Covalent Ethereum Wayback Machine Takes You Back in Time and Provides AI Models With Structured Data.
I'm 99.9% sure that you clicked on this article wondering how the Ethereum Wayback Machine can possibly take you back in time.
I gladly welcome you to my blog—all you have to do is stick with me.
If this is your first time hearing about this machine, you are right in time. And yes! This is a time machine, but not the type that takes you back to where you lost your ETH to help you recover it. Lol.
Let’s find out what it actually is:
Understanding the Problem
To understand the Ethereum Wayback Machine, we first have to understand the problem.
Ethereum has experienced frequent cases of network congestion over the years, especially when the transactions on-chain are higher than usual, leading to high gas fees and low latency.
This has led to the development of rollups that serve as secondary layers for execution and data availability, while referring back to Ethereum for consensus and final validation.
The division of labor helps in boosting Ethereum’s scalability and efficiency. But then, issues are raised pertaining to the concerns on how data are made available on these roll-ups.
Instead of storing archival data within client nodes, rollups are driven by the Ethereum Improvement Proposal 4844 (EIP-4844) which uses a novel data structure referred to as “blobs.”
These blobs save on-chain data and are required for validation. Additionally, they come with a state expiry programmed to delete recorded data after a specific period of time, typically between 1 – 3 months.
Client nodes only focus on execution, consensus, and validation, while the storing of historical data is outsourced to outside infrastructure actors. All this helps to improve the Ethereum network sufficiently.
However, deleting historical data? I'm not sure I like that.
It defeats one main purpose of blockchain—a decentralized system of trackable historical data, which includes immutable record of transactions, smart contracts, and activities on a blockchain network. Once data is deleted, the traceability of that data back in time is breached.
The roll-up era has introduced this new challenge — ensuring the continued accessibility of historical blockchain data while relying on rent-seeking, centralized intermediaries for data storage.
Therefore, long-term data availability is crucial!
But why should anyone give a damn about long-term data availability?
First thing first — as I said earlier, blockchain wouldn’t be blockchain when on-chain data can’t be tracked way back. What’s the essence of the word: “chain”?
To add, long-term data availability ensures transparency because the full historical transaction record of blockchain, dating back to the beginning, can be accessed and analyzed by anyone. With these records, a blockchain or dApps can adhere to regulation compliance.
Frauds and other illegal activities are easy to investigate with historical data. Also, investors could require historical data on a dApp’s blockchain activities to make investment decisions. When it dates back to only 3 months ago, it could be a total disaster.
Then, you think of adoption? If I were a developer, I wouldn't go for blockchain that wipes all my on-chain data after just few months.
Please spare me the short term memory and restore my history!
Even as a user, I'm quite disturbed by this. Why would I go looking for transaction I made 4 months ago and can never find it?
No, please, thank you. At that point, even centralized systems offer better storage programs.
Thankfully, Covalent saw these issues and offered Ethereum Wayback Machine (EWM) as the messiah.
What is Ethereum Wayback Machine?
Ethereum Wayback Machine (EWM) was launched as the data storage infrastructure for on-chain archival/historical data.
The specifications, source codes, and the reference implementation that powers EWM is open-source and empowers blockchain users, devs, researchers, etc., to access and retrieve cryptographically secure historical data readily without relying on centralized, rent-seeking intermediaries.
The Language Model
The Ethereum blockchain is made up of raw unstructured data (transactions, blocks, and the overall arrangement of data) and structured data (organized, formatted, processed, and stored data).
To query this data, a language model is used called the Ethereum Query Language (EQL).
Just like how SQL functions, EQL uses queries to retrieve information from the Ethereum chain. With the language, users can access the chain’s database, locate, and retrieve information through the EWM.
The Covalent Ethereum Wayback Machine Will Improve Web3 AI Models.
AI has been among the top trends on Web3 this season, but as you should know, AI relies heavily on data availability, especially structured data, to function properly. Therefore, Covalent is one of the most suitable data structures that Web3 AI needs.
Covalent has a unified API that spans over 225 blockchains and 240 million wallets.
It provides a pool of cryptographically secure, standardized, and structured data that AI can easily access, enabling the creation of AI-driven solutions that use both historical records and real-time data.
How Does EWM Provide AI Models With Structured Data?
EWM utilizes two(2) basic components to achieve long-term data availability:
- Block Specimens Producer (BSP)
- Block Results Producer (BRP)
In addition, there are other components that make up the process such as:
- Query Nodes,
- Delegators
- Event Sequencers,
- Refiners,
- IPFS
Let’s see how all these are interlinked in the EVM and AI process.
The step-by-step is illustrated as follows:
- An AI application writes (requests) to retrieve a particular data from Ethereum blockchain:
- The RPC nodes receive the request.
- The BSP (which runs on top of popular Ethereum clients like GETH, reads from the RPC) uses cryptographic proofs to specifically extract data from the blockchain, and output it to a storage format known as ‘Block Specimens’.
- The BRP collects the block specimens and uses ‘the Refiner’ to process and transform the block specimens into enriched data outputs called ‘Block Results’.
- The Query node operator interacts with the BRP and collects the block results to send the final result to the AI application.
And like that, the EWM has been able to provide the AI access to available historical blockchain data.
The Role of Delegators, Event Sequencers, Refiners, and IPFS.
Delegators stake Covalent Query Tokens (CQT) to network operators (BSPs) to support their tasks and in turn receive rewards and extra benefits like logs from event sequencers.
The event sequencers process and organize on-chain events, like smart contract and transaction interactions, in a structured format.
So typically, whatever query a BSP or BRP logs in, the sequencers process and logs out the block specimens and the block results, respectively.
Covalent also integrates the refiner into the BRP, to facilitate the transmission of the refined output data (block result) to the IPFS for decentralized storage purposes.
Why Does The Ethereum Wayback Machine Stand Out for AI Models?
Aside from having long-term historical data readily available, data is also accurate because they are generated through cryptographic proofs which ensures data integrity and allows users to verify the accuracy of the data.
EWM offers AI with enrichment abilities, such as contract tracing and off-chain NFT metadata which empowers developers and users with in-depth insights and analytics—something AI models should not lack.
The Fact? More improvement to the Ethereum Wayback Machine will continue to be updated, which will in turn boost AI’s presence in Web3.
This is only the beginning and maybe, one day, we'll be able to go back in time and restore lost ETH (don't quote me anywhere), but until then, I hope you enjoy EWM.
Learn more about Covalent and follow them for more updates:
#CovalentAlchemist Covalent
