EyeQ: (An engineer’s approach to) Taming network performance unpredictability in the Cloud Vimal Mohammad Alizadeh Balaji Prabhakar David Mazières Changhoon.

Slides:



Advertisements
Similar presentations
Fast Data at Massive Scale Lessons Learned at Facebook Bobby Johnson.
Advertisements

Elastic Provisioning In Virtual Private Clouds
Network Resource Broker for IPTV in Cloud Computing Lei Liang, Dan He University of Surrey, UK OGF 27, G2C Workshop 15 Oct 2009 Banff,
Data Center Networking with Multipath TCP
All rights reserved © 2006, Alcatel Grid Standardization & ETSI (May 2006) B. Berde, Alcatel R & I.
Wireless Networks Should Spread Spectrum On Demand Ramki Gummadi (MIT) Joint work with Hari Balakrishnan.
Improving Datacenter Performance and Robustness with Multipath TCP
Storage Design for Agile VDI Alex Danilychev, PhD.
The Platform as a Service Model for Networking Eric Keller, Jennifer Rexford Princeton University INM/WREN 2010.
Cloud Service Models and Performance Ang Li 09/13/2010.
Virtual Switching Without a Hypervisor for a More Secure Cloud Xin Jin Princeton University Joint work with Eric Keller(UPenn) and Jennifer Rexford(Princeton)
Mohammad Alizadeh, Albert Greenberg, David A. Maltz, Jitendra Padhye Parveen Patel, Balaji Prabhakar, Sudipta Sengupta, Murari Sridharan Microsoft Research.
Ananta: Cloud Scale Load Balancing
1 School of Computing Science Simon Fraser University CMPT 771/471: Internet Architecture & Protocols TCP-Friendly Transport Protocols.
Copyright 2010 ITRI 工業技術研究院 11 ITRI Cloud OS & Virtual Resource Management Patrick Fu System Software Division, CCMA/ITRI
1 Copyright © 2012 Juniper Networks, Inc. Executive Intro Slide Turn Trends into Opportunities Vertical Wide Michael Tjon-En-Fa Industry,
Scalable Rule Management for Data Centers Masoud Moshref, Minlan Yu, Abhishek Sharma, Ramesh Govindan 4/3/2013.
CloudStack Scalability Testing, Development, Results, and Futures Anthony Xu Apache CloudStack contributor.
Deconstructing Datacenter Packet Transport Mohammad Alizadeh, Shuang Yang, Sachin Katti, Nick McKeown, Balaji Prabhakar, Scott Shenker Stanford University.
Mohammad Alizadeh, Albert Greenberg, David A. Maltz, Jitendra Padhye Parveen Patel, Balaji Prabhakar, Sudipta Sengupta, Murari Sridharan Presented by Shaddi.
The Case for Enterprise Ready Virtual Private Clouds Timothy Wood, Alexandre Gerber *, K.K. Ramakrishnan *, Jacobus van der Merwe *, and Prashant Shenoy.
Stratos: A Network-Aware Orchestration Layer for Middleboxes in the Cloud Aditya Akella, Aaron Gember, Anand Krishnamurthy, Saul St. John University of.
Data Center Fabrics. Forwarding Today Layer 3 approach: – Assign IP addresses to hosts hierarchically based on their directly connected switch. – Use.
Improving Datacenter Performance and Robustness with Multipath TCP Costin Raiciu, Sebastien Barre, Christopher Pluntke, Adam Greenhalgh, Damon Wischik,
PFabric: Minimal Near-Optimal Datacenter Transport Mohammad Alizadeh Shuang Yang, Milad Sharif, Sachin Katti, Nick McKeown, Balaji Prabhakar, Scott Shenker.
Alan Shieh Cornell University Srikanth Kandula Albert Greenberg Changhoon Kim Bikas Saha Microsoft Research, Azure, Bing Sharing the Datacenter Network.
“It’s going to take a month to get a proof of concept going.” “I know VMM, but don’t know how it works with SPF and the Portal” “I know Azure, but.
Congestion Control An Overview -Jyothi Guntaka. Congestion  What is congestion ?  The aggregate demand for network resources exceeds the available capacity.
Alan Shieh Cornell University Srikanth Kandula Albert Greenberg Changhoon Kim Microsoft Research Seawall: Performance Isolation for Cloud Datacenter Networks.
Virtual Layer 2: A Scalable and Flexible Data-Center Network Work with Albert Greenberg, James R. Hamilton, Navendu Jain, Srikanth Kandula, Parantap Lahiri,
Defense: Christopher Francis, Rumou duan Data Center TCP (DCTCP) 1.
SEDCL: Stanford Experimental Data Center Laboratory.
Jennifer Rexford Princeton University MW 11:00am-12:20pm Data-Center Traffic Management COS 597E: Software Defined Networking.
Jennifer Rexford Fall 2010 (TTh 1:30-2:50 in COS 302) COS 561: Advanced Computer Networks Data.
Mohammad Alizadeh, Abdul Kabbani, Tom Edsall,
An Introduction to Cloud Computing. The challenge Add new services for your users quickly and cost effectively.
Mohammad Alizadeh Stanford University Joint with: Abdul Kabbani, Tom Edsall, Balaji Prabhakar, Amin Vahdat, Masato Yasuda HULL: High bandwidth, Ultra Low-Latency.
Sharing the Data Center Network Alan Shieh, Srikanth Kandula, Albert Greenberg, Changhoon Kim, Bikas Saha Microsoft Research, Cornell University, Windows.
Curbing Delays in Datacenters: Need Time to Save Time? Mohammad Alizadeh Sachin Katti, Balaji Prabhakar Insieme Networks Stanford University 1.
MDC417 Follow me on Working as Practice Manager for Insight, he is a subject matter expert in cloud, virtualization and management.
CON Software-Defined Networking in a Hybrid, Open Data Center Krishna Srinivasan Senior Principal Product Strategy Manager Oracle Virtual Networking.
VL2: A Scalable and Flexible Data Center Network Albert Greenberg, James R. Hamilton, Navendu Jain, Srikanth Kandula, Changhoon Kim, Parantap Lahiri, David.
Windows Azure Virtual Machines Anton Boyko. A Continuous Offering From Private to Public Cloud.
1 Agility in Virtualized Utility Computing Hangwei Qian, Elliot Miller, Wei Zhang Michael Rabinovich, Craig E. Wills {EECS Department, Case Western Reserve.
Data Center Load Balancing T Seminar Kristian Hartikainen Aalto University, Helsinki, Finland
Theophilus Benson*, Ashok Anand*, Aditya Akella*, Ming Zhang + *University of Wisconsin, Madison + Microsoft Research.
Slide 1/20 "PerfSight: Performance Diagnosis for Software Dataplanes." Wu, Wenfei, Keqiang He, and Aditya Akella ACM ICM, Presented by: Ayush Patwari.
MMPTCP: A Multipath Transport Protocol for Data Centres 1 Morteza Kheirkhah University of Edinburgh, UK Ian Wakeman and George Parisis University of Sussex,
6.888 Lecture 6: Network Performance Isolation Mohammad Alizadeh Spring
T3: TCP-based High-Performance and Congestion-aware Tunneling Protocol for Cloud Networking Satoshi Ogawa† Kazuki Yamazaki† Ryota Kawashima† Hiroshi Matsuo†
VL2: A Scalable and Flexible Data Center Network
Data Center Architectures
Chen Qian, Xin Li University of Kentucky
Data Center TCP (DCTCP)
CIS 700-5: The Design and Implementation of Cloud Networks
How I Learned to Stop Worrying About the Core and Love the Edge
Heitor Moraes, Marcos Vieira, Italo Cunha, Dorgival Guedes
An Introduction to Cloud Computing
ECE 544: Traffic engineering (supplement)
Improving Datacenter Performance and Robustness with Multipath TCP
Improving Datacenter Performance and Robustness with Multipath TCP
Congestion-Aware Load Balancing at the Virtual Edge
Azure Container Instances
NTHU CS5421 Cloud Computing
VL2: A Scalable and Flexible Data Center Network
Internet and Web Simple client-server model
Congestion-Aware Load Balancing at the Virtual Edge
CS 401/601 Computer Network Systems Mehmet Gunes
Lecture 8, Computer Networks (198:552)
Data Center Traffic Engineering
Presentation transcript:

