Often termed the Google of blockchains, The Graph is a decentralized protocol for indexing and querying data from blockchains. Sounds complicated? Think of it in the way Google indexes the Internet. In indexing, the Googlebot analyses a web page for content and meaning, stores it in the Google index, which then shows up in Google Search results. Quite similar to a books library.
Books are organised as per a code, labelled by category, and so on, thus making retrieval easy. The Graph, on its part indexes blockchain data, grouping it into open application programming interfaces (API) called subgraphs, accessible for use by anyone. The utilisation of the platform has grown rapidly, as indicated by usage among projects including Uniswap, Synthetix, Gnosis, AAVE, Decentraland. If, for example, you’ve used Uniswap, or delved deeper into it, chances are you’ve come across Uniswap Analytics. Much of that is powered by The Graph.
Understanding The Graph
Before delving further into this project, a quick understanding of how The Graph works may be helpful. Here’s a look at the protocol roles within this ecosystem:
The Graph makes difficult-to-query information, accessible in an easily queried way, for individuals or organizations (Consumer). How a typical transaction occurs is spelt out below:
- The consumer asks/queries The Graph which Indexers have the data they are seeking
- A list of Indexers is provided
- The consumer selects an Indexer based on which they deem most likely to provide the highest quality service at the best price
- Pricing is set between an Indexer and a Consumer, rather than at the network level
- If the Indexer accepts the price offered by the consumer, then they process the query and respond with the resulting data
Thus, The Graph functions as a marketplace for specific data on a blockchain like Ethereum. Each data set on this marketplace is called a Subgraph and can be accessed through the Graph Explorer. Each Subgraph carries a description of specific smart contracts within those dApps. This includes details like the subgraph description, developers, project links, query fee data and charts, deployment ID, API endpoint and other metadata. Below, is an example of the Subgraph for Curve (Source: Graph Explorer).
Using the Graph Explorer, an app developer, for example, can query the data needed for their dApp. They do so using The Graph’s querying language called GraphQL. Once a data request is received, the Indexer searches through relevant subgraphs to find the information. Indexers decide on the subgraphs based on what’s referred to as a Curations Signal. This has been put in place by Curators who develop subgraphs and evaluate them for quality.
The journey so far
The Graph was founded in 2018 by Yaniv Tal (Project Lead), Jannis Pohlmann and Brandon Ramirez. The trio has previously worked together on several startups. Having begun diving into Ethereum in early 2017, they realised soon after that the tooling and lack of mature protocols made it difficult to build dApps on Ethereum. This inspired the want to make open data more accessible, leading to the incorporation of The Graph.
The first prototype was built soon after, in late 2017. The mainnet launched in 2020. Over 200 Indexers deployed their nodes in the testnet. The Graph has, since, grown to support indexing data from Ethereum, IPFS (InterPlanetary File System), and PoA. The current total number of subgraphs is more than 22,000 and the network has over 8091 delegators; 157 indexers and 2322 Curators (Source: Graph Explorer). Over 23,000 developers have built subgraphs for applications, such as Uniswap, Synthetix, Foundation, Zora, KnownOrigin, Gnosis, Balancer, Livepeer, DAOstack, Audius, Decentraland, and many others.
As per company reports (dated May 2021), from an approximate 30 million daily queries in June 2020, daily query volume in 2021 has grown to over 675 million queries per day. Between May 2020 to April 2021, over 110 billion queries were processed. Query volume has also increased month-over-month (MoM) with 11.7B in January, 14.5B in February, 19B in March, and 20.3B in April.
As of Q1 2021, DeFi makes up 57% of queried subgraphs. 11% of subgraphs serve NFT marketplaces, arts, and gaming. 7% of subgraphs are related to cross-chain use cases like synthetic bitcoin and protocol bridges, signifying an increase in interoperability between dapps.
DeFi has also dominated dApp adoption, making up 90% of the hosted service query volume. Network-data focused subgraphs make up 4% of query volume. Query volume for NFT use cases like art profiles, NFT minting, and marketplaces have risen in the last few months with increasing mainstream traction. NFTs make up 2.5% of the current query volume.
