\documentclass{beamer} \usepackage{xcolor} \definecolor{primary}{rgb}{0.125, 0.424, 0.243} \usepackage[default]{opensans} \usepackage{graphicx} \usepackage{textcomp} \title{Blocktree} \subtitle{A platform for distributed computing.} \author{Matt Carr \\ matt@carrfound.org} \begin{document} \frame{\titlepage} \begin{frame} \frametitle{Problem} Distributed systems are hard. \begin{itemize} \item<2-> High performance messaging is often outsourced (RabbitMQ, Amazon SQS). \item<3-> Scaling stateful systems continues to be a challenge. \item<4-> The speed of light is a fundamental limit on latency which forces us to move data closer to the user. \end{itemize} \end{frame} \begin{frame}{} \frametitle{Proposed Solution} Reduce latency and the scope of the problem by moving data onto a user's own devices. \begin{itemize} \item<2-> Create a high performance message passing API. Zero copy locally, with transparent forwarding over the network. \item<3-> Create an end-to-end encrypted distributed file system which is served by a user's devices. \item<4-> Allow user's to securely share files using cryptographic mechanisms. \item<5-> Allow a WebAssembly runtime and native containers to access this API. \end{itemize} \end{frame} \begin{frame} \frametitle{User Value Proposition} There are many reasons why user's might prefer this to the status quo. \begin{itemize} \item<2-> Control over personal data. \item<3-> Privacy provided by cryptography, not legislation. \item<4-> Freedom from manipulative advertising. \end{itemize} \end{frame} \begin{frame} \frametitle{The Core Idea} \begin{itemize} \item<2-> All data and running programs are identified by paths in a single global namespace. \begin{center} \scriptsize{/0!dSip4J0kurN5VhVo\_aTipM-ywOOWrqJuRRVQ7aa-bew/docs/otf\_app.ods} \end{center} \item<3-> The first component of the path specifies a signing key which is the root of trust for for all paths under it. \item<4-> Messages are only trusted if they're signed by a key which is transitively trusted by the root. \item<5-> Anyone can verify a file by checking the digital signature it bears is transitively trusted. \end{itemize} \end{frame} \begin{frame} \frametitle{OSS Alternatives} There are many other decentralized systems out there. Blocktree is different in several ways. \begin{itemize} \item<2-> It avoids the scaling issues of blockchains by limiting the scope of consensus (Etherium). \item<3-> It offers mutable storage, which is what real-world applications need. This is difficult with content addressing systems (IPFS, WNFS). \item<4-> Approaches which standardize the format of user data (Solid Pods) fail to address the question of who will store it. \end{itemize} \end{frame} \begin{frame} \frametitle{Current Status and Use of Funds} \begin{itemize} \item<3-> The metadata handling, encryption, and integrity protection mechanisms have been implemented. \item<4-> Support for securely using keys in a TPM is supported. \item<5-> A FUSE daemon which allows the file system to be mounted under Linux has been written. \item<6-> Funds from the OTF would allow me to continue working on the project full-time and allow me to hire additional developers. \end{itemize} \end{frame} \end{document}