Solidity: calldata, memory, and storage

Solidity: calldata, memory, and storage

Variables

There are three types of variables in Solidity:

  • Global variables
    • Provide information about the blockchain
    • For example, block.number, block.timestamp, or msg.sender
  • State variables
    • Declared outside a function
    • Stored on the blockchain
    • Also called "storage"
  • Local variables
    • Declared inside a function
    • Not stored on the blockchain, stored in memory instead
    • Erased between function calls

References:
https://docs.soliditylang.org/en/v0.7.6/units-and-global-variables.html
https://solidity-by-example.org/variables

Data Types

There are two types of data:

  • Value types: YourContract, address, bool, uint256, int256, enum, and bytes32 (fixed-size byte arrays)
  • Reference types: array, mapping, and struct

It's worth noting that bytes and string are dynamically-sized byte arrays and are considered reference types. However, byte (bytes1), bytes2, ..., bytes32 are value types since they're fixed-size byte arrays.

References:
https://docs.soliditylang.org/en/v0.7.6/types.html#value-types
https://docs.soliditylang.org/en/v0.7.6/types.html#reference-types

Data Location

When using a reference type, you must explicitly provide the data location where the type is stored. There are three data locations:

  • storage is where the state variables are stored, and its lifetime is the lifetime of the contract
  • memory means variables are temporary and erased between external function calls
  • calldata behaves mostly like memory, but is immutable

For reference types in function arguments, you must declare them as memory or calldata.

If possible, use calldata as the data location because it avoids copying, reduces gas usage, and ensures that the data cannot be modified. Arrays and structs with calldata data location can also be returned from functions, but it is not possible to allocate such types.

References:
https://docs.soliditylang.org/en/v0.7.6/types.html#data-location
https://medium.com/coinmonks/solidity-storage-vs-memory-vs-calldata-8c7e8c38bce
https://gist.github.com/hrkrshnn/ee8fabd532058307229d65dcd5836ddc

It is also a best practice to use external if you expect that the function will only ever be called externally, and use public if you need to call the function internally. The difference between both is that public function arguments are copied to memory, while in external functions, arguments are read directly from calldata, which is cheaper than memory allocation. external functions are sometimes more efficient when they receive large arrays.

References:
https://medium.com/newcryptoblock/best-practices-in-solidity-b324b65d33b1

Assignment Behaviour

  • Assignments between storage and memory (or from calldata) always create an independent copy.
  • Assignments from memory to memory only create references.
  • Assignments from storage to a local storage variable also only assign a reference.
  • All other assignments to storage always copy.

References:
https://docs.soliditylang.org/en/v0.7.6/types.html#data-location-and-assignment-behaviour

hardhat-deploy: Upgradeable Contracts with Linked Libraries

hardhat-deploy: Upgradeable Contracts with Linked Libraries

Library

Assume ContractA imports LibraryA, when deploying ContractA, LibraryA is embedded into ContractA if LibraryA contains only internal functions.

If LibraryA contains at least one external function, LibraryA must be deployed first, and then linked when deploying ContractA.

ref:
https://solidity-by-example.org/library/
https://docs.soliditylang.org/en/v0.7.6/using-the-compiler.html

Foundry

In Foundry tests, Foundry will automatically deploy libraries if they have external functions, so you don't need to explicitly link them.

hardhat-deploy

Whenever the library is changed, hardhat-deploy will deploy a new implementation and upgrade the proxy:

import { DeployFunction } from "hardhat-deploy/dist/types"
import { QuoteVault } from "../typechain-types"

const func: DeployFunction = async function (hre) {
    const { deployments, ethers } = hre

    // deploy library
    await deployments.deploy("PerpFacade", {
        from: deployerAddress,
        contract: "contracts/lib/PerpFacade.sol:PerpFacade",
    })
    const perpFacadeDeployment = await deployments.get("PerpFacade")

