Bring on the Stateful Internet
Why I’m for Ethereum, Hyperledger and so much more.
Because I was part of the team that conceived of and built Hyperledger Fabric, and now I build things primarily on Ethereum, people ask me:
Are you a Hyperledger guy or an Ethereum guy?
Here’s my answer:
I’m a Stateful Internet guy.
I’m working at ConsenSys today because I believe Ethereum has the ecosystem, the historical and technical foundations, and the innovative staying power to be a good candidate for the role of “root chain” in what helps me to think of as the Stateful Internet:
An Internet that can handle data, and rules for changing the state of that data, in the way the Internet handles messaging today —more or less decentralized — AND that allows private stateful networks to interoperate with each other and the public network through standard protocols.
But I also believe that the engineering patterns that went into Hyperledger Fabric channels — as well as the ideas being cooked up in hundreds of other chain schemes like Tendermint, Coda, HashGraph, even Tangle — could lead to better thinking about implementing the billions of peer-to-peer state engines that one day could use the Ethereum main-net as the root chain to prevent shenanigans and coordinate code execution, timing and parameter passing between them.
Chain Freedom Needs an Umpire
I think I once wrote the line IBM’s Jerry Cuomo often uses when asked which blockchain platform will win: “There won’t be one chain to rule them all.” But now, I believe that in order for the Stateful Internet to avoid descending into a fight at the gates of Mordor, we do in fact need a trustless, permissionless, decentralized root chain playing umpire. Only then will each innovative protocol, platform and network be able to do its thing for its users’ special requirements while still being able to interop with the larger ecosystem.
The Stateful Internet Delivers The Two Sides of Security
In blockchain circles, we talk about security a lot. The public network crowd extols the records of Bitcoin and Ethereum for having never been “hacked” (meaning specifically that nobody has successfully tampered with the integrity of the ledger without consensus…big fork events notwithstanding). The private network crowd is all about making sure that private data and agreements between known parties stay confidential.
But security is about BOTH things:
- anti-tampering
- anti-surveillance
Public main-nets are good at #1 and for obvious reasons bad at #2. Private databases are good at #2, bad at #1. Private blockchains aren’t very good at either one: 10 parties maintaining a distributed ledger on a permissioned network can easily collude and tamper with history or censor an otherwise valid transaction, but now you have 10 mirrors to hack into and see the data…10x more threat surface than a conventional Oracle database. Better hope every one of the IT managers in your consortium is as good as the best one!
But what if you could get the anti-surveillance power of a private network that only you and your counterparties ever see but have it secured against tampering and collusion (without leaking private information) by a strong global pubic main-net? And what if you could use that same root network to help avoid some of the hazards of coordinating logic and data across multiple state machines (e.g., race conditions, deadlocks, non-determinism)?
So that’s the “what” but where’s the “how”?
There are a variety of plans for cross-chain interoperability. Some, to me, are reminiscent of Hypercard before HTML, kind of a sovereign approach. Then there are “Layer 2” protocols like Plasma. Lots of stuff is emerging, which is exciting. The one that excites me right now is still just a paper by a brilliant researcher in Australia, Peter Robinson: https://arxiv.org/abs/1806.09834. Check it out, but put on your reading glasses…it’s dense but worth it.
Whatever the solution, you’ll know we’re getting there when you don’t need a different “client” or node or peer or other runtime for every n-lateral network/protocol/platform/ledger (maybe we should just call these “Channels”) you participate in. Imagine having to spin up a different Apache or NGINX Web Server every time you host a different web site! I’m hoping that PegaSys’s Pantheon project will go in this direction.
Connect the Patterns, Forget the Flags
I wish we could take all the good work out there — the patterns each team in the blockchain space has explored for the past several years — and lop off all the brands, the flags, the preciousness we all get when looking at our own babies. We would see it all as a bag full of Legos, a set of potential standards converging on what we really need in order to build awesome new applications that transcend the limitations and troubling central control issues of client/server: #statefulInternet.
John Wolpert is at ConsenSys building a team of top developers, designers and product people to bootstrap new ventures on Ethereum. Prior to joining ConsenSys, John was the global head of blockchain products at IBM and one the founders of Hyperledger Fabric. If you have a new venture you want to talk about, message @jwolpert on Twitter.
Disclaimer: The views expressed by the author above do not necessarily represent the views of Consensys AG. ConsenSys is a decentralized community with ConsenSys Media being a platform for members to freely express their diverse ideas and perspectives. To learn more about ConsenSys and Ethereum, please visit our website.
