765887Z--b749d3d83376276ab4ddef2d9300fb5ce70ebafe"
#Set the password to decrypt oracle server key file.
keyPassword: "123%&%^$" # Password for the key file
#Set the WS-RPC server listening interface and port of the connected Autonity Client node.
autonityWSUrl: "ws://127.0.0.1:8546"
#Set the directory of the data plugins.
pluginDir: "./plugins" # Directory for plugins
#Set the profiling report directory, where some runtime state will be saved at.
profileDir: "." # Profile directory
#Set the confidence strategy, available strategies are: 0: linear, 1: fixed.
confidenceStrategy: 0 # 0: linear, 1: fixed
#Set the plugin configs.
# The forex data plugins are used to fetch realtime rate of currency pairs:
# EUR-USD, JPY-USD, GBP-USD, AUD-USD, CAD-USD and SEK-USD from commercial data providers. There are 4 implemented forex
# data plugins, each of them requires the end user to apply for their own service key from the selected data provider.
# The selection of which forex data plugin(s) to use is for the end user to decide. The user can use any one of them,
# or he/she can use multiple forex data plugins in the setup.
#
# The crypto data plugins are used to fetch market prices for the crypto currency pairs: ATN-USDC, NTN-USDC, NTN-ATN and
# USDC-USD. USDC liquidity is bridged to the Autonity public testnet from the Polygon Amoy testnet via a bridge service.
# Out-the-box plugins for collecting ATN-USDC and NTN-USDC market data are available for UniSwap V2 and AirSwap protocols. NTN-ATN market price is derived from
# that market data, and USDC pricing is converted to USD. ATN-NTN, ATN-USD, and NTN-USD prices are then submitted on-chain.
# To retrieve ATN and NTN prices, put the `crypto_uniswap` plugin and `crypto_airswap` plugin in your plugin directory.
# Oracle server can then discover and load them. Configuring the `crypto_uniswap` and `crypto_airswap` plugin does not
# require an API key, it is an open and free data source of a standard EVM RPC websocket service endpoint. The
# end user can connect to specific EVM RPC endpoint base on the blockchain which hosts the uniswap and airswap contracts.
# USDC-USD prices are required by the protocol to convert the ATN-USDC and NTN-USDC to ATN-USD and NTN-USD. This enables
# the reporting of ATN and NTN prices in USD to the ASM. Three plugins are implemented to source the USDC-USD datapoint
# from open and free data sources: coinbase, coingecko, and kraken. To prevent single data source failure, putting all
# 3 plugins of CEX into your plugin directory is recommended. Oracle server can then discover and load them.
# You don't need to configure the CEX plugins (crypto_coinbase, crypto_coingecko, crypto_kraken) in your oracle server
# plugin configuration file.
# For the forex data plugin default configuration is set, so the end user just needs to configure required settings,
# namely `name` and `key`. The configuration settings of a plugin are:
#
# // PluginConfig carry the configuration of plugins.
# type PluginConfig struct {
# Name string `json:"name" yaml:"name"` // the name of the plugin binary.
# Key string `json:"key" yaml:"key"` // the API key granted by your data provider to access their data API.
# Scheme string `json:"scheme" yaml:"scheme"` // the data service scheme, http or https.
# Endpoint string `json:"endpoint" yaml:"endpoint"` // the data service endpoint url of the data provider.
# Timeout int `json:"timeout" yaml:"timeout"` // the timeout period in seconds that an API request is lasting for.
# DataUpdateInterval int `json:"refresh" yaml:"refresh"` // the interval in seconds to fetch data from data provider due to rate limit.
# NTNTokenAddress string `json:"ntnTokenAddress" yaml:"ntnTokenAddress"` // The NTN erc20 token address on the target blockchain.
# ATNTokenAddress string `json:"atnTokenAddress" yaml:"atnTokenAddress"` // The Wrapped ATN erc20 token address on the target blockchain.