The GRT token
The Graph’s native cryptocurrency, GRT, is used to ensure the integrity of the data queried on the network. All users on the platform, Indexers, Curators or Delegators, are required to stake GRT tokens to perform their roles, in return, earning fees from the network.
Two crucial functions of the GRT token within the protocol include:
- Indexer Staking: Indexers deposit GRT to be discoverable in the query market and to provide economic security for the work they are performing
- Curator Signaling: Curators deposit GRT in a curation market, where they are rewarded for correctly predicting which subgraphs will be valuable to the network
GRT is an ERC20 token with an initial supply of 10 billion. It has an inflation rate of 3% per year, used to pay indexing rewards to Indexers. 1% of all query fees go towards burning GRT. The withdrawal tax charge to Curators is also burned and the same goes for any unclaimed rewards from the rebate pool distributed to network participants. The initial token sale was at $0.03. The sale saw a very encouraging response, and the token registered a high of $2.84 (February 2021) soon after listing. At the time of writing (2 December), it trades at $0.90 with a market capitalization of $4.49 billion and a circulating supply of 4.95 billion.
Here’s a look at the tokenomics:
Given that there’s a sizable proportion of the tokens allocated to parties beyond the general public, the continuous inflow of tokens upon completion of the respective vesting schedule can threaten to keep sell pressures high. The GRT vesting and distribution schedules range between 6 months to 10 years. Here’s a visual look at the vesting schedule,
Token holders can stake GRT tokens by delegating them to Indexers. In return, they receive a portion of the reward. Tokens can be staked through The Graph staking UI. A word of advice here, the process is not stated upfront and may require some degree of manoeuvring. There’s a detailed guide available on the The Graph Academy. Here’s a quick look at what to expect:
The page lists details of Indexers, highlighting:
- Subgraphs: Showcasing which subgraphs are indexed by the indexer
- Query Fee Cut: Query fees are earned every time an indexer serves data. This shows how much of the total query fees goes to the indexer. The rest is shared among the delegators, proportional to their share of delegated tokens. For example, the first indexer in the screenshot above has a 50% Query Fee Cut. Thus, they would keep 50% of the query fees for themselves. The delegators get to share the remaining 50%. A lower Query Fee Cut is generally better for the delegators.
- Effective Reward Cut: Indexers earn indexing rewards for their own stake as well as delegated tokens. The Effective Reward Cut describes the commission paid out to the indexer for providing their services.
- Stake Owned: This is the amount of GRT the indexer is “staking” on The Graph Network. If the indexer indulges in some kind of malpractice, these tokens can be slashed by the protocol.
- Delegated: This is the number of tokens currently delegated to the indexer. A lower amount of delegated tokens translates to a higher share of the fees and rewards to the delegator.
- Query Fees: While not visible in the screenshot above, the table also shows the total amount of query fees an indexer has earned. Ideally, a high amount of Query Fees, combined with a low Query Fee Cut and a small delegation pool would be ideal (Source: The Graph Portal).
- Indexer Rewards: This is the total amount of indexing rewards the indexer has accrued. Again, a high amount of indexing rewards earned, combined with a low Effective Reward Cut and a small delegation pool would be ideal.
Thus, based on the metrics above, staking GRT can necessitate some degree of research work. There are, however, secondary alternatives to staking GRT. Binance, for one, offers a Locked Staking option of GRT, with an estimated APY of 13.56% and locked in periods of 30, 60, and 90 days.
Now that the role of an indexer is clear, let’s take a quick look at the role of a Curator and see if it’s a profitable one. As mentioned earlier, Curators analyze subgraphs and curate high-quality subgraphs. Thus, for example, once a Curator has found a potentially attractive subgraph, they curate it by signalling on it. In doing so, Curators highlight to Indexers, which subgraphs are valuable and should be indexed accordingly.
