There are different projects for cloud computing, such as Akash and Flux Network. Are there any similar projects on Substrate? Creating a Substrate pallet can be reused in different projects for scalable cloud computing. Applications don’t have to rely on the Akash network for hosting their apps; they can build their custom incentive system and pay with their own token, and provide all offchain computing features of blockchain dapp like indexing, database etc.
In Substrate cloud computing projects like Phala exist, but they focus on maintaining confidentiality with Trusted Execution Environments (TEEs).
Additionally, Docker container hosting is not required; hosting with WebAssembly (e.g using wasmedge with server solutions like Tokio, Actix, etc.) with memory and local storage will suffice. An incentive system pallet is also needed for trading computation and storage.
To clarify for Phala, we do not focus on confidentiality for computation. We focus on secure off-chain computation to connect our secure off-chain programs (Phat Contract) with the Internet via HTTPS requests.
We advise users to use Phala like serverless functions via Phat Contract queries & not like an on-chain private smart contract protocol like Secret Network. This is an important differentiation given we function as a chain agnostic solution that can be used to deploy Customized Oracles (ink! Or JS), compute intensive tasks, batch https requests, scheduler, connect cross-chain, etc.
Heres a starter kit to get familiar with the type of off-chain computation we provide for users to deploy.
Example to connect to S3 API on 4Everland from Phat Contract Rust SDK
To close, privacy by default is just a side effect of our secure off-chain computation model. Hope this helps clarify what Phala’s main utility and focus is.
Good question.
We thought about the direction of “De-AWS” but we passed that idea.
Reasons?
For generic computation, we bet Web3 apps don’t need docker-servers running on a de-network. Instead, we think ppl need a entire brand new level computation, which should fit with following qualities:
Verifiable for computation execution results
Edge
Function as a service
More flexible than smart contract
The world also need specific computation protocols, e.g:
Yes, for web3 computation we don’t need docker containers.
But what about running the blockchain itself? We need to run substrate in docker or maybe through WasmEdge as substrate is just a rust code.
We do need decentralized hosting services for substrate chains, without getting tied to Amazon or Google hosting services.
Also, polkadot can subsidize the hosting fees (substrate chain only), so that one can run e.g. 500GB of storage, 1-2 cores for free.
Further hosting on a service provider requires a lot of system administration work such as Linux commands etc and maintenance throughout the time, providing specialized infrastructure for substrate chain can reduce the work, just upload once, and it’s done.
Though one can host substrate chains on flux network with affordable pricing, and documentation says 99% uptime (not tested), more option is better.
For generic computation, we bet Web3 apps don’t need docker-servers running on a de-network. Instead, we think ppl need a entire brand new level computation, which should fit with following qualities: