In the rapidly evolving landscape of Web3 solutions, Tashi has emerged as a pioneer in leaderless consensus algorithms. We had the privilege of sitting down with Tashi’s Chief Technology Officer, Ken Anderson, who shared insights into the company’s vision, technological innovations, and its versatile applications in both Web2 and Web3 spaces.

How did the vision for Tashi come about in the first place?

Ken: Tashi’s founders recognized the potential of leaderless consensus algorithms as the fundamental building block for practical and viable protocols in the Web2 space. Plenty of projects had shown feasibility but viability was still lacking; meaning that plenty of projects showed that we could decentralize certain business processes, but few, if any, have shown to do it in a way that is cost effective or fast enough for mainstream use cases. We aimed to bridge that gap by providing performant solutions without compromising the fairness and democratization principles of Web3.

What sets Tashi apart from other paradigms in the industry?

Ken: Tashi embodies the essence of Web3 while meeting the speed and resource requirements of most Web2 use cases. Tashi Consensus Engine (TCE) is the core of our solution. It fundamentally allows multiple parties to submit messages or transactions as fast as possible and results in a fair ordering of those messages on which all parties can act. We proved this by building what we call “Tashi Network Transport” (TNT), a plugin for Unity’s gaming engine that allowed game developers to build multiplayer games without servers or leaders (hosts). TNT demonstrated that we could accomplish this with incredibly low latency, as low as 29 ms, making it compatible with most types of multiplayer games. After we built our first very hacked-together multiplayer game using TNT we got very excited and socialized our discovery with a number of the top game development studios in the world. Their responses were encouraging, but we also knew that there was so much potential for TCE outside of gaming. We’ve explored use cases including a decentralized message queue, decentralized file sharing, decentralized databases, etc.

What is the strategy behind mainstreaming Tashi for users?

Ken: We’ve followed a strategic three-phase approach. First, we built Tashi as a consensus engine. We bundled TCE as a Rust library and have multi-language support on the horizon. This would allow other companies to use TCE in their own projects, benefiting from our progress.

Second, we are building a handful of stand-alone projects that use TCE under the hood. We built TNT for the gaming industry and we are finalizing a decentralized message queue. We call these apps “MeshApps”. These apps can be used off the shelf as ad hoc private multiparty applications. The idea is that you can connect with any set of peers all running the same MeshApp, talk only to your local installation of your MeshApp and everyone ends up with the same state. The message queue is a great example of this. Each peer submits their client messages to their local instance, messages from all of the peers are processed with TCE, resulting in a single ordered queue of messages which can be read locally and operated on. There is no centralized message queue, infrastructure, or controllers of the network. Each peer reads and writes locally but gets the benefit of fair ordering of messages across the entire network of peers.

Third, we are building a platform that allows users to a) manage peer connections, b) initialize a MeshApp between those managed peers, and c) act as a marketplace for MeshApps. Think of the Tashi platform as a type of launcher, much like how Steam or Battle.net are for game and community management. With this platform, Tashi can open up MeshApp development to third parties to solve plenty of use cases we aren’t even thinking about. These could be enterprise solutions, games, or web apps to name a few.

Hear more from Ken at DevComm 2023 on YouTube!