The design described in this document is applicable to any other public EOSIO blockchain. But WAX is the first place where it's actually needed.
Back in 2019, Alexandre Bourget and others have published an EEP for EOS Mainnet Extensions. Some ideas in this proposal are overlapping with that document.
WAX is a public blockchain using a slightly modified EOSIO software (RSA signature validation intrinsic is added for RNG implementation). Due to growing popularity of NFT, easy on-boarding, and low fees it has experienced an enormous growth in the course of 2021. As of August 2021, the total state database is over 32GB in size. The network has over 2 million accounts, and several millions NFT objects in circulation.
In addition to the size of the state memory, CPU and RAM costs have grown dramatically.
One of significant growth factors is easy user on-boarding via WAX Web Wallet. The users are authenticating via social networks, and WCW is managing the keys for the users.
The public API infrastructure is heavily loaded with user traffic, mostly related to AlienWorlds mining. This motivates the projects to build and maintain their own API infrastructure.
The blockchain CPU resource is quite well congested, so the users and dapps have to stake significant amounts of tokens to be able to operate.
-
Transparency for the user: the WAX user should operate on the sidechain seamlessly, using the same private keys and account names.
-
Predictable resource costs: WAX mainnet is growing fast and in an unpredictable way. It makes it difficult for a business to plan the costs and future growth. The sidechain should provide a clear and predictable cost structure. The sidechain design should also allow building chains that are dedicated for one business application.
-
WAX as a primary currency: the sidechain services should be paid in WAX. This increases the market demand for the token and unifies the business landscape.
-
Standalone and long term operation: the need for teleporting resources, such as fungible or non-fungible tokens, between the side and main chain should be minimal, if existing at all.
-
Playground for alternative resource models: a sidechain can introduce a new way of charging for resources. For example, transaction fees could replace staking and RAM purchasing.
-
Independent infrastructure budgeting: the sidechain may define its own way of defining the number of block producers, the method of selecting them, and rewards structure. Ideally, a sidechain should operate as a self-sustainable business.
The sidechain system contract should prevent normal accounts from creating new accounts and changing their permissions. These operations should only be allowed to the oracle accounts.
Several independent oracles are reading the account information updates from WAX state history nodes and create the sidechain multisig transactions that create new accounts or update the permissions on existing ones. The goal is to reflect all WAX accounts on the sidechain with identical permissions.
A sidechain may also choose to synchronize only a subset of WAX accounts. For example, only owners of a certain NFT or fungible token would be allowed to operate. Although this would save on system resources, it reduces the transparency and makes uniform front-end operation more complex. The system should be able to utilize any existing information on WAX as a filtering criteria, such as:
-
fungible or non-fungible token ownership;
-
account permission with a specific name;
-
account names from certain realms (such as
*.wam).
The filter should also allow two ways of dealing with the accounts that fell off the filtering criteria:
-
keep them in the system and keep synchronizing their permissions;
-
nullify their permissions on the side chain.
Contracts and memory content should not be automatically mirrored to the sidechain. The contract owners should buy the sidechain resources and deploy their applications according to their needs.
The sidechain infrastructure should provide a uniform way for wallets and dapps to find relevant information in order to auto-configure themselves.
The location service should provide the following information:
-
Chain name: a valid EOSIO name of up to 12 characters long.
-
Chain ID: sha256 hash of the genesis.
-
Link to the founding documentation, terms and conditions.
-
Whether the sidechain imports all accounts, and the filtering criteria.
-
A copy of genesis JSON file.
-
Link to a listing of public API endpoints, such as P2P, nodeos API, history API, and other APIs like Atomic Assets.
-
Billing information: resource prices and the way to purchase them, in some formalized and machine-readable way.
Listing of public endpoints can be implemented in a smart contract where service providers would edit the records for their own services.
Ideally, all payments for the sidechain service should be in WAX and stay on the WAX blockchain, in order to avoid the teleporting infrastructure and unnecessary custodial accounts.
Also it is preferable that the sidechain does not create a tradeable or speculative token.
The WAX user that needs to purchase a resource on the sidechain interacts with the payment gateway contract, such as POS contract made by me, and receives a resource token that can only be spent on particular sidechain resource. The user interacts with the resource management contract and indicates which sidechain accounts should receive the resources and their amounts.