    // deploy upgradeable contract
    await deployments.deploy("QuoteVault", {
        from: deployerAddress,
        contract: "contracts/QuoteVault.sol:QuoteVault",
        proxy: {
            owner: multisigOwnerAddress, // ProxyAdmin.owner
            proxyContract: "OpenZeppelinTransparentProxy",
            viaAdminContract: "DefaultProxyAdmin",
            execute: {
                init: {
                    methodName: "initialize",
                    args: [
                        "Kantaban USDC-ETH QuoteVault",
                        "kUSDC-ETH",
                        usdcDecimals,
                        USDC,
                        WETH,
                    ],
                },
            },
        },
        libraries: {
            PerpFacade: perpFacadeDeployment.address,
        },
    })
    const quoteVaultDeployment = await deployments.get("QuoteVault")

    // must specify library address when instantiating the contract:
    const quoteVaultFactory = await ethers.getContractFactory("contracts/QuoteVault.sol:QuoteVault", {
        libraries: {
            PerpFacade: perpFacadeDeployment.address,
        },
    })
    const quoteVault = quoteVaultFactory.attach(quoteVaultDeployment.address) as unknown as QuoteVault
    console.log(await quoteVault.decimals())
}

export default func

ref:
https://github.com/wighawag/hardhat-deploy#handling-contract-using-libraries

Deploy Ethereum RPC Provider Load Balancer with HAProxy in Kubernetes (AWS EKS)

Deploy Ethereum RPC Provider Load Balancer with HAProxy in Kubernetes (AWS EKS)

To achieve high availability and better performance, we could build a HAProxy load balancer in front of multiple Ethereum RPC providers, and also automatically adjust traffic weights based on the latency and block timestamp of each RPC endpoints.

ref:
https://www.haproxy.org/

Configurations

In haproxy.cfg, we have a backend named rpc-backend, and two RPC endpoints: quicknode and alchemy as upstream servers.

global
    log stdout format raw local0 info
    stats socket ipv4@*:9999 level admin expose-fd listeners
    stats timeout 5s

defaults
    log global
    mode http
    option httplog
    option dontlognull
    timeout connect 10s
    timeout client 60s
    timeout server 60s
    timeout http-request 60s

frontend stats
    bind *:8404
    stats enable
    stats uri /
    stats refresh 10s

frontend http
    bind *:8000
    option forwardfor
    default_backend rpc-backend

backend rpc-backend
    balance leastconn
    server quicknode 127.0.0.1:8001 weight 100
    server alchemy 127.0.0.1:8002 weight 100

frontend quicknode-frontend
    bind *:8001
    option dontlog-normal
    default_backend quicknode-backend

backend quicknode-backend
    balance roundrobin
    http-request set-header Host xxx.quiknode.pro
    http-request set-path /xxx
    server quicknode xxx.quiknode.pro:443 sni str(xxx.quiknode.pro) check-ssl ssl verify none

frontend alchemy-frontend
    bind *:8002
    option dontlog-normal
    default_backend alchemy-backend

backend alchemy-backend
    balance roundrobin
    http-request set-header Host xxx.alchemy.com
    http-request set-path /xxx
    server alchemy xxx.alchemy.com:443 sni str(xxx.alchemy.com) check-ssl ssl verify none

ref:
https://docs.haproxy.org/2.7/configuration.html
https://www.haproxy.com/documentation/hapee/latest/configuration/

Test it on local:

docker run --rm -v $PWD:/usr/local/etc/haproxy \
-p 8000:8000 \
-p 8404:8404 \
-p 9999:9999 \
-i -t --name haproxy haproxy:2.7.0

docker exec -i -t -u 0 haproxy bash

echo "show stat" | socat stdio TCP:127.0.0.1:9999
echo "set weight rpc-backend/quicknode 0" | socat stdio TCP:127.0.0.1:9999

# if you're using a socket file descriptor
apt update
apt install socat -y
echo "set weight rpc-backend/alchemy 0" | socat stdio /var/lib/haproxy/haproxy.sock

ref:
https://www.redhat.com/sysadmin/getting-started-socat

Healtchcheck

Then the important part: we're going to run a simple but flexible healthcheck script, called node weighter, as a sidecar container. So the healthcheck script can access HAProxy admin socket of the HAProxy container through 127.0.0.1:9999.

