SlideShare uma empresa Scribd logo
1 de 24
HTTP at your local BigCo:How the internet sausage gets made	 Peter Griess @pgriess
Goals and non-goals Basics of TCP/IP, DNS and HTTP and how they work together; pitfalls and optimizations A 1,000 foot view of scaling out HTTP infrastructure All manner of load balancing / traffic shaping Living on the edge Not: how to make a fast application (database access, rendering performance, etc)
Background: DNS Map hostnames to IP(s) www.facebook.com 69.171.229.12, 69.171.228.40 Resolution process Recursion (and what does the DNS server see?) Caching Latencies: on-host, cached in LAN, cached at ISP, miss
Background: TCP Stateful protocol Negotiated by a synchronous 3-way handshake: 2xRTT before first byte is sent! e.g. USA => South America ~250ms RTT Seamless failover is hard (but not impossible) Load balancing must be aware of flows
Background: HTTP Layered on top of TCP/TLS Has some useful bits Compression Connection re-use Pipelining Caching Kind of sucks Headers on all requests/responses Compression on bodies only Pipelining has to be disabled most of the time Pipelining suffers from head-of-line blocking
mycutekittens.tv 68.193.17.4 Big bad internet HTTP
Problem?
Problem Availability Server goes down (kernel panic?) Network goes down (cable cut?) Datacenter goes down (EC2?) Overload Shed load (good, can be transparent) Get infinitely slow (not good)
mycutekittens.tv: multi-server Big bad internet ???
We have options DNS load balancing IP load balancing HTTP load balancing
DNS load balancing mycutekittens.tv resolves to IPs: A, B, C, D Add new IPs to scale out Remove IPs when hosts go down Benefits Don’t need extra hardware to do load balancing Can span datacenters DNS servers are cheap / fast Drawbacks Hotspots due to caching Hotspots due to ordering in result list Hotspots due to resolver size TTL / flexibility trade-off
mycutekittens.tv: DNS Big bad internet DNS Server DNS 68.193.17.4 68.193.17.5 68.193.17.6
IP load balancing (1) mycutekittens.tv resolves to 1 public IP owned by an IP load balancer Add new backend hosts w/ private IPs to scale out Load balancer health-checks hosts actively or passively to avoid dead hosts Scheduling policies vs. failover DSR
IP load balancing (2) Benefits Only 1 public IP (high DNS TTL) Backend network capacity/membership transparent to the internet Cheap-ish Failover is possible, not insanely difficult Drawbacks Can’t do what you can with HTTP
mycutekittens.tv: IP 10.0.0.1 Big bad internet 10.0.0.2 GW 68.193.17.4 10.0.0.3 LB
HTTP load balancing (1) mycutekittens.tv resolves to 1 public IP owned by an HTTP load balancer Largely same as IP load balancing Terminates TCP connections (sees all bytes) Can make routing decisions based on HTTP Can autonomously serve requests (caching, access control, etc) Examples: Send requests for /foo/* to pool A 401 requests without cookie Q
HTTP load balancing (2) Benefits Largely the same as IP More flexible rules Can terminate TLS (security+, cost+) Drawbacks No DSR Failover difficult Not as performant as IP
mycutekittens.tv: HTTP 10.0.0.1 Big bad internet 10.0.0.2 68.193.17.4 LB HTTP(S) 10.0.0.3
mycutekittens.tv: MOAR Eventually a single LB is going to be a problem Not enough capacity Availability Turtles all the day way down LB of LBs! DNS load balancing between datacenters …
HTTPS: myths and reality Too computationally expensive Only a few percent (imperialviolet.org); is your webserver actually CPU bound? doubt it SSL acceleration cards, GPUs, etc Too much latency Handshaking is 5-7xRTT Session resume False start Snap start Caching breaks
My latency is huge in Japan RTT to USA is (or any single DC) can be huge Re-use connections (connection: keep-alive) Send work in parallel (pipelining) Use compression (content-encoding) Lots of tricks for static resources (bundling, CDNs, caching, etc) Pre-fetch data
Let’s get crazy: SPDY Don’t limit yourself to HTTP; use a different protocol SPDY developed by Google, supported by Chrome, google.com (and soon facebook.com) Connection re-use w/o head-of-line blocking Headers always compressed Always SSL (but breaks caching)
Let’s get crazy: TCP termination Synchronous RTTs: the silent killer Opening new TCP connections is very costly Run proxies close to users and proxy traffic back to core using optimized protocol Low RTT to proxy Do SPDY-like tricks between edge + core Potentially faster network to core than public internet Advertise these proxies via DNS Geo-targetting AS-adjacency Akamai CDN does this, sort of
Let’s get crazy: DNS anycast Remember how DNS resolutions were slow? DNS servers could be far away from a user Advertise multiple network routes for the same DNS IP, let the IP stack pick the closest one

