I wonder the difference between platforms that use different blockchain protocols.

It's not easy to answer, but it's a good question.

As Mat0 of the stack overflow community mentioned, one of the differences is the consensus algorithm. Some use the proof of work (PoW) method and others use other methods to solve the Byzantine general problem (PBFT, SBFT, PAXOS derivative, etc.). As another example, Hyperledger Sawtooth Lake developed Proof of Elapsed Time (PoET), an innovative algorithm differentiated from the previous one. As another example, a framework such as Hyperledger Fabric also supports consensus algorithms in a plug-in manner.

Another difference is how data and transactions are handled. UTXO There is a platform that handles limited transactions using the protocol, and Hyperledger Fabric, Hyperledger There are also platforms that can handle complex transactions using smart contacts such as eum.

There are also platforms built for specific use cases, while others are designed for a variety of purposes.

Finally, there are platforms created for untrusted network environments (unauthorized participants) and vice versa, for trusted network environments (unauthorized participants).

Hyperledger is a platform designed for enterprise use. The platform is evolving in the hope that companies develop together or in competition will increase interoperability and gradually integrate.

A lot of research is needed to determine the platform, a clear understanding of the use case and requirements of your project, and a sufficient experiment is required.

I agree with most of the articles in which the answers were adopted, but I will explain more about UTXO.

From my point of view, the CTO of R3, the biggest difference between blockchain platforms is what "replicated virtual machines" (which is a metaphor created by the writer) they use, or UTXO models based on transaction results.

Mike Hearn mentioned this while writing about Corda's design: https://www.corda.net/2016/12/09/rationale-tradeoffs-adopting-utxo-style-model/

The key to what I wanted to create is the UTXO model - at least as implemented in Corda, which allows all function codes to be turned without limiting any operations.

The difference is more pronounced in the context of the non-functional nature of the platform.

As the adopted respondent said, the platform should be tailored to the use case: the UTXO model we used in Corda would be well suited to a variety of financial services (because we designed it like that!). But there's no platform that fits everywhere.