The node weighter can be written in any language. Here is a TypeScript example:

in HAProxyConnector.ts which sets weights through HAProxy admin socket:

import net from "net"

export interface ServerWeight {
    backendName: string
    serverName: string
    weight: number
}

export class HAProxyConnector {
    constructor(readonly adminHost = "127.0.0.1", readonly adminPort = 9999) {}

    setWeights(serverWeights: ServerWeight[]) {
        const scaledServerWeights = this.scaleWeights(serverWeights)

        const commands = scaledServerWeights.map(server => {
            return `set weight ${server.backendName}/${server.serverName} ${server.weight}\n`
        })

        const client = net.createConnection({ host: this.adminHost, port: this.adminPort }, () => {
            console.log("HAProxyAdminSocketConnected")
        })
        client.on("error", err => {
            console.log("HAProxyAdminSocketError")
        })
        client.on("data", data => {
            console.log("HAProxyAdminSocketData")
            console.log(data.toString().trim())
        })

        client.write(commands.join(""))
    }

    private scaleWeights(serverWeights: ServerWeight[]) {
        const totalWeight = sum(serverWeights.map(server => server.weight))

        return serverWeights.map(server => {
            server.weight = Math.floor((server.weight / totalWeight) * 256)
            return server
        })
    }
}

in RPCProxyWeighter.ts which calculates weights based a custom healthcheck logic:

import { HAProxyConnector } from "./connectors/HAProxyConnector"
import config from "./config.json"

export interface Server {
    backendName: string
    serverName: string
    serverUrl: string
}

export interface ServerWithWeight {
    backendName: string
    serverName: string
    weight: number
    [metadata: string]: any
}

export class RPCProxyWeighter {
    protected readonly log = Log.getLogger(RPCProxyWeighter.name)
    protected readonly connector: HAProxyConnector

    protected readonly ADJUST_INTERVAL_SEC = 60 // 60 seconds
    protected readonly MAX_BLOCK_TIMESTAMP_DELAY_MSEC = 150 * 1000 // 150 seconds
    protected readonly MAX_LATENCY_MSEC = 3 * 1000 // 3 seconds
    protected shouldScale = false
    protected totalWeight = 0

    constructor() {
        this.connector = new HAProxyConnector(config.admin.host, config.admin.port)
    }

    async start() {
        while (true) {
            let serverWithWeights = await this.calculateWeights(config.servers)
            if (this.shouldScale) {
                serverWithWeights = this.connector.scaleWeights(serverWithWeights)
            }
            this.connector.setWeights(serverWithWeights)

            await sleep(1000 * this.ADJUST_INTERVAL_SEC)
        }
    }

    async calculateWeights(servers: Server[]) {
        this.totalWeight = 0

        const serverWithWeights = await Promise.all(
            servers.map(async server => {
                try {
                    return await this.calculateWeight(server)
                } catch (err: any) {
                    return {
                        backendName: server.backendName,
                        serverName: server.serverName,
                        weight: 0,
                    }
                }
            }),
        )

        // if all endpoints are unhealthy, overwrite weights to 100
        if (this.totalWeight === 0) {
            for (const server of serverWithWeights) {
                server.weight = 100
            }
        }

        return serverWithWeights
    }

    async calculateWeight(server: Server) {
        const healthInfo = await this.getHealthInfo(server.serverUrl)

        const serverWithWeight: ServerWithWeight = {
            ...{
                backendName: server.backendName,
                serverName: server.serverName,
                weight: 0,
            },
            ...healthInfo,
        }

        if (healthInfo.isBlockTooOld || healthInfo.isLatencyTooHigh) {
            return serverWithWeight
        }

        // normalizedLatency: the lower the better
        // blockTimestampDelayMsec: the lower the better
        // both units are milliseconds at the same scale
        // serverWithWeight.weight = 1 / healthInfo.normalizedLatency + 1 / healthInfo.blockTimestampDelayMsec