Mais conteúdo relacionado

Mais procurados

Introducing HTTP/2
Introducing HTTP/2Introducing HTTP/2
Introducing HTTP/2Ido Flatow
 
HTTP/2 Introduction
HTTP/2 IntroductionHTTP/2 Introduction
HTTP/2 IntroductionWalter Liu
 
So that was HTTP/2, what's next?
So that was HTTP/2, what's next?So that was HTTP/2, what's next?
So that was HTTP/2, what's next?Daniel Stenberg
 
Load Balancing with HAproxy
Load Balancing with HAproxyLoad Balancing with HAproxy
Load Balancing with HAproxyBrendan Jennings
 
HTTP/2: What no one is telling you
HTTP/2: What no one is telling youHTTP/2: What no one is telling you
HTTP/2: What no one is telling youFastly
 
HTTP/2 for Developers
HTTP/2 for DevelopersHTTP/2 for Developers
HTTP/2 for DevelopersSvetlin Nakov
 
Introduction to HTTP/2
Introduction to HTTP/2Introduction to HTTP/2
Introduction to HTTP/2Ido Flatow
 
HTTP/2 standard for video streaming
HTTP/2 standard for video streamingHTTP/2 standard for video streaming
HTTP/2 standard for video streamingHung Thai Le
 
HTTP 2.0 – What do I need to know?
HTTP 2.0 – What do I need to know? HTTP 2.0 – What do I need to know?
HTTP 2.0 – What do I need to know? Sigma Software
 
ChinaNetCloud Training - HAProxy Intro
ChinaNetCloud Training - HAProxy IntroChinaNetCloud Training - HAProxy Intro
ChinaNetCloud Training - HAProxy IntroChinaNetCloud
 
Web Server Load Balancer
Web Server Load BalancerWeb Server Load Balancer
Web Server Load BalancerMobME Technical
 
Choosing A Proxy Server - Apachecon 2014
Choosing A Proxy Server - Apachecon 2014Choosing A Proxy Server - Apachecon 2014
Choosing A Proxy Server - Apachecon 2014bryan_call
 
HAProxy scale out using open source
HAProxy scale out using open sourceHAProxy scale out using open source
HAProxy scale out using open sourceIngo Walz
 

Mais procurados (20)

Introducing HTTP/2
Introducing HTTP/2Introducing HTTP/2
Introducing HTTP/2
 
HTTP/2 Introduction
HTTP/2 IntroductionHTTP/2 Introduction
HTTP/2 Introduction
 
So that was HTTP/2, what's next?
So that was HTTP/2, what's next?So that was HTTP/2, what's next?
So that was HTTP/2, what's next?
 
Load Balancing with HAproxy
Load Balancing with HAproxyLoad Balancing with HAproxy
Load Balancing with HAproxy
 
Web server
Web serverWeb server
Web server
 
HTTP/2: What no one is telling you
HTTP/2: What no one is telling youHTTP/2: What no one is telling you
HTTP/2: What no one is telling you
 
Http2 right now
Http2 right nowHttp2 right now
Http2 right now
 
HTTP/2 for Developers
HTTP/2 for DevelopersHTTP/2 for Developers
HTTP/2 for Developers
 
Introduction to HTTP/2
Introduction to HTTP/2Introduction to HTTP/2
Introduction to HTTP/2
 
HTTP2 is Here!
HTTP2 is Here!HTTP2 is Here!
HTTP2 is Here!
 
