We're very transparent, so:
It's not a TM scaling problem. It's a problem with the characteristics of scaled TM for our specific implementation. Maybe. I'm doing a redraft with HL and the first stage of that is simply going to be learning about he comparative performance characteristics of each. We're still reviewing this situation, but basically it does look like we are deciding between TM and HL. We do not have a conventional setup in the least bit, more details of which I'll be releasing soon. I think one significant detail of this that I can share right now is that we'll be using this implementation of google's spanner/F1:
https://github.com/pingcap/tikvSo basically, the KV store is validated by hashes in the blockchain database, but actual content in the KV store is kept off-chain.
Please, keep the questions coming! I'm very happy to ensure that the community has a very clear understanding of the solution we're going to offer.
Why Tendermint?
+ Amazing codebase that makes blockchainizing data structures shockingly easy
+ Friendly and helpful community
+ Awesome founders who both made the majority of code commits to the project
- The extent of the slowdown as it scales, and uncertainty surrounding node counts such as 2048.
Why Hyperledger?
+ Linux foundation support
+ More diverse dvelopment pool VS TM
- Seems more rigid than TM
- Will certainly require more configuration and it's uncertain it will be able to "eat" arbitrary applications written in any programming language to conform to a certain API.
Thanks again
-Jake