        // NOTE: if we're using `balance source` in HAProxy, the weight can only be 100% or 0%,
        // therefore, as long as the RPC endpoint is healthy, we always set the same weight
        serverWithWeight.weight = 100

        this.totalWeight += serverWithWeight.weight

        return serverWithWeight
    }

    protected async getHealthInfo(serverUrl: string): Promise<HealthInfo> {
        const provider = new ethers.providers.StaticJsonRpcProvider(serverUrl)

        // TODO: add timeout
        const start = Date.now()
        const blockNumber = await provider.getBlockNumber()
        const end = Date.now()

        const block = await provider.getBlock(blockNumber)

        const blockTimestamp = block.timestamp
        const blockTimestampDelayMsec = Math.floor(Date.now() / 1000 - blockTimestamp) * 1000
        const isBlockTooOld = blockTimestampDelayMsec >= this.MAX_BLOCK_TIMESTAMP_DELAY_MSEC

        const latency = end - start
        const normalizedLatency = this.normalizeLatency(latency)
        const isLatencyTooHigh = latency >= this.MAX_LATENCY_MSEC

        return {
            blockNumber,
            blockTimestamp,
            blockTimestampDelayMsec,
            isBlockTooOld,
            latency,
            normalizedLatency,
            isLatencyTooHigh,
        }
    }

    protected normalizeLatency(latency: number) {
        if (latency <= 40) {
            return 1
        }

        const digits = Math.floor(latency).toString().length
        const base = Math.pow(10, digits - 1)
        return Math.floor(latency / base) * base
    }
}

in config.json:

Technically, we don't need this config file. Instead, we could read the actual URLs from HAProxy admin socket directly. Though creating a JSON file that contains URLs is much simpler.

{
    "admin": {
        "host": "127.0.0.1",
        "port": 9999
    },
    "servers": [
        {
            "backendName": "rpc-backend",
            "serverName": "quicknode",
            "serverUrl": "https://xxx.quiknode.pro/xxx"
        },
        {
            "backendName": "rpc-backend",
            "serverName": "alchemy",
            "serverUrl": "https://xxx.alchemy.com/xxx"
        }
    ]
}

ref:
https://www.haproxy.com/documentation/hapee/latest/api/runtime-api/set-weight/
https://sleeplessbeastie.eu/2020/01/29/how-to-use-haproxy-stats-socket/

Deployments

apiVersion: v1
kind: ConfigMap
metadata:
  name: rpc-proxy-config-file
data:
  haproxy.cfg: |
    ...
  config.json: |
    ...
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: rpc-proxy
spec:
  replicas: 2
  selector:
    matchLabels:
      app: rpc-proxy
  template:
    metadata:
      labels:
        app: rpc-proxy
    spec:
      volumes:
        - name: rpc-proxy-config-file
          configMap:
            name: rpc-proxy-config-file
      containers:
        - name: haproxy
          image: haproxy:2.7.0
          ports:
            - containerPort: 8000
              protocol: TCP
          resources:
            requests:
              cpu: 200m
              memory: 256Mi
            limits:
              cpu: 1000m
              memory: 256Mi
          volumeMounts:
            - name: rpc-proxy-config-file
              subPath: haproxy.cfg
              mountPath: /usr/local/etc/haproxy/haproxy.cfg
              readOnly: true
        - name: node-weighter
          image: your-node-weighter
          command: ["node", "./index.js"]
          resources:
            requests:
              cpu: 200m
              memory: 256Mi
            limits:
              cpu: 1000m
              memory: 256Mi
          volumeMounts:
            - name: rpc-proxy-config-file
              subPath: config.json
              mountPath: /path/to/build/config.json
              readOnly: true
---
apiVersion: v1
kind: Service
metadata:
  name: rpc-proxy
spec:
  clusterIP: None
  selector:
    app: rpc-proxy
  ports:
    - name: http
      port: 8000
      targetPort: 8000

The RPC load balancer can then be accessed through http://rpc-proxy.default.svc.cluster.local:8000 inside the Kubernetes cluster.

ref:
https://www.containiq.com/post/kubernetes-sidecar-container
https://hub.docker.com/_/haproxy

