Selecting an Oracle
Oracles are smart contracts that publish off-chain data on-chain. Blend pools use price oracles to get the prices of their assets. Pool creators must set a pool's oracle contract when they create a pool. This must be a single contract that can report prices for all assets in the pool.\
Oracles CANNOT be changed after a pool is created, so please be very careful when selecting an oracle for a pool.
Oracles must support the "lastprice" and "decimals" functions on the SEP-40 oracle standard.
Test the Oracle
It is required to verify that "lastprice" works for all potential reserves the pool will maintain. Please note the Blend pool will always invoke the oracle with Asset::Stellar({contract_address})
to fetch last price.
The oracle can be tested by using the PoolOracle
object provided via the Blend JS SDK.
Oracle Adapters
When a generalized oracle is insufficient for a pool (i.e. it doesn't support all assets the pool needs prices for), the pool creator may need to use an oracle adaptor. An oracle adaptor is a custom oracle contract that uses custom logic to aggregate multiple oracle feeds or impose desired behavior, such as reporting TWAP prices.
Sample Oracle Adapter: https://github.com/blend-capital/oracle-aggregator
Types of Price Feeds
Pool creators may find different types of price feeds useful. The two main types used in most lending pools are:
Spot: Spot price feeds report the current price of an asset.
TWAPs (time-weighted-asset-prices): These price feeds report average asset price over a given time period. These price feeds are more difficult to manipulate than spot price feeds; they are preferable for high-volatility or low-liquidity assets. TWAPs come in two types:
GM-TWAPS: geometric mean TWAPs calculate the geometric mean of an asset's price over a given time period. They are more resistant to manipulation than AM-TWAPS.
AM-TWAPS: arithmetic mean TWAPs calculate the arithmetic mean of an asset's price over a given time period.
Price Feed Aggregation
Ideally, price feeds should be aggregated across multiple sources to make them more manipulation-resistant. For example, a price feed reporting the XLM:USD spot price might aggregate the price by taking the average price from the Stellar DEX, Binance orderbook, and Coinbase orderbook.
Well Known Oracles
Last updated