As per the team at The Graph, delegation is a safe and stable way of generating interest while not exposing oneself to the high risks Curators are confronted with. So while Curation offers large potential gains, it is significantly riskier as Curators may be confronted with the loss of capital. Currently, a minimum of 100 GRT (˜$89) is needed to curate a subgraph with no signal. Curators also have to pay gas costs and a curation fee. For this reason, it may not be profitable to signal small amounts of GRT.
Here’s a quick look on how curation actually works:
When Curators signal on a subgraph, they actually deposit GRT into a bonding curve. Doing so mints curation shares. Now, 10% of all the query fees generated on the subgraph are deposited into the bonding curve. Thus, the value of each share increases if a subgraph receives queries. To get rewarded eventually, Curators need to burn their curation shares. (To know more about this, click here). Again, it is important to note that as a Curator, one is confronted with the risk of withdrawing less GRT from the bonding curve than they started with. In the worst-case scenario, the curation shares decrease substantially in value so that one receives only a tiny fraction of the GRT back that was initially deposited.
The Graph has a sizable online social presence, as reflected by the following on its channels. Twitter clearly is the mainstay channel with 168K followers. Updates are regular and include informational, educational content. There are even regular community talks. Other prominent social channels include Telegram (26K), Discord (28K) and Reddit (20K).
While some have linked The Graph as a competitor to Chainlink, there remains a fundamental difference in what the two are doing. While Chainlink acts as an oracle, bridging the gap between real-world data and the blockchain world, The Graph makes blockchain data available off-chain.
Roadmap and future
The Graph doesn’t currently have a clear roadmap. During the current year, the project has increased focus on moving away from having a highly centralized development side. Edge & Node, a software development company launched earlier this year, led by the core protocol team behind The Graph, has a 2-year service agreement with The Graph Foundation to help maintain the core protocol and build new trust-minimized tools and applications alongside The Graph community. More recently, with a focus to decentralize core protocol development, two new development teams have joined the network. They include StreamingFast and Figment.
Developments over the year have included:
Launch of the Subgraph Studio: Enabling developers to deploy and curate on subgraphs, and earn query fees. The two months post-launch in July saw more than 180 subgraphs having migrated from the hosted service to the permissionless Subgraph Studio, a 2150% growth in just two months.
Graph Improvement Proposal (GIP) Process: GIPs are proposed and discussed in The Graph Forum to receive feedback from subgraph developers, Indexers, Curators and Delegators. GIPs are then put up for a Snapshot vote to unofficially gauge community sentiment, followed by a vote by The Graph Council before the upgrades are implemented.
Improvements in costs and performance: StreamingFast, in September, shared their enhancements that could increase indexing speed by 300x. Over the past months, the network has seen query fees range between 0.00002 GRT and 0.0002 GRT per query. The cost per query is expected to change with the shift in network dynamics, such as increased migration, greater query volume and Indexers becoming more efficient in managing their operations.
The Graph has successfully put together an indexing protocol offering developers easy access to on-chain data in a decentralized manner. As an open data marketplace where queries are individually priced and optimized for access not only making the Web3 stack possible, The Graph also simplifies developer work and creates better performance for dapps at scale.
One of the biggest opportunities for The Graph going ahead includes expanding to other blockchains. As stated earlier, between May 2020 to April 2021, over 110 billion queries were processed on The Graph’s hosted service. April this year saw over 20 billion queries registered. And this largely comes from just one blockchain – Ethereum. Over the last year, successfully integrated chains have included EVM-based L1s, Fuse, Celo, Avalanche, Clover, Binance Smart Chain, and Fantom and L2s such as Polygon. Most recently, The Graph announced the indexing of NEAR, a Layer 1 blockchain, and the first non-EVM (Ethereum Virtual Machine) compatible blockchain supported by Graph Node. The team has indicated a focus on multi-blockchain expansion, with a focus on Solana and Polkadot.
Will it emerge as the Google of Blockchains? It’s certainly on the right path. However, what remains to be seen, as stated above, is how successful The Graph is in integrating other blockchains and maintaining its first-mover advantage. The blockchain space is evolving every single day, and competitors threaten to spring out of nowhere. Striving to reach Google’s stature is great. Yet the path is bound to be a long and challenging one.