ReFlex: Remote Flash Local Flash

Size: px
Start display at page:

Download "ReFlex: Remote Flash Local Flash"

Transcription

1 ReFlex: Remote Flash Local Flash Ana Klimovic Heiner Litz Christos Kozyrakis October 28, 216 IAP Cloud Workshop

2 Flash in Datacenters Flash provides 1 higher throughput and 2 lower latency than disk PCIe Flash: 1,, IOPS 1s of µs latency Flash is underunlized due to imbalanced resource requirements 2

3 Datacenter Flash Use-Case Applica(on Tier Datastore Tier Datastore Service App Tier App Clients Servers get(k) put(k,val) TCP/IP NIC Key-Value Store CPU RAM So9ware Hardware Flash 3

4 Imbalanced Resource UNlizaNon Sample unlizanon of Facebook servers hosnng a Flashbased key-value store over 6 months [EuroSys 16] [EuroSys 16] Flash storage disaggrega.on. Klimovic, A., Kozyrakis, C., Thereska, E., John, B., Kumar, S. 4

5 Imbalanced Resource UNlizaNon Sample unlizanon of Facebook servers hosnng a Flashbased key-value store over 6 months [EuroSys 16] [EuroSys 16] Flash storage disaggrega.on. Klimovic, A., Kozyrakis, C., Thereska, E., John, B., Kumar, S. 5

6 Imbalanced Resource UNlizaNon Sample unlizanon of Facebook servers hosnng a Flashbased key-value store over 6 months [EuroSys 16] unlizanon [EuroSys 16] Flash storage disaggrega.on. Klimovic, A., Kozyrakis, C., Thereska, E., John, B., Kumar, S. 6

7 Imbalanced Resource UNlizaNon Flash capacity and IOPS are underunlized for long periods of Nme unlizanon [EuroSys 16] Flash storage disaggrega.on. Klimovic, A., Kozyrakis, C., Thereska, E., John, B., Kumar, S. 7

8 Local Flash Architecture Applica(on Tier Datastore Tier Datastore Service App Tier App Clients Servers get(k) put(k,val) TCP/IP NIC Key-Value Store CPU RAM So9ware Hardware Flash Provision Flash and CPU in a dependent manner. 8

9 Disaggregated Flash Architecture Applica(on Tier Datastore Tier Datastore Service App Tier App Clients Servers get(k) put(k,val) TCP/IP NIC Key-Value Store CPU RAM So5ware Hardware read(blk); write(blk,data) Protocol iscsi Flash Tier Remote Block Service So5ware CPU NIC RAM Flash Hardware 9

10 ExisNng Approaches Why not apply methods for remote disk or remote memory to access remote flash? There are 2 main issues: 1. Performance overhead 2. Interference on shared remote flash device 1

11 Issue 1: Performance Overhead 1 4kB random read p95 read latency (us) latency 75% throughput drop Local Flash iscsi (1 core) libaio+libevent (1core) IOPS (Thousands) TradiNonal network storage protocols (e.g. iscsi) and convennonal Linux mechanisms have high overhead 11

12 Issue 2: Interference Latency depends on IOPS load p95 read latency (us) %read 99%read 95%read 9%read 75%read 5%read Total IOPS (Thousands) Writes impact read tail latency Flash read performance degrades with increasing % write To share flash, need performance isolanon mechanisms 12

13 Requirements Low Latency Cost Scalability QoS & Perf IsolaPon TradiNonal I/O protocols (e.g. iscsi) X X Distributed storage systems (e.g. GFS, BigTable) X X RDMA (e.g. NVMe over Fabrics) ~ ~ X ReFlex 13

14 ReFlex A remote flash system that provides remote local flash performance over commodity networks 1. Low latency, high throughput, low compute overhead 2. Enforce tail latency and throughput guarantees for clients sharing flash 14

15 ContribuNons A remote flash system that provides remote local flash performance over commodity networks 1. Efficient dataplane execunon model à Low latency and high throughput at low compute overhead 2. Novel I/O scheduler à Enforce tail latency and throughput guarantees for tenants sharing flash 15

16 ReFlex Design Separate control & data planes: Control plane: Resource management: cores, network, Flash Allocate for IOPS, capacity, and latency requirements Example SLO: 1ms tail read latency at 1K IOPS Dataplane: Integrate storage & network stack Low latency, high throughput, high efficiency Enforce QoS with I/O scheduling 16

17 System architecture Ring 3 Control Plane App libix Guest Ring IX Host Ring Linux kernel Dune RX TX Core Core 17

18 System architecture Ring 3 Control Plane App libix App libix Guest Ring IX ReFlex Host Ring Linux kernel Dune RX TX RX TX SQ CQ Core Core Core 18

19 ExecuNon Model Ring 3 ReFlex Server 3 Event CondiNons libix Batched Syscalls Guest Ring NVMe TCP/IP 2 TCP/IP Scheduler NVMe CQ 1 RX TX SQ 4 19

20 ExecuNon Model Ring 3 Event CondiNons ReFlex Server libix 7 Batched Syscalls Guest Ring NVMe 6 TCP/IP TCP/IP Scheduler NVMe 8 CQ 5 RX TX SQ 2

21 Principle 1: Process to complenon Ring 3 ReFlex Server Event CondiNons libix Batched Syscalls Guest Ring NVMe TCP/IP TCP/IP Scheduler NVMe CQ RX TX SQ Improve data-cache locality 21

22 Principle 2: Batch adapnvely Ring 3 ReFlex Server Event CondiNons libix Batched Syscalls Guest Ring NVMe TCP/IP Match batch size to system load TCP/IP Scheduler NVMe CQ RX TX SQ Improve instrucnon-cache locality and prefetching 22

23 Principle 3: Avoid data copies Ring 3 ReFlex Server Event CondiNons libix Batched Syscalls Guest Ring NVMe TCP/IP Forward data directly from NIC à Flash TCP/IP NVMe Scheduler CQ RX TX SQ Reduce latency, cache pollunon, and compute overhead 23

24 Principle 4: Schedule I/O Ring 3 ReFlex Server Event CondiNons libix Batched Syscalls Guest Ring NVMe TCP/IP Control rate at which each tenant issues I/O to shared flash device TCP/IP Scheduler NVMe CQ RX TX SQ Provide Quality of Service (QoS) guarantees 24