EyeQ: (An engineer’s approach to) Taming network performance unpredictability in the Cloud Vimal Mohammad Alizadeh Balaji Prabhakar David Mazières Changhoon Kim Albert Greenberg

What are we depending on? lessons-weve-learned-using-aws.html 5 Lessons We’ve Learned Using AWS … in the Netflix data centers, we have a high capacity, super fast, highly reliable network. This has afforded us the luxury of designing around chatty APIs to remote systems. AWS networking has more variable latency. Overhaul apps to deal with variability 2 Many customers don’t even realise network issues: Just “spin up more VMs!” Makes app more network dep.

Cloud: Warehouse Scale Computer Multi-tenancy: To increase cluster utilisation 6/11/123 Provisioning the Warehouse CPU, memory, disk Network

Sharing the Network Policy – Sharing model Mechanism – Computing rates – Enforcing rates on entities… Per-VM (multi-tenant) Per-service (search, map-reduce, etc.) 6/11/124 Can we achieve this? 2Ghz VCPU 15GB memory 1Gbps network Tenant X’s Virtual Switch VM1 VM2 VMn VM3 … Tenant Y’s Virtual Switch VM1 VM2 VMi VM3 … Customer X specifies the thickness of each pipe. No traffic matrix. (Hose Model)

Why is it hard? (1) Bandwidth demands can be… – Random, bursty – Short: few millisecond requests Timescales matter! – Need guarantees on the order of few RTTs (ms) 6/11/125 Default policy insufficient: 1 vs many TCP flows, UDP, etc. Poor scalability of traditional QoS mechanisms 10–100KB 10–100MB

Seconds: Eternity 6/11/126 Switch 1 Long lived TCP flow Bursty UDP session ON: 5ms OFF: 15ms Shared 10G pipe

Under the hood 6/11/127 Switch

Why is it hard? (2) 6/11/128 Switch Switch sees contention, but lacks VM state Receiver-host has VM state, but does not see contention (1) Drops in network: servers don’t see true demand (2) Elusive TCP (back-off) makes true demand detection harder

Key Idea: Bandwidth Headroom Bandwidth guarantees: managing congestion Congestion: link util reaches 100% – At millisecond timescales Don’t allow 100% util – 10% headroom: Early detection at receiver 6/11/129 N x 10G UDP TCP Shared pipe Limit to 9G Single Switch: Headroom What about a network?

Network design: the old 6/11/ for-cloud-and-big-data-interop-2012-session-teaser/ Over-subscription

Network design: the new 6/11/ for-cloud-and-big-data-interop-2012-session-teaser/ (1) Uniform capacity across racks (2) Over-subscription only at Top-of-Rack

Mitigating Congestion in a Network 6/11/1212 Load balancing + Admissibility = Hotspot free network core [VL2, FatTree, Hedera, MicroTE] Aggregate rate > 10Gbps Fabric gets congested Server VM 10Gbps pipe Fabric Aggregate rate < 10Gbps Congestion free Fabric Server VM 10Gbps pipe Fabric Load balancing: ECMP, etc. Admissibility: e2e congestion control (EyeQ)

EyeQ Platform 6/11/1213 TX packets VM TX VM Software VSwitch Adaptive Rate Limiters untrusted RX 3Gbps 6Gbps RX packets Software VSwitch VM Congestion Detectors untrusted VM RX component detects TX component reacts End-to-end flow control (VSwitch—VSwitch) DataCentre Fabric Congestion Feedback

Does it work? 6/11/1214 Without EyeQWith EyeQ Improves utilisation Provides protection TCP: 6Gbps UDP: 3Gbps

State: only at edge 15 EyeQ One Big Switch

Thanks! 16 EyeQ Load balancing + Bandwidth headroom + Admissibility at millisec timescales = Network as one big switch = Bandwidth sharing at edge Linux, Windows implementation for 10Gbps ~1700 lines C code (Linux kmod) No documentation, yet.

6/11/1217