HTTP/2 standard for video streaming
HTTP/2 standard for video streamingHTTP/2 standard for video streaming
HTTP/2 standard for video streaming
 
HTTP 2.0 – What do I need to know?
HTTP 2.0 – What do I need to know? HTTP 2.0 – What do I need to know?
HTTP 2.0 – What do I need to know?
 
Http/2
Http/2Http/2
Http/2
 
ChinaNetCloud Training - HAProxy Intro
ChinaNetCloud Training - HAProxy IntroChinaNetCloud Training - HAProxy Intro
ChinaNetCloud Training - HAProxy Intro
 
Netkitmig
NetkitmigNetkitmig
Netkitmig
 
Web Server Load Balancer
Web Server Load BalancerWeb Server Load Balancer
Web Server Load Balancer
 
Choosing A Proxy Server - Apachecon 2014
Choosing A Proxy Server - Apachecon 2014Choosing A Proxy Server - Apachecon 2014
Choosing A Proxy Server - Apachecon 2014
 
6 technical-dns-workshop-day3
6 technical-dns-workshop-day36 technical-dns-workshop-day3
6 technical-dns-workshop-day3
 
HAProxy scale out using open source
HAProxy scale out using open sourceHAProxy scale out using open source
HAProxy scale out using open source
 
SPDY and HTTP/2
SPDY and HTTP/2SPDY and HTTP/2
SPDY and HTTP/2
 

Destaque

Destaque (7)

A póker titkai mike caro
A póker titkai   mike caroA póker titkai   mike caro
A póker titkai mike caro
 
Itay
ItayItay
Itay
 
Introwebinargettingdiversebuttsthedwcgroup7 11-dc-edits-2
Introwebinargettingdiversebuttsthedwcgroup7 11-dc-edits-2Introwebinargettingdiversebuttsthedwcgroup7 11-dc-edits-2
Introwebinargettingdiversebuttsthedwcgroup7 11-dc-edits-2
 
Slidesharetest
SlidesharetestSlidesharetest
Slidesharetest
 
The New Black Test
The New Black   TestThe New Black   Test
The New Black Test
 
donotusethisversion-dc-edits
donotusethisversion-dc-editsdonotusethisversion-dc-edits
donotusethisversion-dc-edits
 
Augmented Reality
Augmented RealityAugmented Reality
Augmented Reality
 

Semelhante a HTTP at your local BigCo

Next generation web protocols
Next generation web protocolsNext generation web protocols
Next generation web protocolsDaniel Austin
 
HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612
HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612
HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612Avenga Germany GmbH
 
Computer network (10)
Computer network (10)Computer network (10)
Computer network (10)NYversity
 
Data center disaster recovery.ppt
Data center disaster recovery.ppt Data center disaster recovery.ppt
Data center disaster recovery.ppt omalreda
 
HTML5, HTTP2, and You 1.1
HTML5, HTTP2, and You 1.1HTML5, HTTP2, and You 1.1
HTML5, HTTP2, and You 1.1Daniel Austin
 
Scaling Streaming - Concepts, Research, Goals
Scaling Streaming - Concepts, Research, GoalsScaling Streaming - Concepts, Research, Goals
Scaling Streaming - Concepts, Research, Goalskamaelian
 
(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014
(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014
(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014Amazon Web Services
 
Improving performance by changing the rules from fast to SPDY
Improving performance by changing the rules   from fast to SPDYImproving performance by changing the rules   from fast to SPDY
Improving performance by changing the rules from fast to SPDYCotendo
 
From Fast To SPDY
From Fast To SPDYFrom Fast To SPDY
From Fast To SPDYMike Belshe
 
Linux HTTPS/TCP/IP Stack for the Fast and Secure Web
Linux HTTPS/TCP/IP Stack for the Fast and Secure WebLinux HTTPS/TCP/IP Stack for the Fast and Secure Web
Linux HTTPS/TCP/IP Stack for the Fast and Secure WebAll Things Open
 