25 Request Cost Model Determine relanve I/O cost during calibranon Cost = rela.ve impact on tail latency of a concurrent (4kB) read, in unit of tokens Example: 4kB read cost = 1 token, 4kB write cost = X tokens p95 read latency (us) %read 99%read 95%read 9%read 75%read 5%read Total IOPS (Thousands) Find X that unifies curves for all rd/wr ra.os Scale write IOPS by cost factor X p95 Read Latency (us) %read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) For this device, X=1 Rd IOPS + Wr IOPS Rd IOPS + X * (Wr IOPS) 25

26 Token-based Scheduling p95 Read Latency (us) ms tail latency SLO 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) 26

27 Token-based Scheduling p95 Read Latency (us) ms tail latency SLO Device max IOPS: 485K 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) 27

28 Token-based Scheduling p95 Read Latency (us) ms tail latency SLO Tenant reserves 2K Device max IOPS: 485K 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) 28

29 Token-based Scheduling p95 Read Latency (us) ms tail latency SLO Tenant reserves 2K Device max IOPS: 485K 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) 29

30 Token-based Scheduling 2 p95 Read Latency (us) ms tail latency SLO Tenant reserves 2K Device max IOPS: 485K 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) For latency-cripcal tenants: Generate tokens based on IOPS in SLO Avoid overhead by giving token credit limit Rate limit large bursts that exceed IOPS reservanon Don t allow idle tenants to accumulate tokens 3

31 Token-based Scheduling 2 p95 Read Latency (us) ms tail latency SLO Tenant reserves 2K Device max IOPS: 485K 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) For best-effort tenants: Generate tokens based on ler-over IOPS Schedule only if have enough tokens Round-robin for fairness 31

32 Token-based Scheduling 2 p95 Read Latency (us) ms tail latency SLO Tenant reserves 2K Device max IOPS: 485K 1%read 99%read 95%read 9%read 75%read 5%read Weighted IOPS (x 1 3 tokens/s ) Distributed scheduler: Each core manages tokens for own tenants Coordinate only to share unused tokens Scales well for mulncore 32

33 Results: Local Remote Latency p95 Read Latency (us) Linux: 75K IOPS/core ReFlex: 85K IOPS/core Local-1T ReFlex-1T Libaio-1T IOPS (Thousands) 33

34 Results: Local Remote Latency p95 Read Latency (us) Unloaded latency Local Flash 78 µs ReFlex 99 µs Libaio 121 µs Local-1T ReFlex-1T Libaio-1T IOPS (Thousands) 34

35 Results: Local Remote Latency p95 Read Latency (us) ReFlex saturates Flash device Local-1T Local-2T ReFlex-1T ReFlex-2T Libaio-1T Libaio-2T IOPS (Thousands) 35

36 Results: Performance IsolaNon Read p95 latency (us) I/O sched disabled I/O sched enabled Latency SLO IOPS (Thousands) I/O sched disabled Tenant A IOPS SLO I/O sched enabled Tenant B IOPS SLO Tenant A Tenant B Tenant C Tenant D Tenant A Tenant B Tenant C Tenant D 1%rd 8%rd 95%rd 25%rd 1%rd 8%rd 95%rd 25%rd Tenants A & B: latency-crincal; Tenant C + D: best effort 36

37 Results: Performance IsolaNon Read p95 latency (us) I/O sched disabled I/O sched enabled Latency SLO IOPS (Thousands) I/O sched disabled Tenant A IOPS SLO I/O sched enabled Tenant B IOPS SLO Tenant A Tenant B Tenant C Tenant D Tenant A Tenant B Tenant C Tenant D 1%rd 8%rd 95%rd 25%rd 1%rd 8%rd 95%rd 25%rd Tenants A & B: latency-crincal; Tenant C + D: best effort Without scheduler: latency and bandwidth QoS for A/B are violated 37

38 Results: Performance IsolaNon Read p95 latency (us) I/O sched disabled I/O sched enabled Latency SLO IOPS (Thousands) I/O sched disabled Tenant A IOPS SLO I/O sched enabled Tenant B IOPS SLO Tenant A Tenant B Tenant C Tenant D Tenant A Tenant B Tenant C Tenant D 1%rd 8%rd 95%rd 25%rd 1%rd 8%rd 95%rd 25%rd Tenants A & B: latency-crincal; Tenant C + D: best effort Without scheduler: latency and bandwidth QoS for A/B are violated Scheduler rate limits best-effort tenants to enforce SLOs 38

39 Conclusion ReFlex enables flash disaggreganon: Achieves remote local flash by rethinking the OS & network storage sorware stack Provides QoS on shared flash with I/O scheduler Uses commodity networking, has low CPU overhead Future work: Client interface: transacnons, databases? Control plane: policies for resource management QoS with host-side Flash TranslaNon Layer (FTL) 39

ReFlex: Remote Flash Local Flash

ReFlex: Remote Flash Local Flash ReFlex: Remote Flash Local Flash Ana Klimovic Heiner Litz Christos Kozyrakis NVMW 18 Memorable Paper Award Finalist 1 Flash in Datacenters Flash provides 1000 higher throughput and 100 lower latency than

More information

No Tradeoff Low Latency + High Efficiency

No Tradeoff Low Latency + High Efficiency No Tradeoff Low Latency + High Efficiency Christos Kozyrakis http://mast.stanford.edu Latency-critical Applications A growing class of online workloads Search, social networking, software-as-service (SaaS),

More information

Pocket: Elastic Ephemeral Storage for Serverless Analytics

Pocket: Elastic Ephemeral Storage for Serverless Analytics Pocket: Elastic Ephemeral Storage for Serverless Analytics Ana Klimovic*, Yawen Wang*, Patrick Stuedi +, Animesh Trivedi +, Jonas Pfefferle +, Christos Kozyrakis* *Stanford University, + IBM Research 1

More information

IX: A Protected Dataplane Operating System for High Throughput and Low Latency

IX: A Protected Dataplane Operating System for High Throughput and Low Latency IX: A Protected Dataplane Operating System for High Throughput and Low Latency Adam Belay et al. Proc. of the 11th USENIX Symp. on OSDI, pp. 49-65, 2014. Presented by Han Zhang & Zaina Hamid Challenges