The Graph: Subgraph and GRT Token

The Graph: Subgraph and GRT Token

The Graph is a protocol for indexing and querying Blockchain data. Currently, The Graph has the legacy version and the decentralized version. The legacy version is a centralized and managed service hosted by The Graph, and it will eventually be shutdown in the future. The decentralized version (aka The Graph Network) consists of 4 major roles: Developers, Indexers, Curators, and Delegators.

ref:
https://thegraph.com/

Video from Finematics
https://finematics.com/the-graph-explained/

Terminologies

Define a Subgraph

A subgraph defines one or more entities of indexed data: what kind of data on the blockchain you want to index for faster querying. Once deployed, subgraphs could be queried by dApps to fetch blockchain data to power their frontend interfaces. Basically, a subgraph is like a database, and an entity (a GraphQL type) is like a table in RDBMS.

A subgraph definition consists of 3 files:

  • subgraph.yaml: a YAML file that defines the subgraph manifest and metadata.
  • schema.graphql: a GraphQL schema that defines what data (entities) are stored, and how to query it via GraphQL.
  • mappings.ts: AssemblyScript code that translates blockchain data (events, blocks, or contract calls) to GraphQL entities.

GraphQL Schema

First, we need to design our GraphQL entity schemas (the data model) which mainly depends on how you want to query the data instead of how the data are emitted from the blockchain. GraphQL schemas are defined using the GraphQL schema language.

Here're some notes about subgraph's GraphQL schema:

An example of schema.graphql:

type Market @entity {
  id: ID!
  baseToken: Bytes!
  pool: Bytes!
  feeRatio: BigInt!
  tradingFee: BigDecimal!
  tradingVolume: BigDecimal!
  blockNumberAdded: BigInt!
  timestampAdded: BigInt!
}

type Trader @entity {
  id: ID!
  realizedPnl: BigDecimal!
  fundingPayment: BigDecimal!
  tradingFee: BigDecimal!
  badDebt: BigDecimal!
  totalPnl: BigDecimal!
  positions: [Position!]! @derivedFrom(field: "traderRef")
}

type Position @entity {
  id: ID!
  trader: Bytes!
  baseToken: Bytes!
  positionSize: BigDecimal!
  openNotional: BigDecimal!
  openPrice: BigDecimal!
  realizedPnl: BigDecimal!
  tradingFee: BigDecimal!
  badDebt: BigDecimal!
  totalPnl: BigDecimal!
  traderRef: Trader!
}

ref:
https://thegraph.com/docs/developer/create-subgraph-hosted#the-graphql-schema
https://thegraph.com/docs/developer/graphql-api
https://graphql.org/learn/schema/

It's also worth noting that The Graph supports Time-travel queries. We can query the state of your entities for an arbitrary block in the past:

{
  positions(
    block: {
      number: 1234567
    },
    where: {
      trader: "0x5abfec25f74cd88437631a7731906932776356f9"
    }
  ) {
    id
    trader
    baseToken
    positionSize
    openNotional
    openPrice
    realizedPnl
    fundingPayment
    tradingFee
    badDebt
    totalPnl
    blockNumber
    timestamp
  }
}

ref:
https://thegraph.com/docs/developer/graphql-api#time-travel-queries

Subgraph Manifest

Second, we must provide a manifest to tell The Graph which contracts we would like to listen to, which contract events we want to index. Also a mapping file that instructs The Graph on how to transform blockchain data into GraphQL entities.

A template file of subgraph.yaml:

specVersion: 0.0.2
description: Test Subgraph
repository: https://github.com/vinta/my-subgraph
schema:
  file: ./schema.graphql
dataSources:
  - kind: ethereum/contract
    name: ClearingHouse
    network: {{ network }}
    source:
      abi: ClearingHouse
      address: {{ clearingHouse.address }}
      startBlock: {{ clearingHouse.startBlock }}
    mapping:
      kind: ethereum/events
      apiVersion: 0.0.4
      language: wasm/assemblyscript
      file: ./src/mappings/clearingHouse.ts
      entities:
        - Protocol
        - Market
        - Trader
        - Position
      abis:
        - name: ClearingHouse
          file: ./abis/ClearingHouse.json
      eventHandlers:
        - event: PoolAdded(indexed address,indexed uint24,indexed address)
          handler: handlePoolAdded
        - event: PositionChanged(indexed address,indexed address,int256,int256,uint256,int256,uint256)
          handler: handlePositionChanged

Since we would usually deploy our contracts to multiple chains (at least one for mainnet and one for testnet), so we could use a template engine (like mustache.js) to facilate deployment.