Serverless (Distributed computing)
Serverless (Distributed computing)Serverless (Distributed computing)
Serverless (Distributed computing)Sri Prasanna
 
Cleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the Web
Cleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the WebCleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the Web
Cleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the WebSteffen Gebert
 
My other computer_is_a_datacentre
My other computer_is_a_datacentreMy other computer_is_a_datacentre
My other computer_is_a_datacentreSteve Loughran
 
HBaseCon 2015: HBase at Scale in an Online and High-Demand Environment
HBaseCon 2015: HBase at Scale in an Online and  High-Demand EnvironmentHBaseCon 2015: HBase at Scale in an Online and  High-Demand Environment
HBaseCon 2015: HBase at Scale in an Online and High-Demand EnvironmentHBaseCon
 
Stephan Ewen - Experiences running Flink at Very Large Scale
Stephan Ewen -  Experiences running Flink at Very Large ScaleStephan Ewen -  Experiences running Flink at Very Large Scale
Stephan Ewen - Experiences running Flink at Very Large ScaleVerverica
 
Network and distributed systems
Network and distributed systemsNetwork and distributed systems
Network and distributed systemsSri Prasanna
 
MNPHP Scalable Architecture 101 - Feb 3 2011
MNPHP Scalable Architecture 101 - Feb 3 2011MNPHP Scalable Architecture 101 - Feb 3 2011
MNPHP Scalable Architecture 101 - Feb 3 2011Mike Willbanks
 
Knowledge share about scalable application architecture
Knowledge share about scalable application architectureKnowledge share about scalable application architecture
Knowledge share about scalable application architectureAHM Pervej Kabir
 

Semelhante a HTTP at your local BigCo (20)

Next generation web protocols
Next generation web protocolsNext generation web protocols
Next generation web protocols
 
HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612
HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612
HTTP/2 turns 3 years old // Web Performance Meetup wao.io 20180612
 
Computer network (10)
Computer network (10)Computer network (10)
Computer network (10)
 
Data center disaster recovery.ppt
Data center disaster recovery.ppt Data center disaster recovery.ppt
Data center disaster recovery.ppt
 
Http2
Http2Http2
Http2
 
HTML5, HTTP2, and You 1.1
HTML5, HTTP2, and You 1.1HTML5, HTTP2, and You 1.1
HTML5, HTTP2, and You 1.1
 
Scaling Streaming - Concepts, Research, Goals
Scaling Streaming - Concepts, Research, GoalsScaling Streaming - Concepts, Research, Goals
Scaling Streaming - Concepts, Research, Goals
 
(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014
(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014
(WEB401) Optimizing Your Web Server on AWS | AWS re:Invent 2014
 
Improving performance by changing the rules from fast to SPDY
Improving performance by changing the rules   from fast to SPDYImproving performance by changing the rules   from fast to SPDY
Improving performance by changing the rules from fast to SPDY
 
From Fast To SPDY
From Fast To SPDYFrom Fast To SPDY
From Fast To SPDY
 
Linux HTTPS/TCP/IP Stack for the Fast and Secure Web
Linux HTTPS/TCP/IP Stack for the Fast and Secure WebLinux HTTPS/TCP/IP Stack for the Fast and Secure Web
Linux HTTPS/TCP/IP Stack for the Fast and Secure Web
 
Serverless (Distributed computing)
Serverless (Distributed computing)Serverless (Distributed computing)
Serverless (Distributed computing)
 
Cleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the Web
Cleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the WebCleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the Web
Cleaning Up the Dirt of the Nineties - How New Protocols are Modernizing the Web
 
My other computer_is_a_datacentre
My other computer_is_a_datacentreMy other computer_is_a_datacentre
My other computer_is_a_datacentre
 
HBaseCon 2015: HBase at Scale in an Online and High-Demand Environment
HBaseCon 2015: HBase at Scale in an Online and  High-Demand EnvironmentHBaseCon 2015: HBase at Scale in an Online and  High-Demand Environment
HBaseCon 2015: HBase at Scale in an Online and High-Demand Environment
 