More information

2017 Storage Developer Conference. Mellanox Technologies. All Rights Reserved.

2017 Storage Developer Conference. Mellanox Technologies. All Rights Reserved. Ethernet Storage Fabrics Using RDMA with Fast NVMe-oF Storage to Reduce Latency and Improve Efficiency Kevin Deierling & Idan Burstein Mellanox Technologies 1 Storage Media Technology Storage Media Access

More information

Kernel Bypass. Sujay Jayakar (dsj36) 11/17/2016

Kernel Bypass. Sujay Jayakar (dsj36) 11/17/2016 Kernel Bypass Sujay Jayakar (dsj36) 11/17/2016 Kernel Bypass Background Why networking? Status quo: Linux Papers Arrakis: The Operating System is the Control Plane. Simon Peter, Jialin Li, Irene Zhang,

More information

Ziye Yang. NPG, DCG, Intel

Ziye Yang. NPG, DCG, Intel Ziye Yang NPG, DCG, Intel Agenda What is SPDK? Accelerated NVMe-oF via SPDK Conclusion 2 Agenda What is SPDK? Accelerated NVMe-oF via SPDK Conclusion 3 Storage Performance Development Kit Scalable and

More information

NFS/RDMA over 40Gbps iwarp Wael Noureddine Chelsio Communications

NFS/RDMA over 40Gbps iwarp Wael Noureddine Chelsio Communications NFS/RDMA over 40Gbps iwarp Wael Noureddine Chelsio Communications Outline RDMA Motivating trends iwarp NFS over RDMA Overview Chelsio T5 support Performance results 2 Adoption Rate of 40GbE Source: Crehan

More information

RDMA and Hardware Support

RDMA and Hardware Support RDMA and Hardware Support SIGCOMM Topic Preview 2018 Yibo Zhu Microsoft Research 1 The (Traditional) Journey of Data How app developers see the network Under the hood This architecture had been working

More information

Low-Overhead Flash Disaggregation via NVMe-over-Fabrics Vijay Balakrishnan Memory Solutions Lab. Samsung Semiconductor, Inc.

Low-Overhead Flash Disaggregation via NVMe-over-Fabrics Vijay Balakrishnan Memory Solutions Lab. Samsung Semiconductor, Inc. Low-Overhead Flash Disaggregation via NVMe-over-Fabrics Vijay Balakrishnan Memory Solutions Lab. Samsung Semiconductor, Inc. 1 DISCLAIMER This presentation and/or accompanying oral statements by Samsung

More information

Low-Overhead Flash Disaggregation via NVMe-over-Fabrics

Low-Overhead Flash Disaggregation via NVMe-over-Fabrics Low-Overhead Flash Disaggregation via NVMe-over-Fabrics Vijay Balakrishnan Memory Solutions Lab. Samsung Semiconductor, Inc. August 2017 1 DISCLAIMER This presentation and/or accompanying oral statements

More information

Low-Latency Datacenters. John Ousterhout

Low-Latency Datacenters. John Ousterhout Low-Latency Datacenters John Ousterhout The Datacenter Revolution Phase 1: Scale How to use 10,000 servers for a single application? New storage systems: Bigtable, HDFS,... New models of computation: MapReduce,

More information

THE STORAGE PERFORMANCE DEVELOPMENT KIT AND NVME-OF

THE STORAGE PERFORMANCE DEVELOPMENT KIT AND NVME-OF 14th ANNUAL WORKSHOP 2018 THE STORAGE PERFORMANCE DEVELOPMENT KIT AND NVME-OF Paul Luse Intel Corporation Apr 2018 AGENDA Storage Performance Development Kit What is SPDK? The SPDK Community Why are so

More information

IX: A Protected Dataplane Operating System for High Throughput and Low Latency

IX: A Protected Dataplane Operating System for High Throughput and Low Latency IX: A Protected Dataplane Operating System for High Throughput and Low Latency Adam Belay 1 George Prekas 2 Ana Klimovic 1 Samuel Grossman 1 Christos Kozyrakis 1 1 Stanford University Edouard Bugnion 2

More information

Storage Systems for Serverless Analytics

Storage Systems for Serverless Analytics Storage Systems for Serverless Analytics Ana Klimovic * Yawen Wang * Christos Kozyrakis * Patrick Stuedi ⱡ Jonas Pfefferle ⱡ Animesh Trivedi ⱡ * ⱡ Serverless: a new cloud computing paradigm Users write

More information

FlashShare: Punching Through Server Storage Stack from Kernel to Firmware for Ultra-Low Latency SSDs

FlashShare: Punching Through Server Storage Stack from Kernel to Firmware for Ultra-Low Latency SSDs FlashShare: Punching Through Server Storage Stack from Kernel to Firmware for Ultra-Low Latency SSDs Jie Zhang, Miryeong Kwon, Donghyun Gouk, Sungjoon Koh, Changlim Lee, Mohammad Alian, Myoungjun Chun,

More information

IX: A Protected Dataplane Operating System for High Throughput and Low Latency

IX: A Protected Dataplane Operating System for High Throughput and Low Latency IX: A Protected Dataplane Operating System for High Throughput and Low Latency Belay, A. et al. Proc. of the 11th USENIX Symp. on OSDI, pp. 49-65, 2014. Reviewed by Chun-Yu and Xinghao Li Summary In this

More information

Storage Protocol Offload for Virtualized Environments Session 301-F

Storage Protocol Offload for Virtualized Environments Session 301-F Storage Protocol Offload for Virtualized Environments Session 301-F Dennis Martin, President August 2016 1 Agenda About Demartek Offloads I/O Virtualization Concepts RDMA Concepts Overlay Networks and

More information

Fast packet processing in the cloud. Dániel Géhberger Ericsson Research

Fast packet processing in the cloud. Dániel Géhberger Ericsson Research Fast packet processing in the cloud Dániel Géhberger Ericsson Research Outline Motivation Service chains Hardware related topics, acceleration Virtualization basics Software performance and acceleration

More information

I N V E N T I V E. SSD Firmware Complexities and Benefits from NVMe. Steven Shrader

I N V E N T I V E. SSD Firmware Complexities and Benefits from NVMe. Steven Shrader I N V E N T I V E SSD Firmware Complexities and Benefits from NVMe Steven Shrader Agenda Introduction NVMe architectural issues from NVMe functions Structures to model the problem Methods (metadata attributes)

More information

NVM Express over Fabrics Storage Solutions for Real-time Analytics

NVM Express over Fabrics Storage Solutions for Real-time Analytics NVM Express over Fabrics Storage Solutions for Real-time Analytics Presented by Paul Prince, CTO Santa Clara, CA 1 NVMe Over Fabrics NVMf Why do we need NVMf? What is it? How does it fit in the Market?

