Data Center Middleboxes Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking November 24,

Slides:



Advertisements
Similar presentations
Elastic Provisioning In Virtual Private Clouds
Advertisements

Dynamic Replica Placement for Scalable Content Delivery Yan Chen, Randy H. Katz, John D. Kubiatowicz {yanchen, randy, EECS Department.
Module 13: Implementing ISA Server 2004 Enterprise Edition: Site-to-Site VPN Scenario.
Economics of stub network multihoming and link load balancing INTERIM RESULTS AND NEXT STEPS Henna Warma Aalto University - COMNET December, 7 th 2011.
Justine Sherry*, Shaddi Hasan*, Colin Scott*, Arvind Krishnamurthy†,
All Rights Reserved © Alcatel-Lucent 2009 Enhancing Dynamic Cloud-based Services using Network Virtualization F. Hao, T.V. Lakshman, Sarit Mukherjee, H.
Barracuda Link Balancer Link Reliability and Bandwidth Optimization.
Practical and Incremental Convergence between SDN and Middleboxes 1 Zafar Qazi, Cheng-Chun Tu, Luis Chiang Vyas Sekar Rui Miao Minlan Yu.
Logically Centralized Control Class 2. Types of Networks ISP Networks – Entity only owns the switches – Throughput: 100GB-10TB – Heterogeneous devices:
SIMPLE-fying Middlebox Policy Enforcement Using SDN
Copyright © 2004 Juniper Networks, Inc. Proprietary and Confidentialwww.juniper.net 1 E-VPN and Data Center R. Aggarwal
MUNIS Platform Migration Project WELCOME. Agenda Introductions Tyler Cloud Overview Munis New Features Questions.
The Case for Enterprise Ready Virtual Private Clouds Timothy Wood, Alexandre Gerber *, K.K. Ramakrishnan *, Jacobus van der Merwe *, and Prashant Shenoy.
Practical and Incremental Convergence between SDN and Middleboxes 1 Zafar Qazi Cheng-Chun Tu Luis Chiang Vyas Sekar Rui Miao Minlan Yu.
Design and Implementation of a Consolidated Middlebox Architecture 1 Vyas SekarSylvia RatnasamyMichael ReiterNorbert Egi Guangyu Shi.
Data Center Storage and Networking Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking December.
15-744: Computer Networking
Highly Available Central Services An Intelligent Router Approach Thomas Finnern Thorsten Witt DESY/IT.
Web Caching Schemes1 A Survey of Web Caching Schemes for the Internet Jia Wang.
CSCD 433/533 Advanced Computer Networks Lecture 1 Course Overview Fall 2011.
This work is supported by the National Science Foundation under Grant Number DUE Any opinions, findings and conclusions or recommendations expressed.
Data Center Traffic and Measurements Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking.
Alternative Switching Technologies: Optical Circuit Switches Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance.
Data Center Networks and Basic Switching Technologies Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems.
Data Center Virtualization: Open vSwitch Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking.
A Scalable, Commodity Data Center Network Architecture Mohammad Al-Fares, Alexander Loukissas, Amin Vahdat Presented by Gregory Peaker and Tyler Maclean.
Data Center Traffic and Measurements: Available Bandwidth Estimation Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance.
Data Center Virtualization: VirtualWire Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking.
Week #10 Objectives: Remote Access and Mobile Computing Configure Mobile Computer and Device Settings Configure Remote Desktop and Remote Assistance for.
LB VIP:Input Endpoint Internal Endpoint foo.cloudapp.net  VIP.
Justine Sherry*, Shaddi Hasan*, Colin Scott*, Arvind Krishnamurthy†,
CS426Fall 2010/Lecture 361 Computer Security CS 426 Lecture 36 Perimeter Defense and Firewalls.
Chapter 1: Hierarchical Network Design
Software Routers: NetSlice Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking October 15,
Supporting Strong Cache Coherency for Active Caches in Multi-Tier Data-Centers over InfiniBand S. Narravula, P. Balaji, K. Vaidyanathan, S. Krishnamoorthy,
Data Center Virtualization: Xen and Xen-blanket
1 Chapter 6: Proxy Server in Internet and Intranet Designs Designs That Include Proxy Server Essential Proxy Server Design Concepts Data Protection in.
ESubnet Enterprises Inc. Richard Danielli, eSubnet Higher sales volumes through high network availability INTIX 2010.
© 2006 Cisco Systems, Inc. All rights reserved.Cisco Public 1 Version 4.0 Identifying Application Impacts on Network Design Designing and Supporting Computer.
Module 8 Configuring Mobile Computing and Remote Access in Windows® 7.
Module 3: Designing IP Addressing. Module Overview Designing an IPv4 Addressing Scheme Designing DHCP Implementation Designing DHCP Configuration Options.
Security Patterns in Wireless Sensor Networks By Y. Serge Joseph October 8 th, 2009 Part I.
608D CloudStack 3.0 Omer Palo Readiness Specialist, WW Tech Support Readiness May 8, 2012.
© 2006 Cisco Systems, Inc. All rights reserved.Cisco PublicITE I Chapter 6 1 Identifying Application Impacts on Network Design Designing and Supporting.
PowerLink Bandwidth Aggregation Redundant WAN Link and VPN Fail-Over Solutions.
Kiew-Hong Chua a.k.a Francis Computer Network Presentation 12/5/00.
Vytautas Valancius, Nick Feamster, Akihiro Nakao, and Jennifer Rexford.
Zibin Zheng DR 2 : Dynamic Request Routing for Tolerating Latency Variability in Cloud Applications CLOUD 2013 Jieming Zhu, Zibin.
Unleashing the Power of IP Communications™ Calling Across The Boundaries Mike Burkett, VP Products September 2002.
SOFTWARE DEFINED NETWORKING/OPENFLOW: A PATH TO PROGRAMMABLE NETWORKS April 23, 2012 © Brocade Communications Systems, Inc.
Benefits For hoster or private cloud: Multi-tenant gateway reduces cost Enhanced Reliability with Clustering Works Seamlessly with Hyper-V network.
VPN’s & Remote Access Issues David Trepp VP of Technology
Office 365 Performance Management. Meet Paul Andrew Office 365 Technical Product Manager – Office 365 datacenter, networking, identity management.
Microsoft Cloud Solution.  What is the cloud?  Windows Azure  What services does it offer?  How does it all work?  How to go about using it  Further.
Going Hybrid – part 1 Moving to Hybrid Cloud with Windows Azure Virtual Machines & System Center 2012 R2.
Marin Franković MVP: SCCDM Algebra visoko učilište What’s new in Azure for IT Pro.
Appendix B Advanced Topics in Exchange Server 2010.
By Harshal Ghule Guided by Mrs. Anita Mahajan G.H.Raisoni Institute Of Engineering And Technology.
Preliminaries: EE807 Software-defined Networked Computing KyoungSoo Park Department of Electrical Engineering KAIST.
Xin Li, Chen Qian University of Kentucky
Overview: Cloud Datacenters II
A Survey of Network Function Placement
15-744: Computer Networking
Authors: Justine Sherry. , Shaddi Hasan. , Colin Scott
15-744: Computer Networking
of Dynamic NFV-Policies
Wide Area Network Fundamentals
Lecture 21, Computer Networks (198:552)
Dynamic WAN Selection Optimize Your Business & Cloud Networks
Microsoft Virtual Academy
Presentation transcript:

Data Center Middleboxes Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking November 24, 2014 Slides from ACM SIGCOMM 2012 presentation on “Making middleboxes someone else's problem: network processing as a cloud service”

Overview and Basics Data Center Networks – Basic switching technologies – Data Center Network Topologies (today and Monday) – Software Routers (eg. Click, Routebricks, NetMap, Netslice) – Alternative Switching Technologies – Data Center Transport Data Center Software Networking – Software Defined networking (overview, control plane, data plane, NetFGPA) – Data Center Traffic and Measurements – Virtualizing Networks – Middleboxes Advanced Topics Where are we in the semester?

Goals for Today Making middleboxes someone else's problem: network processing as a cloud service, – J. Sherry, S. Hasan, C. Scott, A. Krishnamurthy, S. Ratnasamy, and V. Sekar. ACM SIGCOMM Computer Communication Review (CCR) Volume 42, Issue 4 (August 2012), pages

APLOMB “Appliance for Outsourcing Middleboxes” Place middleboxes in the cloud. Use APLOMB devices and DNS to redirect traffic to and from the cloud. That’s it.

Typical Enterprise Networks Internet

Typical Enterprise Networks Internet

A Survey 57 enterprise network administrators Small ( 100k hosts) Asked about deployment size, expenses, complexity, and failures.

Typically on par with # routers and switches. How many middleboxes do you deploy?

Many kinds of devices, all with different functions and management expertise required. What kinds of middleboxes do you deploy?

Average salary for a network engineer - $60-80k USD How many networking personnel are there?