Http2 in practice
Http2 in practiceHttp2 in practice
Http2 in practice
 
Stephan Ewen - Experiences running Flink at Very Large Scale
Stephan Ewen -  Experiences running Flink at Very Large ScaleStephan Ewen -  Experiences running Flink at Very Large Scale
Stephan Ewen - Experiences running Flink at Very Large Scale
 
Network and distributed systems
Network and distributed systemsNetwork and distributed systems
Network and distributed systems
 
MNPHP Scalable Architecture 101 - Feb 3 2011
MNPHP Scalable Architecture 101 - Feb 3 2011MNPHP Scalable Architecture 101 - Feb 3 2011
MNPHP Scalable Architecture 101 - Feb 3 2011
 
Knowledge share about scalable application architecture
Knowledge share about scalable application architectureKnowledge share about scalable application architecture
Knowledge share about scalable application architecture
 

Último

原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量
原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量
原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量208367051
 
UNIT-V-ELECTRIC AND HYBRID VEHICLES.pptx
UNIT-V-ELECTRIC AND HYBRID VEHICLES.pptxUNIT-V-ELECTRIC AND HYBRID VEHICLES.pptx
UNIT-V-ELECTRIC AND HYBRID VEHICLES.pptxDineshKumar4165
 
UNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGER
UNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGERUNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGER
UNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGERDineshKumar4165
 
(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一
(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一
(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一ejgeojhg
 
Independent Andheri Call Girls 9833363713
Independent Andheri Call Girls 9833363713Independent Andheri Call Girls 9833363713
Independent Andheri Call Girls 9833363713Komal Khan
 
办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书
办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书
办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书zdzoqco
 
UNIT-1-VEHICLE STRUCTURE AND ENGINES.ppt
UNIT-1-VEHICLE STRUCTURE AND ENGINES.pptUNIT-1-VEHICLE STRUCTURE AND ENGINES.ppt
UNIT-1-VEHICLE STRUCTURE AND ENGINES.pptDineshKumar4165
 
suspensionretardersystem-220726153321-060937ae.pptx
suspensionretardersystem-220726153321-060937ae.pptxsuspensionretardersystem-220726153321-060937ae.pptx
suspensionretardersystem-220726153321-060937ae.pptxbasant11731
 
定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制
定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制
定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制miwhjk
 
UNIT-III-TRANSMISSION SYSTEMS REAR AXLES
UNIT-III-TRANSMISSION SYSTEMS REAR AXLESUNIT-III-TRANSMISSION SYSTEMS REAR AXLES
UNIT-III-TRANSMISSION SYSTEMS REAR AXLESDineshKumar4165
 
UNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHER
UNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHERUNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHER
UNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHERunosafeads
 
-The-Present-Simple-Tense.pdf english hh
-The-Present-Simple-Tense.pdf english hh-The-Present-Simple-Tense.pdf english hh
-The-Present-Simple-Tense.pdf english hhmhamadhawlery16
 
2.2 CLSS hydraulic Pumps on komatsu pc200 series
2.2 CLSS hydraulic Pumps on komatsu pc200 series2.2 CLSS hydraulic Pumps on komatsu pc200 series
2.2 CLSS hydraulic Pumps on komatsu pc200 seriesdatazaky
 
Dubai Call Girls Services Call 09900000000
Dubai Call Girls Services Call 09900000000Dubai Call Girls Services Call 09900000000
Dubai Call Girls Services Call 09900000000Komal Khan
 
原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证
原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证
原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证jdkhjh
 
办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书
办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书
办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书zdzoqco
 
如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一
如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一
如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一fjjwgk
 
Call Girls Vastrapur 7397865700 Ridhima Hire Me Full Night
Call Girls Vastrapur 7397865700 Ridhima Hire Me Full NightCall Girls Vastrapur 7397865700 Ridhima Hire Me Full Night
Call Girls Vastrapur 7397865700 Ridhima Hire Me Full Nightssuser7cb4ff
 