$ cat configs/arbitrum-rinkeby.json
{
    "network": "arbitrum-rinkeby",
    "clearingHouse": {
        "address": "0xYourContractAddress",
        "startBlock": 1234567
    }
}

# generate the subgraph manifest for different networks
$ mustache configs/arbitrum-rinkeby.json subgraph.template.yaml > subgraph.yaml
$ mustache configs/arbitrum-one.json subgraph.template.yaml > subgraph.yaml

It's worth noting that The Graph Legacy (the Hosted Service) supports most of common networks, for instance, mainnet, rinkeby, bsc, matic, arbitrum-one, and optimism. However, The Graph Network (the decentralized version) only supports Ethereum mainnet and rinkeby.

You could find the full list of supported networks on the document:
https://thegraph.com/docs/developer/create-subgraph-hosted#from-an-existing-contract

Mappings

Mappings are written in AssemblyScript and will be compiled to WebAssembly (WASM) when deploying. AssemblyScript's syntax is similar to TypeScript, but it's actually a completely different language.

For each event handler defined in subgraph.yaml under mapping.eventHandlers, we must create an exported function of the same name. What we do in a event handler is basically:

  1. Creating new entities or loading existed ones by id.
  2. Updating fields of entities from a blockchain event.
  3. Saving entities to The Graph.
    • It's not necessary to load an entity before updating it. It's fine to simply create the entity, set properties, then save. If the entity already exists, changes will be merged automatically.
export function handlePoolAdded(event: PoolAdded): void {
    // upsert Protocol
    const protocol = getOrCreateProtocol()
    protocol.publicMarketCount = protocol.publicMarketCount.plus(BI_ONE)

    // upsert Market
    const market = getOrCreateMarket(event.params.baseToken)
    market.baseToken = event.params.baseToken
    market.pool = event.params.pool
    market.feeRatio = BigInt.fromI32(event.params.feeRatio)
    market.blockNumberAdded = event.block.number
    market.timestampAdded = event.block.timestamp

    // commit changes
    protocol.save()
    market.save()
}

export function handlePositionChanged(event: Swapped): void {
    // upsert Market
    const market = getOrCreateMarket(event.params.baseToken)
    market.tradingFee = market.tradingFee.plus(swappedEvent.fee)
    market.tradingVolume = market.tradingVolume.plus(abs(swappedEvent.exchangedPositionNotional))
    ...

    // upsert Trader
    const trader = getOrCreateTrader(event.params.trader)
    trader.tradingFee = trader.tradingFee.plus(swappedEvent.fee)
    trader.realizedPnl = trader.realizedPnl.plus(swappedEvent.realizedPnl)
    ...

    // upsert Position
    const position = getOrCreatePosition(event.params.trader, event.params.baseToken)
    const side = swappedEvent.exchangedPositionSize.ge(BD_ZERO) ? Side.BUY : Side.SELL
    ...

    // commit changes
    market.save()
    trader.save()
    position.save()
}

We can also access contract states and call contract functions at the current block (even ). Though the functionality of calling contract functions is limited by @graphprotocol/graph-ts, it's not as powerful as libraries like ethers.js. And no, we cannot import ethers.js in mappings, as mappings are written in AssemblyScript. However, contract calls are quite "expensive" in terms of indexing performance. In extreme cases, some indexers might avoid syncing a very slow subgraph, or charge a premium for serving queries.

