Abstract

FireBox is a new project at UC Berkeley proposing a new system architecture for third-generation Warehouse-Scale Computers (WSCs). We envision 2020 WSCs to be composed of multiple “FireBoxes”, a basic building block containing a thousand compute sockets and 100 Petabytes of non-volatile memory connected via a low-latency, high-bandwidth optical switch. FireBoxes are connected to each other, peripherals and the outside world through a WSC -level network, to form a 1MW WSC with a million cores and an exabyte of non-volatile storage.

Within a FireBox, each compute socket contains a System-on-a-Chip (SoC) with around 100 cores connected to high-bandwidth on-package DRAM . Fast SoC network interfaces reduce the software overhead of communicating between application services, and high-radix network backplane switches connected by Terabit/sec optical fibers reduce the network’s contribution to tail latency. The very large non-volatile store directly supports in-memory databases, and pervasive encryption ensures that data is always protected in transit and in storage.

In this talk, I will present our initial ideas for the FireBox software stack. Our goal is to enable FireBox-based WSCs to efficiently execute both latency-sensitive and batch workloads, written by both efficiency and productivity programmers. In doing so, we want to exploit FireBox’s novel design to address today’s and tomorrow’s challenges in cloud computing, specifically how to efficiently share a FireBox between multiple customers, how to support interactive jobs while controlling tail-latencies, and how to improve hardware utilization. At the same time, we need to solve new challenges, such as how to manage FireBox’s non-volatile memory or dealing with its massive scale.

These ideas are early days, and we are hoping for feedback to further refine them. The talk should therefore be very informal and interactive.

Bio: Martin is a third-year graduate student at the Computer Science department at UC Berkeley. He is working with Krste Asanović and John Kubiatowicz. His main research interests are in managed language runtimes and operating systems. Currently, he is doing an internship with Tim Harris in Cambridge.