More information

ibench: Quantifying Interference in Datacenter Applications

ibench: Quantifying Interference in Datacenter Applications ibench: Quantifying Interference in Datacenter Applications Christina Delimitrou and Christos Kozyrakis Stanford University IISWC September 23 th 2013 Executive Summary Problem: Increasing utilization

More information

EXPERIENCES WITH NVME OVER FABRICS

EXPERIENCES WITH NVME OVER FABRICS 13th ANNUAL WORKSHOP 2017 EXPERIENCES WITH NVME OVER FABRICS Parav Pandit, Oren Duer, Max Gurtovoy Mellanox Technologies [ 31 March, 2017 ] BACKGROUND: NVME TECHNOLOGY Optimized for flash and next-gen

More information

W H I T E P A P E R. What s New in VMware vsphere 4: Performance Enhancements

W H I T E P A P E R. What s New in VMware vsphere 4: Performance Enhancements W H I T E P A P E R What s New in VMware vsphere 4: Performance Enhancements Scalability Enhancements...................................................... 3 CPU Enhancements............................................................

More information

Network Requirements for Resource Disaggregation

Network Requirements for Resource Disaggregation Network Requirements for Resource Disaggregation Peter Gao (Berkeley), Akshay Narayan (MIT), Sagar Karandikar (Berkeley), Joao Carreira (Berkeley), Sangjin Han (Berkeley), Rachit Agarwal (Cornell), Sylvia

More information

NVMe Direct. Next-Generation Offload Technology. White Paper

NVMe Direct. Next-Generation Offload Technology. White Paper NVMe Direct Next-Generation Offload Technology The market introduction of high-speed NVMe SSDs and 25/40/50/100Gb Ethernet creates exciting new opportunities for external storage NVMe Direct enables high-performance

More information

At the heart of a new generation of data center infrastructures and appliances. Sept 2017

At the heart of a new generation of data center infrastructures and appliances. Sept 2017 / At the heart of a new generation of data center infrastructures and appliances Sept 2017 VIRTUALIZED DATACENTER: THE BLENDER EFFECT FOR STORAGE I/O OPERATIONS 10, 000 VMs MIOPs in random Aggregate switch

More information

SPDK China Summit Ziye Yang. Senior Software Engineer. Network Platforms Group, Intel Corporation

SPDK China Summit Ziye Yang. Senior Software Engineer. Network Platforms Group, Intel Corporation SPDK China Summit 2018 Ziye Yang Senior Software Engineer Network Platforms Group, Intel Corporation Agenda SPDK programming framework Accelerated NVMe-oF via SPDK Conclusion 2 Agenda SPDK programming

More information

SELF-LEARNING CACHES IRFAN AHMAD CACHEPHYSICS. Copyright 2017 CachePhysics.

SELF-LEARNING CACHES IRFAN AHMAD CACHEPHYSICS. Copyright 2017 CachePhysics. SELF-LEARNING CACHES IRFAN AHMAD CACHEPHYSICS Copyright 217 CachePhysics. ABOUT CachePhysics Irfan Ahmad CachePhysics Cofounder CloudPhysics Cofounder VMware (Kernel, Resource Management), Transmeta, 3+

More information

NAS for Server Virtualization Dennis Chapman Senior Technical Director NetApp

NAS for Server Virtualization Dennis Chapman Senior Technical Director NetApp NAS for Server Virtualization Dennis Chapman Senior Technical Director NetApp Agenda The Landscape has Changed New Customer Requirements The Market has Begun to Move Comparing Performance Results Storage

More information

WORKLOAD CHARACTERIZATION OF INTERACTIVE CLOUD SERVICES BIG AND SMALL SERVER PLATFORMS

WORKLOAD CHARACTERIZATION OF INTERACTIVE CLOUD SERVICES BIG AND SMALL SERVER PLATFORMS WORKLOAD CHARACTERIZATION OF INTERACTIVE CLOUD SERVICES ON BIG AND SMALL SERVER PLATFORMS Shuang Chen*, Shay Galon**, Christina Delimitrou*, Srilatha Manne**, and José Martínez* *Cornell University **Cavium

More information

LegUp: Accelerating Memcached on Cloud FPGAs

LegUp: Accelerating Memcached on Cloud FPGAs 0 LegUp: Accelerating Memcached on Cloud FPGAs Xilinx Developer Forum December 10, 2018 Andrew Canis & Ruolong Lian LegUp Computing Inc. 1 COMPUTE IS BECOMING SPECIALIZED 1 GPU Nvidia graphics cards are

More information

Fusion Engine Next generation storage engine for Flash- SSD and 3D XPoint storage system

Fusion Engine Next generation storage engine for Flash- SSD and 3D XPoint storage system Fusion Engine Next generation storage engine for Flash- SSD and 3D XPoint storage system Fei Liu, Sheng Qiu, Jianjian Huo, Shu Li Alibaba Group Santa Clara, CA 1 Software overhead become critical Legacy

More information

Why NVMe/TCP is the better choice for your Data Center

Why NVMe/TCP is the better choice for your Data Center Why NVMe/TCP is the better choice for your Data Center Non-Volatile Memory express (NVMe) has transformed the storage industry since its emergence as the state-of-the-art protocol for high-performance

More information

NVMe over Fabrics. High Performance SSDs networked over Ethernet. Rob Davis Vice President Storage Technology, Mellanox

NVMe over Fabrics. High Performance SSDs networked over Ethernet. Rob Davis Vice President Storage Technology, Mellanox NVMe over Fabrics High Performance SSDs networked over Ethernet Rob Davis Vice President Storage Technology, Mellanox Ilker Cebeli Senior Director of Product Planning, Samsung May 3, 2017 Storage Performance

More information

Reducing CPU and network overhead for small I/O requests in network storage protocols over raw Ethernet

Reducing CPU and network overhead for small I/O requests in network storage protocols over raw Ethernet Reducing CPU and network overhead for small I/O requests in network storage protocols over raw Ethernet Pilar González-Férez and Angelos Bilas 31 th International Conference on Massive Storage Systems

More information

SoftRDMA: Rekindling High Performance Software RDMA over Commodity Ethernet

SoftRDMA: Rekindling High Performance Software RDMA over Commodity Ethernet SoftRDMA: Rekindling High Performance Software RDMA over Commodity Ethernet Mao Miao, Fengyuan Ren, Xiaohui Luo, Jing Xie, Qingkai Meng, Wenxue Cheng Dept. of Computer Science and Technology, Tsinghua

More information

Scaling Internet TV Content Delivery ALEX GUTARIN DIRECTOR OF ENGINEERING, NETFLIX

Scaling Internet TV Content Delivery ALEX GUTARIN DIRECTOR OF ENGINEERING, NETFLIX Scaling Internet TV Content Delivery ALEX GUTARIN DIRECTOR OF ENGINEERING, NETFLIX Inventing Internet TV Available in more than 190 countries 104+ million subscribers Lots of Streaming == Lots of Traffic

More information

Farewell to Servers: Hardware, Software, and Network Approaches towards Datacenter Resource Disaggregation

Farewell to Servers: Hardware, Software, and Network Approaches towards Datacenter Resource Disaggregation Farewell to Servers: Hardware, Software, and Network Approaches towards Datacenter Resource Disaggregation Yiying Zhang Datacenter 3 Monolithic Computer OS / Hypervisor 4 Can monolithic Application Hardware

More information

École Polytechnique Fédérale de Lausanne. Porting a driver for the Intel XL710 40GbE NIC to the IX Dataplane Operating System

École Polytechnique Fédérale de Lausanne. Porting a driver for the Intel XL710 40GbE NIC to the IX Dataplane Operating System École Polytechnique Fédérale de Lausanne Semester Project Porting a driver for the Intel XL710 40GbE NIC to the IX Dataplane Operating System Student: Andy Roulin (216690) Direct Supervisor: George Prekas

More information

RAMCloud and the Low- Latency Datacenter. John Ousterhout Stanford University

RAMCloud and the Low- Latency Datacenter. John Ousterhout Stanford University RAMCloud and the Low- Latency Datacenter John Ousterhout Stanford University Most important driver for innovation in computer systems: Rise of the datacenter Phase 1: large scale Phase 2: low latency Introduction

More information

Real-Time Internet of Things

Real-Time Internet of Things Real-Time Internet of Things Chenyang Lu Cyber-Physical Systems Laboratory h7p://www.cse.wustl.edu/~lu/ Internet of Things Ø Convergence of q Miniaturized devices: integrate processor, sensors and radios.

More information

InfiniBand Networked Flash Storage

InfiniBand Networked Flash Storage InfiniBand Networked Flash Storage Superior Performance, Efficiency and Scalability Motti Beck Director Enterprise Market Development, Mellanox Technologies Flash Memory Summit 2016 Santa Clara, CA 1 17PB

More information

Fast and Easy Persistent Storage for Docker* Containers with Storidge and Intel

Fast and Easy Persistent Storage for Docker* Containers with Storidge and Intel Solution brief Intel Storage Builders Storidge ContainerIO TM Intel Xeon Processor Scalable Family Intel SSD DC Family for PCIe*/NVMe Fast and Easy Persistent Storage for Docker* Containers with Storidge

More information

Designing Next Generation FS for NVMe and NVMe-oF

Designing Next Generation FS for NVMe and NVMe-oF Designing Next Generation FS for NVMe and NVMe-oF Liran Zvibel CTO, Co-founder Weka.IO @liranzvibel Santa Clara, CA 1 Designing Next Generation FS for NVMe and NVMe-oF Liran Zvibel CTO, Co-founder Weka.IO

More information

Infiniswap. Efficient Memory Disaggregation. Mosharaf Chowdhury. with Juncheng Gu, Youngmoon Lee, Yiwen Zhang, and Kang G. Shin

Infiniswap. Efficient Memory Disaggregation. Mosharaf Chowdhury. with Juncheng Gu, Youngmoon Lee, Yiwen Zhang, and Kang G. Shin Infiniswap Efficient Memory Disaggregation Mosharaf Chowdhury with Juncheng Gu, Youngmoon Lee, Yiwen Zhang, and Kang G. Shin Rack-Scale Computing Datacenter-Scale Computing Geo-Distributed Computing Coflow

More information

Oracle Exadata: Strategy and Roadmap

Oracle Exadata: Strategy and Roadmap Oracle Exadata: Strategy and Roadmap - New Technologies, Cloud, and On-Premises Juan Loaiza Senior Vice President, Database Systems Technologies, Oracle Safe Harbor Statement The following is intended

More information

What s New in VMware vsphere 4.1 Performance. VMware vsphere 4.1

What s New in VMware vsphere 4.1 Performance. VMware vsphere 4.1 What s New in VMware vsphere 4.1 Performance VMware vsphere 4.1 T E C H N I C A L W H I T E P A P E R Table of Contents Scalability enhancements....................................................................

More information

N V M e o v e r F a b r i c s -

N V M e o v e r F a b r i c s - N V M e o v e r F a b r i c s - H i g h p e r f o r m a n c e S S D s n e t w o r k e d f o r c o m p o s a b l e i n f r a s t r u c t u r e Rob Davis, VP Storage Technology, Mellanox OCP Evolution Server

More information

IO virtualization. Michael Kagan Mellanox Technologies

IO virtualization. Michael Kagan Mellanox Technologies IO virtualization Michael Kagan Mellanox Technologies IO Virtualization Mission non-stop s to consumers Flexibility assign IO resources to consumer as needed Agility assignment of IO resources to consumer

More information

Impact of Cache Coherence Protocols on the Processing of Network Traffic

Impact of Cache Coherence Protocols on the Processing of Network Traffic Impact of Cache Coherence Protocols on the Processing of Network Traffic Amit Kumar and Ram Huggahalli Communication Technology Lab Corporate Technology Group Intel Corporation 12/3/2007 Outline Background

More information

MQSim: A Framework for Enabling Realistic Studies of Modern Multi-Queue SSD Devices

MQSim: A Framework for Enabling Realistic Studies of Modern Multi-Queue SSD Devices MQSim: A Framework for Enabling Realistic Studies of Modern Multi-Queue SSD Devices Arash Tavakkol, Juan Gómez-Luna, Mohammad Sadrosadati, Saugata Ghose, Onur Mutlu February 13, 2018 Executive Summary