办理克莱姆森大学毕业证成绩单|购买美国文凭证书
办理克莱姆森大学毕业证成绩单|购买美国文凭证书办理克莱姆森大学毕业证成绩单|购买美国文凭证书
办理克莱姆森大学毕业证成绩单|购买美国文凭证书zdzoqco
 

Último (20)

原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量
原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量
原版1:1定制(IC大学毕业证)帝国理工学院大学毕业证国外文凭复刻成绩单#电子版制作#留信入库#多年经营绝对保证质量
 
UNIT-V-ELECTRIC AND HYBRID VEHICLES.pptx
UNIT-V-ELECTRIC AND HYBRID VEHICLES.pptxUNIT-V-ELECTRIC AND HYBRID VEHICLES.pptx
UNIT-V-ELECTRIC AND HYBRID VEHICLES.pptx
 
UNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGER
UNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGERUNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGER
UNIT-II-ENGINE AUXILIARY SYSTEMS &TURBOCHARGER
 
(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一
(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一
(Griffith毕业证)格里菲斯大学毕业证毕业证成绩单修改留信学历认证原版一比一
 
Independent Andheri Call Girls 9833363713
Independent Andheri Call Girls 9833363713Independent Andheri Call Girls 9833363713
Independent Andheri Call Girls 9833363713
 
办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书
办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书
办理乔治布朗学院毕业证成绩单|购买加拿大文凭证书
 
UNIT-1-VEHICLE STRUCTURE AND ENGINES.ppt
UNIT-1-VEHICLE STRUCTURE AND ENGINES.pptUNIT-1-VEHICLE STRUCTURE AND ENGINES.ppt
UNIT-1-VEHICLE STRUCTURE AND ENGINES.ppt
 
suspensionretardersystem-220726153321-060937ae.pptx
suspensionretardersystem-220726153321-060937ae.pptxsuspensionretardersystem-220726153321-060937ae.pptx
suspensionretardersystem-220726153321-060937ae.pptx
 
定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制
定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制
定制(MSVU毕业证书)圣文森山大学毕业证毕业证成绩单原版一比一定制
 
UNIT-III-TRANSMISSION SYSTEMS REAR AXLES
UNIT-III-TRANSMISSION SYSTEMS REAR AXLESUNIT-III-TRANSMISSION SYSTEMS REAR AXLES
UNIT-III-TRANSMISSION SYSTEMS REAR AXLES
 
UNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHER
UNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHERUNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHER
UNOSAFE ELEVATOR PRIVATE LTD BANGALORE BROUCHER
 
Hot Sexy call girls in Pira Garhi🔝 9953056974 🔝 escort Service
Hot Sexy call girls in Pira Garhi🔝 9953056974 🔝 escort ServiceHot Sexy call girls in Pira Garhi🔝 9953056974 🔝 escort Service
Hot Sexy call girls in Pira Garhi🔝 9953056974 🔝 escort Service
 
-The-Present-Simple-Tense.pdf english hh
-The-Present-Simple-Tense.pdf english hh-The-Present-Simple-Tense.pdf english hh
-The-Present-Simple-Tense.pdf english hh
 
2.2 CLSS hydraulic Pumps on komatsu pc200 series
2.2 CLSS hydraulic Pumps on komatsu pc200 series2.2 CLSS hydraulic Pumps on komatsu pc200 series
2.2 CLSS hydraulic Pumps on komatsu pc200 series
 
Dubai Call Girls Services Call 09900000000
Dubai Call Girls Services Call 09900000000Dubai Call Girls Services Call 09900000000
Dubai Call Girls Services Call 09900000000
 
原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证
原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证
原版1:1复刻俄亥俄州立大学毕业证OSU毕业证留信学历认证
 
办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书
办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书
办理萨省大学毕业证成绩单|购买加拿大USASK文凭证书
 
如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一
如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一
如何办理(UC毕业证书)堪培拉大学毕业证毕业证成绩单原版一比一
 