export function handlePoolAdded(event: PoolAdded): void {
    ...
    const pool = UniswapV3Pool.bind(event.params.pool)
    market.poolTickSpacing = pool.tickSpacing()
    ...
}

ref:
https://thegraph.com/docs/developer/create-subgraph-hosted#writing-mappings
https://thegraph.com/docs/developer/assemblyscript-api

In addition to event handlers, we're also able to define call handlers and block handlers. A call handler listens to a specific contract function call, and receives input and output of the call as the handler argument. On the contrary, a block handler will be called after every block or after blocks that match a predefined filter - for every block which contains a call to the contract listed in dataSources.

ref:
https://thegraph.com/docs/developer/create-subgraph-hosted#defining-a-call-handler
https://thegraph.com/docs/developer/create-subgraph-hosted#block-handlers

Here're references to how other projects organize their subgraphs:
https://github.com/Uniswap/uniswap-v3-subgraph
https://github.com/Synthetixio/synthetix-subgraph
https://github.com/mcdexio/mai3-perpetual-graph

Deploy a Subgraph

Deploy to Legacy Explorer

Before deploying your subgraph to the Legacy Explorer (the centralized and hosted version of The Graph), we need to create it on the Legacy Explorer dashboard.

Then run the following commands to deploy:

$ mustache configs/arbitrum-rinkeby.json subgraph.template.yaml > subgraph.yaml

$ graph auth --product hosted-service <YOUR_THE_GRAPH_ACCESS_TOKEN>
$ graph deploy --product hosted-service <YOUR_GITHUB_USERNAME>/<YOUR_SUBGRAPH_REPO>

ref:
https://thegraph.com/docs/developer/deploy-subgraph-hosted

Deploy to Subgraph Studio

When we deploy a subgraph to Subgraph Studio (the decentralized version of The Graph), we just push it to the Studio where we're able to test it. Versus, when we "publish" a subgraph in Subgraph Studio, we are publishing it on-chain. Unfortunately, Subgraph Studio only supports Ethereum Mainnet and Rinkeby testnet currently.

$ graph auth --studio <YOUR_SUBGRAPH_DEPLOY_KEY>
$ graph deploy --studio <YOUR_SUBGRAPH_SLUG>

ref:
https://thegraph.com/docs/developer/deploy-subgraph-studio

Token Economics

Before we talk about the token economics of GRT token, it's important to know that the following description only applies to The Graph Network, the decentralized version of The Graph. Also, the name of The Graph Network is a bit ambiguous, it is not a new network or a new blockchain, instead, it is a web service that charges HTTP/WebSocket API calls in GRT token.

To make GRT token somehow valuable, when you query data (through GraphQL APIs) from The Graph Network, you need to pay for each query in GRT. First, you have to connect your wallet and create an account on Subgraph Studio to obtain an API key, then you deposit some GRT tokens into the account's billing balance on Polygon since their billing contract is built on Polygon. At the end of each week, if you used your API keys to query data, you will receive an invoice based on the query fees you have generated during this period. This invoice will be paid using GRT available in your balance.

ref:
https://thegraph.com/docs/studio/billing

When it comes to token economics:

  • Indexers earn query fees and indexing rewards. GRT would be slashed if indexers are malicious or serve incorrect data. Though, there's no documentation about how exactly slashing works.
  • Delegators earn a portion of query fees and indexing rewards by delegating GRT to existing indexers.
  • Curators earn a portion of query fees for the subgraphs they signal on by depositing GRT into a bonding curve of a specific subgraph.

ref:
https://thegraph.com/blog/the-graph-grt-token-economics
https://thegraph.com/blog/the-graph-network-in-depth-part-1
https://thegraph.com/blog/the-graph-network-in-depth-part-2

Query Fee

The price of queries will be set by indexers and vary based on cost to index the subgraph, the demand for queries, the amount of curation signal and the market rate for blockchain queries. Though querying data from the hosted version of The Graph is free now.

The Graph has developed a Cost Model (Agora) for pricing queries, and there is also a microtransaction system (Scalar) that uses state channels to aggregate and compress transactions before being finalized on-chain.

ref:
https://github.com/graphprotocol/agora
https://thegraph.com/blog/scalar