More information

PRESENTATION TITLE GOES HERE

PRESENTATION TITLE GOES HERE Performance Basics PRESENTATION TITLE GOES HERE Leah Schoeb, Member of SNIA Technical Council SNIA EmeraldTM Training SNIA Emerald Power Efficiency Measurement Specification, for use in EPA ENERGY STAR

More information

Next Generation Architecture for NVM Express SSD

Next Generation Architecture for NVM Express SSD Next Generation Architecture for NVM Express SSD Dan Mahoney CEO Fastor Systems Copyright 2014, PCI-SIG, All Rights Reserved 1 NVMExpress Key Characteristics Highest performance, lowest latency SSD interface

More information

Arrakis: The Operating System is the Control Plane

Arrakis: The Operating System is the Control Plane Arrakis: The Operating System is the Control Plane Simon Peter, Jialin Li, Irene Zhang, Dan Ports, Doug Woos, Arvind Krishnamurthy, Tom Anderson University of Washington Timothy Roscoe ETH Zurich Building

More information

Forget IOPS: A Proper Way to Characterize & Test Storage Performance Peter Murray SwiftTest

Forget IOPS: A Proper Way to Characterize & Test Storage Performance Peter Murray SwiftTest Forget IOPS: A Proper Way to Characterize & Test Storage Performance Peter Murray peter@swifttest.com SwiftTest Storage Performance Validation Rely on vendor IOPS claims Test in production and pray Validate

More information

D E N A L I S T O R A G E I N T E R F A C E. Laura Caulfield Senior Software Engineer. Arie van der Hoeven Principal Program Manager

D E N A L I S T O R A G E I N T E R F A C E. Laura Caulfield Senior Software Engineer. Arie van der Hoeven Principal Program Manager 1 T HE D E N A L I N E X T - G E N E R A T I O N H I G H - D E N S I T Y S T O R A G E I N T E R F A C E Laura Caulfield Senior Software Engineer Arie van der Hoeven Principal Program Manager Outline Technology

More information

Application Acceleration Beyond Flash Storage

Application Acceleration Beyond Flash Storage Application Acceleration Beyond Flash Storage Session 303C Mellanox Technologies Flash Memory Summit July 2014 Accelerating Applications, Step-by-Step First Steps Make compute fast Moore s Law Make storage

More information

Disclaimer This presentation may contain product features that are currently under development. This overview of new technology represents no commitme

Disclaimer This presentation may contain product features that are currently under development. This overview of new technology represents no commitme SER2734BU Extreme Performance Series: Byte-Addressable Nonvolatile Memory in vsphere VMworld 2017 Content: Not for publication Qasim Ali and Praveen Yedlapalli #VMworld #SER2734BU Disclaimer This presentation

More information

Application Advantages of NVMe over Fabrics RDMA and Fibre Channel

Application Advantages of NVMe over Fabrics RDMA and Fibre Channel Application Advantages of NVMe over Fabrics RDMA and Fibre Channel Brandon Hoff Broadcom Limited Tuesday, June 14 2016 10:55 11:35 a.m. Agenda r Applications that have a need for speed r The Benefits of

More information

Speeding up Linux TCP/IP with a Fast Packet I/O Framework

Speeding up Linux TCP/IP with a Fast Packet I/O Framework Speeding up Linux TCP/IP with a Fast Packet I/O Framework Michio Honda Advanced Technology Group, NetApp michio@netapp.com With acknowledge to Kenichi Yasukata, Douglas Santry and Lars Eggert 1 Motivation

More information

Oracle Enterprise Architecture. Software. Hardware. Complete. Oracle Exalogic.

Oracle Enterprise Architecture. Software. Hardware. Complete. Oracle Exalogic. Oracle Enterprise Architecture Software. Hardware. Complete Oracle Exalogic edward.zhang@oracle.com Exalogic Exalogic Exalogic -- Exalogic Design Center Exalogic - Sun Oracle - - - CPU/Memory/Networking/Storage

More information

Modern hyperconverged infrastructure. Karel Rudišar Systems Engineer, Vmware Inc.

Modern hyperconverged infrastructure. Karel Rudišar Systems Engineer, Vmware Inc. Modern hyperconverged infrastructure Karel Rudišar Systems Engineer, Vmware Inc. 2 What Is Hyper-Converged Infrastructure? - The Ideal Architecture for SDDC Management SDDC Compute Networking Storage Simplicity

More information

Announcements. Reading. Project #1 due in 1 week at 5:00 pm Scheduling Chapter 6 (6 th ed) or Chapter 5 (8 th ed) CMSC 412 S14 (lect 5)

Announcements. Reading. Project #1 due in 1 week at 5:00 pm Scheduling Chapter 6 (6 th ed) or Chapter 5 (8 th ed) CMSC 412 S14 (lect 5) Announcements Reading Project #1 due in 1 week at 5:00 pm Scheduling Chapter 6 (6 th ed) or Chapter 5 (8 th ed) 1 Relationship between Kernel mod and User Mode User Process Kernel System Calls User Process

More information

How to Network Flash Storage Efficiently at Hyperscale. Flash Memory Summit 2017 Santa Clara, CA 1

How to Network Flash Storage Efficiently at Hyperscale. Flash Memory Summit 2017 Santa Clara, CA 1 How to Network Flash Storage Efficiently at Hyperscale Manoj Wadekar Michael Kagan Flash Memory Summit 2017 Santa Clara, CA 1 ebay Hyper scale Infrastructure Search Front-End & Product Hadoop Object Store

More information

Networking at the Speed of Light

Networking at the Speed of Light Networking at the Speed of Light Dror Goldenberg VP Software Architecture MaRS Workshop April 2017 Cloud The Software Defined Data Center Resource virtualization Efficient services VM, Containers uservices

More information

Empirical Evaluation of Latency-Sensitive Application Performance in the Cloud

Empirical Evaluation of Latency-Sensitive Application Performance in the Cloud Empirical Evaluation of Latency-Sensitive Application Performance in the Cloud Sean Barker and Prashant Shenoy University of Massachusetts Amherst Department of Computer Science Cloud Computing! Cloud

More information

Network Support for Multimedia

Network Support for Multimedia Network Support for Multimedia Daniel Zappala CS 460 Computer Networking Brigham Young University Network Support for Multimedia 2/33 make the best of best effort use application-level techniques use CDNs

