README
TApps are designed for Web3
As is the case with all TApps, the TEA Party showcases the special features that are beyond the capabilities of other cloud based internet (web 2.0) applications. Instead of centralized server(s) hosting the app, the individual miners of the TEA network host TApps based solely on their own hosting profitability. The inherent decentralization that all TApps including the TEA Party share gives these apps even more unique features:
They cannot be turned off by any centralized power. As long as there are a minimal number of miners hosting any particular application, it will continue to run forever.
No one, including the host miner, can control or censor the content. The content is owned and protected by its creator's private key. A miner can choose to stop hosting the TApp, but it cannot selectively choose what content to show or hide.
There's no free lunch. Every action that costs any computing resources needs to be paid by someone. In TEA Party's particular case, every message sent needs to be paid for. Additional charges also apply to store the message or to notify the recipient. For further information, read more about where the messages are stored.
In order to get the features above, the underlying technical layer is very different from the existing cloud computing and blockchain tech stacks. It's a new tech stack that's based on recent technologies. However, the developers do not need to understand the complicated low level distributed system, they can build application as if it is still a centralized cloud computing architecture. This is the charm of the TEA Project.
The following sections will explain the cutting edge technologies used in the TEA Party. We hope explaining the underlying technologies and how they work together will help you make your own TEA applications (TApps).
Three Major Parts
The full architecture of the TEA Project is complicated but there are only three majors parts where an application developer will be working. These three major parts all run in different locations just like the traditional 3-tier-architecture of cloud computing web apps.
The front-end is typically a JS application (for webapps) or a mobile application (for mobile apps). But the front-end isn't dictated by TEA, and the developer can use whatever front-end they're comfortable with.
This WebAssembly code is running inside of a hosting node. The hosting node is a miner's computer which has a CML planted. It's similar to the server logic running in back-end servers or application servers in the traditional cloud computing architecture.
This WebAssembly code is running inside the state machine's mini-runtime. It's equivalent to the stored procedure (SQL for example code) in the traditional 3-tier architecture's database.
Three-tier architecture basic workflow
The above 3 components are directly mapped to the traditional 3-tier architecture of a typical cloud computing application.
The basic workflow would look like this: (this example uses a web-based TApp)
The user generates a user action in the front-end. The Javascript web client catches the user action, generates a web request, and sends it to the backend.
The back-end receives the web request and runs the Tea Party back-end code (we call it the back-end actor) to handle anything that does not need the state machine (traditionally, this is referred to as a database). But when it needs to query or update a state in the state machine, it will need to generate a request to the state machine tier. These can be broken down into queries (will not change the state) and commands (potentially could change the state). Commands are typically called txns in the blockchain industry.
The queries and commands are handled by the state machine replications. For queries, it will look up the local state and send the result back. For commands, as one of the replications, it should not modify on its own. Instead, it generates a txn and puts it in a global queue that we call the conveyor. The replicas run a Proof of Time consensus to guarantee that all state machines in all replicas get the same order_of_txns. This ensures that their state can always be kept identical after executing the command. This is the same methodology as is typically used by a distributed database system.
Storage
There are three types of storage options for different use cases.
orbitdb: Based on IPFS / used for large blob storage. It's running on the hosting CML.
State: Usually used to store account balance. It runs inside the state machine.
gluesql: Distributed SQL server instances. It's located inside the state machine.
Comparison between the three storage options
Comparison with a cloud webapp's 3-tier architecture
The requirements for building TApps
In this section, we'll list the knowledge and tools you'll need to build TApps.
Tools
To build and run the demo locally, you'll need:
A Mac or Linux machine.
Docker and docker-compose installed.
Rust compiler.
Web browser.
After building your own TApp, you can try hosting it by launching your own mining node. A mining node is any type of TEA node with a CML planted in it. If you don't own a physical TEA machine, you can rent an Amazon Nitro VM which is TEA-compatible.
Programming languages
The demo TEA Party app is written in the following languages:
Front-end is written in JS and the Vue framework. But this is just what we chose for this example: the TEA Project is agnostic as far as front-ends.
Back-end and State machine actors are written using Rust and then compiled to WebAssembly.
The TEA Project doesn't require the developer to use the Rust programming language. You can use any programming language that compiles to WebAssembly. But at this moment, in order to understand our existing demo code, you'll need to use the Rust langauge.
Architecture knowledge
Layer2 without Rollups
The TEA Project is considered a layer2 solution, but it has been designed with completely different mindset in comparison with existing roll-up solutions. We focus on providing a trustable computing infrastructure, hence why there's [no need to verify the computing results](The_future_and_innovation_of_layer2.md#We Trust the Result by Verifying the Environment). This allows the dApps running on our infrastructure to run at full speed, similar to cloud computing.
Layer2 decoupled with layer1
TEA Project runs [on top of different kinds of blockchains](The_future_and_innovation_of_layer2.md#Layer1 Agnostic) interchangeably due to there being no rollup required. The layer 1 blockchain provides one of the three Roots of Trust, with the other two roots of trust coming from hardware.
Hardware
The TEA Project is very different from many other blockchain projects. TEA relies on two types of hardware in order to reach a special type of consensus:
Please click the above links to learn more about how and why the TEA Project uses these technologies.
If you just want to run the code in your local simulator, you don't need any of the above hardware. You can run our simulator using Docker to run a test environment.
If you want to host your application in a production environment, you'll need a TEA node. If you don't own one, the easiest way is to rent an Amazon Nitro VM.
Code walk through
TEA Party is a demo Web3 app running on the TEA Project. To explore a walk through of the TEA Party application's sample code, please continue reading the code walkthrough.
Last updated