# USDCTokenAddress string `json:"usdcTokenAddress" yaml:"usdcTokenAddress"` // USDCx erc20 token address on the target blockchain.
# SwapAddress string `json:"swapAddress" yaml:"swapAddress"` // UniSwap factory contract address or AirSwap SwapERC20 contract address on the target blockchain.
# Disabled bool `json:"disabled" yaml:"disabled"` // The flag to disable/enable a plugin.
#}
# Un-comment below lines to enable your forex data plugin's configuration on demand. Your production configurations start from below:
#pluginConfigs:
# - name: forex_currencyfreaks # required, it is the plugin file name in the plugin directory.
# key: 175aab9e47e54790bf6d502c48407c10 # required, visit https://currencyfreaks.com to get your key, and replace it.
# refresh: 3600 # optional, buffered data within 3600s, recommended for API rate limited data source.
# - name: forex_openexchange # required, it is the plugin file name in the plugin directory.
# key: 1be02ca33c4843ee968c4cedd2686f01 # required, visit https://openexchangerates.org to get your key, and replace it.
# refresh: 3600 # optional, buffered data within 3600s, recommended for API rate limited data source.
# - name: forex_currencylayer # required, it is the plugin file name in the plugin directory.
# key: 105af082ac7f7d150c87303d4e2f049e # required, visit https://currencylayer.com to get your key, and replace it.
# refresh: 3600 # optional, buffered data within 3600s, recommended for API rate limited data source.
# - name: forex_exchangerate # required, it is the plugin file name in the plugin directory.
# key: 111f04e4775bb86c20296530 # required, visit https://www.exchangerate-api.com to get your key, and replace it.
# refresh: 3600 # optional, buffered data within 3600s, recommended for API rate limited data source.
# Un-comment below lines to config the RPC endpoint of a Piccadilly Network Full Node for your AMM plugin which sources ATN & NTN market data from an on-chain AMM.
# - name: crypto_uniswap
# scheme: "wss" # Available values are: "http", "https", "ws" or "wss", default value is "wss".
# endpoint: "rpc-internal-1.piccadilly.autonity.org/ws" # The default URL might not be stable for public usage, we recommend you to change it with your validator node's RPC endpoint.
#Enable the metric collection for oracle server, supported TS-DB engines are influxDB v1 and v2.
#metricConfigs:
# influxDBEndpoint: "http://localhost:8086"
# influxDBTags: "host=localhost"
# enableInfluxDB: false
# influxDBDatabase: "oracle"
# influxDBUsername: "test"
# influxDBPassword: "test"
# enableInfluxDBV2: false
# influxDBToken: "test"
# influxDBBucket: "oracle"
# influxDBOrganization: "oracle"Print the version of the oracle server:
$./autoracle version
v0.2.4
Run the server:
$./autoracle ./oracle_config.ymlDownload the Autonity client to generate the private key from console, and set the password to encode the key file, the
key file path will display, and remember the password that encrypted the key file.
$./autonity --datadir ./keys/ account new
Your new account is locked with a password. Please give a password. Do not forget this password.
Password:xxxxxx
Repeat password:xxxxxx
Your new key was generated
Public address of the key: 0x7C785Fe9404574AaC7daf2FF30637546493900d1
Path of the secret key file: key-data/keystore/UTC--2023-02-28T11-40-15.383709761Z--7c785fe9404574aac7daf2ff30637546493900d1
- You can share your public address with anyone. Others need it to interact with you.
- You must NEVER share the secret key with anyone! The key controls access to your funds!
- You must BACKUP your key file! Without the key, it's impossible to access account funds!
- You must REMEMBER your password! Without the password, it's impossible to decrypt the key!Prepare the plugin binaries, and save them into the plugins directory.
$./autoracle ./oracle_config.yml── profiles
└── 2024-11-19
├── cpu.profile_1
├── goroutines.txt_1
├── mem.profile_1
└── trace.out_1
To add a new data source, just put the new plugin into the service's plugins directory. The oracle service auto discovers and manages it. There are no other operations required from the operator.
To replace running plugins with new ones, just replace the binary in the plugins directory. The oracle service auto discovers it by checking the modification time of the binary and does the plugin replacement itself. There are no other operations required from the operator.
One can remove the plugin binary from the plugin directory to remove a plugin from the server during runtime, it will also stop and unload the plugin from the oracle server.
A disabled plugin will be unloaded from the oracle server, one can enable it again once get the plugin and its configuration ready, then the oracle server will load and start it.
cpuSysLoad = GetOrRegisterGauge("system/cpu/sysload", DefaultRegistry)
cpuSysWait = GetOrRegisterGauge("system/cpu/syswait", DefaultRegistry)
cpuProcLoad = GetOrRegisterGauge("system/cpu/procload", DefaultRegistry)
cpuThreads = GetOrRegisterGauge("system/cpu/threads", DefaultRegistry)
cpuGoroutines = GetOrRegisterGauge("system/cpu/goroutines", DefaultRegistry)
cpuSchedLatency = getOrRegisterRuntimeHistogram("system/cpu/schedlatency", secondsToNs, nil)
memPauses = getOrRegisterRuntimeHistogram("system/memory/pauses", secondsToNs, nil)
memAllocs = GetOrRegisterMeter("system/memory/allocs", DefaultRegistry)
memFrees = GetOrRegisterMeter("system/memory/frees", DefaultRegistry)
memTotal = GetOrRegisterGauge("system/memory/held", DefaultRegistry)
heapUsed = GetOrRegisterGauge("system/memory/used", DefaultRegistry)
heapObjects = GetOrRegisterGauge("system/memory/objects", DefaultRegistry)
diskReads = GetOrRegisterMeter("system/disk/readcount", DefaultRegistry)
diskReadBytes = GetOrRegisterMeter("system/disk/readdata", DefaultRegistry)
diskReadBytesCounter = GetOrRegisterCounter("system/disk/readbytes", DefaultRegistry)
diskWrites = GetOrRegisterMeter("system/disk/writecount", DefaultRegistry)
diskWriteBytes = GetOrRegisterMeter("system/disk/writedata", DefaultRegistry)
diskWriteBytesCounter = GetOrRegisterCounter("system/disk/writebytes", DefaultRegistry)oracle-server metrics:
numOfPlugins = metrics.GetOrRegisterGauge("oracle/plugins", nil)
oracleRound = metrics.GetOrRegisterGauge("oracle/round", nil)
slashEventCounter = metrics.GetOrRegisterCounter("oracle/slash", nil)
l1ConnectivityErrs = metrics.GetOrRegisterCounter("oracle/l1/errs", nil)
accountBalance = metrics.GetOrRegisterGauge("oracle/balance", nil)
isVoterFlag = metrics.GetOrRegisterGauge("oracle/isVoter", nil)plugin metrics:
All the data points collected from the plugin are tracked in metrics with such id pattern: oracle/plugin_name/symbol/price:
func (pw *PluginWrapper) updateMetrics(prices []types.Price) {
for _, p := range prices {
m, ok := pw.priceMetrics[p.Symbol]
if !ok {
name := strings.Join([]string{"oracle", pw.Name(), p.Symbol, "price"}, "/")
gauge := metrics.GetOrRegisterGaugeFloat64(name, nil)
gauge.Update(p.Price.InexactFloat64())
pw.priceMetrics[p.Symbol] = gauge
continue
}
m.Update(p.Price.InexactFloat64())
}
}make autoracle-bakerloomake autoracleTo build the data source simulator run
make simulatorTo run e2e test use
make e2e-testTo run all tests use
make testTo lint code run
make lintTo generate mocks for unit test
make mockTo build docker image
make build-docker-imageTo build a plugin, please refer to How to build a plugin
Built binaries are presented at: ./build/bin under which there is a plugins directory for the built plugins as well.