What kind of resources does a node need to participate?
The answer is minimal, most cloud instances can easily perform. The biggest drawback in-terms of performance is networking. I’ve seen when there are BGP leaks happening around the internet, some routes become susceptible to DDoS type behavior causing the process to slow down or halt.
What about TPS?
See this statement by Gav with Polkadot.
Answered by Gav: Reason for using asynchronous rather than synchronous communication? Difference in terms of TPS?
First some terminology. Cross-chain communication has two paradigms: active and passive. They might also be named “mutable” and “immutable” or “push” and “pull” or “write” and “read”. Active is where one chain, actually sends a message into another chain which would then have some effect in that receiving chain. Passive, on the other hand, is where one chain simply reads some recent information about some other chain and takes some action itself.
In wholly synchronous systems like Ethereum, both are synchronous. One smart contract sends a message to another, and is halted until any effects are caused and it receives any reply from them. Achieving wholly synchronous inter-chain communication is something that I threw out of the window 30 seconds into the design of Polkadot. For active communication, Polkadot’s parallelisation means that it must be asynchronous, so token transfers between chains, for example, will need to be async.
For passive communication however, there is no such requirement needed. Given any witness data, Polkadot parachains will be able to interpret other parachains’ recent states instantly and trustlessly.
Transaction throughput doesn’t really come into it - it’s all about the fact that it’s an architecture built for large-scale parallelism from the ground-up.
A lot of the same concepts will apply to DeepBrain Chain as well, so take note of the Transaction throughput doesn’t come into play. It scales as much as we want to scale the network.
Currently DeepBrain Chain nodes will only run on Linux, macOS, Docker and Kubernetes . We plan to support Windows in the near future.