More information

Red Hat Ceph Storage and Samsung NVMe SSDs for intensive workloads

Red Hat Ceph Storage and Samsung NVMe SSDs for intensive workloads Red Hat Ceph Storage and Samsung NVMe SSDs for intensive workloads Power emerging OpenStack use cases with high-performance Samsung/ Red Hat Ceph reference architecture Optimize storage cluster performance

More information

Advanced Computer Networks. End Host Optimization

Advanced Computer Networks. End Host Optimization Oriana Riva, Department of Computer Science ETH Zürich 263 3501 00 End Host Optimization Patrick Stuedi Spring Semester 2017 1 Today End-host optimizations: NUMA-aware networking Kernel-bypass Remote Direct

More information

SENIC: Scalable NIC for End-Host Rate Limiting

SENIC: Scalable NIC for End-Host Rate Limiting : Scalable NIC for End-Host Rate Limiting Sivasankar Radhakrishnan, Yilong Geng +, Vimalkumar Jeyakumar +, Abdul Kabbani, George Porter, Amin Vahdat University of California, San Diego + Stanford University

More information

Implica(ons of Non Vola(le Memory on So5ware Architectures. Nisha Talagala Lead Architect, Fusion- io

Implica(ons of Non Vola(le Memory on So5ware Architectures. Nisha Talagala Lead Architect, Fusion- io Implica(ons of Non Vola(le Memory on So5ware Architectures Nisha Talagala Lead Architect, Fusion- io Overview Non Vola;le Memory Technology NVM in the Datacenter Op;mizing sobware for the iomemory Tier

More information

Data Processing at the Speed of 100 Gbps using Apache Crail. Patrick Stuedi IBM Research

Data Processing at the Speed of 100 Gbps using Apache Crail. Patrick Stuedi IBM Research Data Processing at the Speed of 100 Gbps using Apache Crail Patrick Stuedi IBM Research The CRAIL Project: Overview Data Processing Framework (e.g., Spark, TensorFlow, λ Compute) Spark-IO FS Albis Streaming

More information

Exploring System Challenges of Ultra-Low Latency Solid State Drives

Exploring System Challenges of Ultra-Low Latency Solid State Drives Exploring System Challenges of Ultra-Low Latency Solid State Drives Sungjoon Koh Changrim Lee, Miryeong Kwon, and Myoungsoo Jung Computer Architecture and Memory systems Lab Executive Summary Motivation.

More information

Providing Near-Optimal Fair- Queueing Guarantees at Round-Robin Amortized Cost

Providing Near-Optimal Fair- Queueing Guarantees at Round-Robin Amortized Cost Providing Near-Optimal Fair- Queueing Guarantees at Round-Robin Amortized Cost Paolo Valente Department of Physics, Computer Science and Mathematics Modena - Italy Workshop PRIN SFINGI October 2013 2 Contributions

More information

Building a High IOPS Flash Array: A Software-Defined Approach

Building a High IOPS Flash Array: A Software-Defined Approach Building a High IOPS Flash Array: A Software-Defined Approach Weafon Tsao Ph.D. VP of R&D Division, AccelStor, Inc. Santa Clara, CA Clarification Myth 1: S High-IOPS SSDs = High-IOPS All-Flash Array SSDs

More information

SmartNICs: Giving Rise To Smarter Offload at The Edge and In The Data Center

SmartNICs: Giving Rise To Smarter Offload at The Edge and In The Data Center SmartNICs: Giving Rise To Smarter Offload at The Edge and In The Data Center Jeff Defilippi Senior Product Manager Arm #Arm Tech Symposia The Cloud to Edge Infrastructure Foundation for a World of 1T Intelligent

More information

Enabling NVMe I/O Scale

Enabling NVMe I/O Scale Enabling NVMe I/O Determinism @ Scale Chris Petersen, Hardware System Technologist Wei Zhang, Software Engineer Alexei Naberezhnov, Software Engineer Facebook Facebook @ Scale 800 Million 1.3 Billion 2.2

More information

Jim Harris Principal Software Engineer Intel Data Center Group

Jim Harris Principal Software Engineer Intel Data Center Group Jim Harris Principal Software Engineer Intel Data Center Group Legal Disclaimer INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR

More information

IOFlow. A Software-Defined Storage Architecture

IOFlow. A Software-Defined Storage Architecture IOFlow A Software-Defined Storage Architecture ACM SOSP 2013 Authors: Eno Thereska, Hitesh Ballani, Greg O'Shea, Thomas Karagiannis, Antony Rowstron, Tom Talpey, and Timothy Zhu Presented by yifan wu,

More information

14th ANNUAL WORKSHOP 2018 NVMF TARGET OFFLOAD. Liran Liss. Mellanox Technologies. April 2018

14th ANNUAL WORKSHOP 2018 NVMF TARGET OFFLOAD. Liran Liss. Mellanox Technologies. April 2018 14th ANNUAL WORKSHOP 2018 NVMF TARGET OFFLOAD Liran Liss Mellanox Technologies April 2018 AGENDA Introduction NVMe NVMf NVMf target driver Offload model Verbs interface Status 2 OpenFabrics Alliance Workshop

More information

Extremely Fast Distributed Storage for Cloud Service Providers

Extremely Fast Distributed Storage for Cloud Service Providers Solution brief Intel Storage Builders StorPool Storage Intel SSD DC S3510 Series Intel Xeon Processor E3 and E5 Families Intel Ethernet Converged Network Adapter X710 Family Extremely Fast Distributed

More information

Maximum Performance. How to get it and how to avoid pitfalls. Christoph Lameter, PhD

Maximum Performance. How to get it and how to avoid pitfalls. Christoph Lameter, PhD Maximum Performance How to get it and how to avoid pitfalls Christoph Lameter, PhD cl@linux.com Performance Just push a button? Systems are optimized by default for good general performance in all areas.

