A tutorial, Server Benchmarking with CloudSuite 3.0, will be presented on January 20th at HiPEAC 2016.

  • Where? In conjunction with HiPEAC in Prague, Czech Republic.
  • When? January 20th, 2016.
  • Intended audience: Academic/industrial researchers interested in datacenter workloads and their performance evaluation via both existing state- of-the-art servers and cycle-accurate simulation.
  • Team: Alexandros Daglis, Mario Paulo Drumond, Javier Picorel, Babak Falsafi
  • Keywords: Scale-out workloads, server benchmarking, rigorous measurement methodologies, performance evaluation.
  • Registration: Please follow the registration link on the conference web page.
  • Slides: Download

Tutorial In Brief

Since its inception, CloudSuite (cloudsuite.ch) has emerged as a popular suite of benchmarks both in industry and among academics for the performance evaluation of cloud services. The EuroCloud Server project blueprinted key optimizations in server SoCs based on the salient features of CloudSuite workloads that lead to an order of magnitude improvement in efficiency while preserving QoS. ARM-based server products (e.g., Cavium ThunderX) have now emerged following these guidelines and showcasing the improved efficiency.

CloudSuite 3.0 is a major enhancement over prior releases both in workloads and infrastructure. It includes benchmarks that represent massive data manipulation with tight latency constraints such as in-memory data analytics using Apache Spark, a new real-time video streaming benchmark following today’s most popular video-sharing website setups, and a new web serving benchmark mirroring today’s multi-tier web server software stacks with a caching layer.

To ease the deployment of CloudSuite into private and public cloud systems, the benchmarks are integrated into the Docker container system and Google’s PerfKit Benchmarker. PerfKit helps at automating the process of benchmarking with a performance comparison across existing cloud server systems. CloudSuite 3.0 will be released and supported to run both on real hardware and a QEMU-based emulation platform.

Organizers

  • Alexandros Daglis is a fourth year PhD student that works at EPFL under the supervision of Babak Falsafi. His research interests include rack-scale computing and datacenter architectures. His current focus is on system design for high performance remote memory access.
  • Javier Picorel is a fifth year PhD student working at EPFL under the supervision of Babak Falsafi. His research interests are in computer architecture, especially architectures and system support for processing-in-memory systems.
  • Mario Drumond is a second year PhD student that works at EPFL under the supervision of Babak Falsafi. His research interests include reconfigurable computing and computer architecture. His current focus is on using FPGAs to acelerate machine learning algorithms.

Agenda

Time Slot Topic Material
14:00 – 15:00 CloudSuite 3.0 Overview of CloudSuite 3.0
15:00 – 15:15 Coffee break
15:15 – 16:15 CloudSuite 3.0 on Real Hardware (Docker) Running CloudSuite on real hardware
16:15 – 16:30 Coffee break
16:30 – 17:30 CloudSuite 3.0 Full-System Simulation Introduction to Flexus and its interaction with the QEMU full-system simulator