In OSDI 2016, Google published a paper describing Slicer, a system to make it easy to build stateful applications by solving the distributed problems associated with sharding applications.

https://www.usenix.org/system/files/conference/osdi16/osdi16-adya.pdf

My current team at Square is building a (simplified) clone of Slicer to simplify feature development, improve load balancing, and reduce latency by relying on soft local state. Examples of our use cases include replacing Redis (for Rpc response caching) with application memory and allowing chunked image uploads to be coalesced in application memory. Since Square's tech stack is built on gRPC, HTTP, and Envoy, our clone will be built on these open source technologies.

We hypothesize that such a clone might be useful outside of Square. However, we want to validate this hypothesis and understand the use cases, so that we can build a system that might be useful to the open source community.

I ask the Hacker News community: 1. Would the systems you are building be able to take advantage of an open source clone of Slicer built on Envoy with support for gRPC and HTTP? 2. If so, please name your company, a brief description of your system and use cases. Additionally, what minimal subset of Slicer's features are sufficient for your use case?