Misconfig.OverloadPhysical/ Electrical Firewalls67.3%16.3% Proxies63.2%15.7%21.1% IDS54.45%11.4%34% Most administrators spent 1-5 hrs/week dealing with failures; 9% spent 6-10 hrs/week. How do administrators spend their time?

Recap High Capital and Operating Expenses Time Consuming and Error-Prone Physical and Overload Failures

How can we improve this?

Proposal Internet

Proposal Internet Cloud Provider

High Capital and Operating Expenses Time Consuming and Error Prone Physical and Overload Failures Economies of scale and pay-per use Simplifies configuration and deployment Redundant resources for failover A move to the cloud

Design

Challenges Minimal Complexity at the Enterprise Functional Equivalence Low Performance Overhead

APLOMB “Appliance for Outsourcing Middleboxes”

Outsourcing Middleboxes with APLOMB Internet Cloud Provider APLOMB Gateway NAT

Inbound Traffic Internet Cloud Provider Web Server: Enterprise Network Admin. Register:

Inbound Traffic Internet Cloud Provider DNS Register: enterprise.com

Minimizing latency?

External Client Choosing a Datacenter Cloud Provider East Cloud Provider West Enterprise Route through cloud datacenter that minimizes end to end latency. APLOMB Gateway keeps a “routing table” to select best tunnel for every Internet prefix. External Client

Caches and “Terminal Services” Traffic destined to services like caches should be redirected to the nearest node. Cloud Provider West

APLOMB “Appliance for Outsourcing Middleboxes” Place middleboxes in the cloud. Use APLOMB devices and DNS to redirect traffic to and from the cloud. That’s it.

Firewalls IDSes Load Balancers VPNs Proxy/Caches WAN Optimizers ✔ ✔ ✔ ✔ ✗ Bandwidth? ✗ Compression? Can we outsource all middleboxes?

I APLOMB+ for Compression Add generic compression to APLOMB gateway to reduce bandwidth consumption. Cloud Provider Internet

Firewalls IDSes Load Balancers VPNs Proxy/Caches WAN Optimizers ✔ ✔ ✔ ✔ ✗ Bandwidth? ✗ Compression? ✔ ✔ Can we outsource all middleboxes?

Does it work?

Deployment Cloud provider: EC2 – 7 Datacenters OpenVPN for tunneling, Vyatta for middlebox services Two Types of Clients: – Software VPN client on laptops – Tunneling software router for wired hosts

Implementation & Deployment Performance metrics Case Study of a Large Enterprise Impact in a real usage scenario Wide-Area Measurements Network latency Three Part Evaluation

Does APLOMB inflate latency?

For PlanetLab nodes, 60% of pairs’ latency improves with redirection through EC2.

Latency at a Large Enterprise Measured redirection latency between enterprise sites. Median latency inflation: 1.13 ms Sites experiencing inflation were primarily in areas where EC2 does not have a wide footprint.

How does APLOMB impact other quality metrics, like bandwidth and jitter?

Bandwidth: download times with BitTorrent increased on average 2.3% Jitter: consistently within industry standard bounds of 30ms

Does APLOMB negate the benefits of bandwidth-saving devices?

APLOMB+ incurs a median penalty of 3.8% bandwidth inflation over traditional WAN Optimizers.

Does “elastic scaling” at the cloud provide real benefits?

Some sites generate as much as 13x traffic more than average at peak hours.

Recap Good application performance – Latency median inflation 1.1ms – Download times increased only 2.3% Generic redundancy elimination saves bandwidth costs Strong benefits from elasticity

Moving middleboxes to the cloud seems to be practical and feasible solution to the complexity of enterprise networks. Conclusion and Discussion

Did the soln make the problem simpler? – How to measure simplicity/complexity? Does the soln also make security problems someone else's problems. – Do we trust the cloud provider? Privacy concerns? – Do we trust the cloud provider Monetary cost: Is APLOMB cheaper or more expensive? Precedence – Zscalar – Ariaka – Total uptime Middleboxes not at the edge of your network – APLOMB cannot outsource these middleboxes Conclusion and Discussion

Before Next time Project Interim report – Due Today, Monday, November 24. – And meet with groups, TA, and professor Fractus Upgrade: Should be back online Required review and reading for Monday, December 1 – IOFlow: a software-defined storage architecture, E. Thereska, H. Ballani, G. O'Shea, T. Karagiannis, A. Rowstron, T. Talpey, R. Black, T. Zhu. ACM Symposium on Operating Systems Principles (SOSP), October 2013, pages – Check piazza: Check website for updated schedule