Oracles on the sidechain detect the instruction to top-up a certain sidechain account resource, and report back to WAX contract when the operation succeeds and irreversible.
The resource token should never be transferable from the sidechain back onto WAX blockchain.
Infrastructure providers get paid from the resource management account in WAX, according to their business agreement. It makes sense to indicate several categories of infrastructure service, and bill them separately:
-
Block signing and chain security. The infrastructure providers do their best to distribute the block production servers geographically and place them in legally independent locations. High-volume chains require also that network latency is within certain limits, so in some cases the production servers should reside within the same continent, following major Internet backbone destinations. As I have shown in an article, a chain of 5 producers tolerates an outage of only one of them. A network of 9 producers would still be alive if two producers go offline.
-
Public p2p service. As it is essential to have p2p access as much distributed as block production, this service would normally be provided by every infrastructure team.
-
Public EOSIO API service. A small blockchain does not necessarily need every BP to set up their own API service. It could be one or two instances with necessary level of redundancy.
-
Public history service, such as Hyperion. History services require certain skills and amount of hardware, and it makes sense to motivate a skilled team with an adequate budget.
All in all, the network should provide a service that it can afford.
Also the infrastructure providers want transparency and adequate payment for their service, in a token that is easy to sell on the market. Thus, having all payments in WAX is the easiest way to perform the billing. As an alternative, a stablecoin on WAX could be chosen as a means of payments for the sidechain resource.
Sidechains open up a larger field for experimenting and trying different resource models. Here are some possible options:
User stakes the system token for CPU and NET, and buys RAM from Bancor algorithm contract. Major drawbacks:
-
The network does not earn from usage. Free transactions become a heavy burden for the infrastructure.
-
Unpredictable CPU and RAM costs, which makes business planning difficult.
The benefit is that the technology is proven and not much needs to be changed in vanilla system contracts.
There are several approaches to introducing transaction fees in an EOSIO blockchain. To name a few,
-
FIO is running a heavily customized version of nodeos. Also it's an application-specific network, so only the system contracts are allowed. Each of system actions has an associated price factor, and the final price is defined by a median from prices suggested by block producers. Each action has also a spending limit, so that the user controls the costs on their side, disallowing the chain to charge more than expected.
-
Letting the contracts access the resource usage. This option needs a feasibility research.
-
Setting the replenishing window to infinity. The change itself is not difficult, but implications need to be researched.
If a sidechain is dedicated to a particular application, the billing can be as simple as a monthly invoice to the customer. The total bill would be determined by current infrastructure costs, and the customer would just order additional capacity if required.
The difference of a sidechain from a public decentralized chain is that it's dedicated to a particular business purpose. In such an occasion, you would not expect anyone to register a BP and ask for votes. Even more, voting is not needed at all. It's rather a business agreement (or a legal contract) between infrastructure providers, and between the network and its customers.
Such agreements should clearly document both the process of adding a new infrastructure provider, removing it, or a way for the provider to voluntarily leave the service.
One of potential applications for a sidechain could be a service that is only available to KYC'ed accounts. But this would make the KYC flag associated with WAX accounts publicly readable, which is a serious privacy and security concern.
Probably the best case for a KYC'ed network would be to enforce generation of random account names and using the keys that are not known on any other public blockchain. Of course there would not be any account synchronization with WAX mainnet, but the sidechain would still be able to reuse the core components for gatewaying and billing.
-
Uniformity and reusability. The gateway contracts should be open for anyone to set up their own sidechain and billing workflow without having to copy and deploy them.
-
Open source permissive licenses. The integration software, such as oracles and smart contracts, should be available for anyone without limits.
-
A blockchain dedicated to a game or some other high-traffic application: transactions are frequent and bulky, and probably it's not important to keep them forever. The blockchain can define a rotation period, and all history solutions would only be indexing for a certain time interval.
-
A blockchain dedicated to an application requiring large amounts of RAM.
-
Archival blockchain: transactions would be relatively expensive, but the blockchain would guarantee long-term durability. This kind of service could be of demand for financial applications.
Copyright 2021 cc32d9@gmail.com
This work is licensed under a Creative Commons Attribution 4.0 International License.