More information

Workspace & Storage Infrastructure for Service Providers

Workspace & Storage Infrastructure for Service Providers Workspace & Storage Infrastructure for Service Providers Garry Soriano Regional Technical Consultant Citrix Cloud Channel Summit 2015 @rhipecloud #RCCS15 The industry s most complete Mobile Workspace solution

More information

Evaluation of the Chelsio T580-CR iscsi Offload adapter

Evaluation of the Chelsio T580-CR iscsi Offload adapter October 2016 Evaluation of the Chelsio T580-CR iscsi iscsi Offload makes a difference Executive Summary As application processing demands increase and the amount of data continues to grow, getting this

More information

Introduction to Operating Systems

Introduction to Operating Systems Introduction to Operating Systems Chapter 1 1.3 Chapter 1.5 1.9 Learning Outcomes High-level understand what is an operating system and the role it plays A high-level understanding of the structure of

More information

Titan: Fair Packet Scheduling for Commodity Multiqueue NICs. Brent Stephens, Arjun Singhvi, Aditya Akella, and Mike Swift July 13 th, 2017

Titan: Fair Packet Scheduling for Commodity Multiqueue NICs. Brent Stephens, Arjun Singhvi, Aditya Akella, and Mike Swift July 13 th, 2017 Titan: Fair Packet Scheduling for Commodity Multiqueue NICs Brent Stephens, Arjun Singhvi, Aditya Akella, and Mike Swift July 13 th, 2017 Ethernet line-rates are increasing! 2 Servers need: To drive increasing

More information

Farewell to Servers: Resource Disaggregation

Farewell to Servers: Resource Disaggregation Farewell to Servers: Hardware, Software, and Network Approaches towards Datacenter Resource Disaggregation Yiying Zhang 2 Monolithic Computer OS / Hypervisor 3 Can monolithic Application Hardware servers

More information

VM Migration Acceleration over 40GigE Meet SLA & Maximize ROI

VM Migration Acceleration over 40GigE Meet SLA & Maximize ROI VM Migration Acceleration over 40GigE Meet SLA & Maximize ROI Mellanox Technologies Inc. Motti Beck, Director Marketing Motti@mellanox.com Topics Introduction to Mellanox Technologies Inc. Why Cloud SLA

More information

Accelerate block service built on Ceph via SPDK Ziye Yang Intel

Accelerate block service built on Ceph via SPDK Ziye Yang Intel Accelerate block service built on Ceph via SPDK Ziye Yang Intel 1 Agenda SPDK Introduction Accelerate block service built on Ceph SPDK support in Ceph bluestore Summary 2 Agenda SPDK Introduction Accelerate

More information

Enabling Fast, Dynamic Network Processing with ClickOS

Enabling Fast, Dynamic Network Processing with ClickOS Enabling Fast, Dynamic Network Processing with ClickOS Joao Martins*, Mohamed Ahmed*, Costin Raiciu, Roberto Bifulco*, Vladimir Olteanu, Michio Honda*, Felipe Huici* * NEC Labs Europe, Heidelberg, Germany

More information

S K T e l e c o m : A S h a r e a b l e D A S P o o l u s i n g a L o w L a t e n c y N V M e A r r a y. Eric Chang / Program Manager / SK Telecom

S K T e l e c o m : A S h a r e a b l e D A S P o o l u s i n g a L o w L a t e n c y N V M e A r r a y. Eric Chang / Program Manager / SK Telecom S K T e l e c o m : A S h a r e a b l e D A S P o o l u s i n g a L o w L a t e n c y N V M e A r r a y Eric Chang / Program Manager / SK Telecom 2/23 Before We Begin SKT NV-Array (NVMe JBOF) has been

More information

Multifunction Networking Adapters

Multifunction Networking Adapters Ethernet s Extreme Makeover: Multifunction Networking Adapters Chuck Hudson Manager, ProLiant Networking Technology Hewlett-Packard 2004 Hewlett-Packard Development Company, L.P. The information contained

More information

pblk the OCSSD FTL Linux FAST Summit 18 Javier González Copyright 2018 CNEX Labs

pblk the OCSSD FTL Linux FAST Summit 18 Javier González Copyright 2018 CNEX Labs pblk the OCSSD FTL Linux FAST Summit 18 Javier González Read Latency Read Latency with 0% Writes Random Read 4K Percentiles 2 Read Latency Read Latency with 20% Writes Random Read 4K + Random Write 4K

More information

Why AI Frameworks Need (not only) RDMA?

Why AI Frameworks Need (not only) RDMA? Why AI Frameworks Need (not only) RDMA? With Design and Implementation Experience of Networking Support on TensorFlow GDR, Apache MXNet, WeChat Amber, and Tencent Angel Bairen Yi (byi@connect.ust.hk) Jingrong

More information

DPDK Summit China 2017

DPDK Summit China 2017 Summit China 2017 Embedded Network Architecture Optimization Based on Lin Hao T1 Networks Agenda Our History What is an embedded network device Challenge to us Requirements for device today Our solution

More information

Part 1: Introduction to device drivers Part 2: Overview of research on device driver reliability Part 3: Device drivers research at ERTOS

Part 1: Introduction to device drivers Part 2: Overview of research on device driver reliability Part 3: Device drivers research at ERTOS Some statistics 70% of OS code is in device s 3,448,000 out of 4,997,000 loc in Linux 2.6.27 A typical Linux laptop runs ~240,000 lines of kernel code, including ~72,000 loc in 36 different device s s

More information

Accelerating NVMe I/Os in Virtual Machine via SPDK vhost* Solution Ziye Yang, Changpeng Liu Senior software Engineer Intel

Accelerating NVMe I/Os in Virtual Machine via SPDK vhost* Solution Ziye Yang, Changpeng Liu Senior software Engineer Intel Accelerating NVMe I/Os in Virtual Machine via SPDK vhost* Solution Ziye Yang, Changpeng Liu Senior software Engineer Intel @optimistyzy Notices & Disclaimers Intel technologies features and benefits depend

More information

LITE Kernel RDMA. Support for Datacenter Applications. Shin-Yeh Tsai, Yiying Zhang

LITE Kernel RDMA. Support for Datacenter Applications. Shin-Yeh Tsai, Yiying Zhang LITE Kernel RDMA Support for Datacenter Applications Shin-Yeh Tsai, Yiying Zhang Time 2 Berkeley Socket Userspace Kernel Hardware Time 1983 2 Berkeley Socket TCP Offload engine Arrakis & mtcp IX Userspace

More information

Falcon: Scaling IO Performance in Multi-SSD Volumes. The George Washington University

Falcon: Scaling IO Performance in Multi-SSD Volumes. The George Washington University Falcon: Scaling IO Performance in Multi-SSD Volumes Pradeep Kumar H Howie Huang The George Washington University SSDs in Big Data Applications Recent trends advocate using many SSDs for higher throughput

More information