Call Girls Vastrapur 7397865700 Ridhima Hire Me Full Night
Call Girls Vastrapur 7397865700 Ridhima Hire Me Full NightCall Girls Vastrapur 7397865700 Ridhima Hire Me Full Night
Call Girls Vastrapur 7397865700 Ridhima Hire Me Full Night
 
办理克莱姆森大学毕业证成绩单|购买美国文凭证书
办理克莱姆森大学毕业证成绩单|购买美国文凭证书办理克莱姆森大学毕业证成绩单|购买美国文凭证书
办理克莱姆森大学毕业证成绩单|购买美国文凭证书
 

HTTP at your local BigCo

  • 1. HTTP at your local BigCo:How the internet sausage gets made Peter Griess @pgriess
  • 2. Goals and non-goals Basics of TCP/IP, DNS and HTTP and how they work together; pitfalls and optimizations A 1,000 foot view of scaling out HTTP infrastructure All manner of load balancing / traffic shaping Living on the edge Not: how to make a fast application (database access, rendering performance, etc)
  • 3. Background: DNS Map hostnames to IP(s) www.facebook.com 69.171.229.12, 69.171.228.40 Resolution process Recursion (and what does the DNS server see?) Caching Latencies: on-host, cached in LAN, cached at ISP, miss
  • 4. Background: TCP Stateful protocol Negotiated by a synchronous 3-way handshake: 2xRTT before first byte is sent! e.g. USA => South America ~250ms RTT Seamless failover is hard (but not impossible) Load balancing must be aware of flows
  • 5. Background: HTTP Layered on top of TCP/TLS Has some useful bits Compression Connection re-use Pipelining Caching Kind of sucks Headers on all requests/responses Compression on bodies only Pipelining has to be disabled most of the time Pipelining suffers from head-of-line blocking
  • 8. Problem Availability Server goes down (kernel panic?) Network goes down (cable cut?) Datacenter goes down (EC2?) Overload Shed load (good, can be transparent) Get infinitely slow (not good)
  • 10. We have options DNS load balancing IP load balancing HTTP load balancing
  • 11. DNS load balancing mycutekittens.tv resolves to IPs: A, B, C, D Add new IPs to scale out Remove IPs when hosts go down Benefits Don’t need extra hardware to do load balancing Can span datacenters DNS servers are cheap / fast Drawbacks Hotspots due to caching Hotspots due to ordering in result list Hotspots due to resolver size TTL / flexibility trade-off
  • 12. mycutekittens.tv: DNS Big bad internet DNS Server DNS 68.193.17.4 68.193.17.5 68.193.17.6
  • 13. IP load balancing (1) mycutekittens.tv resolves to 1 public IP owned by an IP load balancer Add new backend hosts w/ private IPs to scale out Load balancer health-checks hosts actively or passively to avoid dead hosts Scheduling policies vs. failover DSR
  • 14. IP load balancing (2) Benefits Only 1 public IP (high DNS TTL) Backend network capacity/membership transparent to the internet Cheap-ish Failover is possible, not insanely difficult Drawbacks Can’t do what you can with HTTP
  • 15. mycutekittens.tv: IP 10.0.0.1 Big bad internet 10.0.0.2 GW 68.193.17.4 10.0.0.3 LB
  • 16. HTTP load balancing (1) mycutekittens.tv resolves to 1 public IP owned by an HTTP load balancer Largely same as IP load balancing Terminates TCP connections (sees all bytes) Can make routing decisions based on HTTP Can autonomously serve requests (caching, access control, etc) Examples: Send requests for /foo/* to pool A 401 requests without cookie Q
  • 17. HTTP load balancing (2) Benefits Largely the same as IP More flexible rules Can terminate TLS (security+, cost+) Drawbacks No DSR Failover difficult Not as performant as IP
  • 18. mycutekittens.tv: HTTP 10.0.0.1 Big bad internet 10.0.0.2 68.193.17.4 LB HTTP(S) 10.0.0.3
  • 19. mycutekittens.tv: MOAR Eventually a single LB is going to be a problem Not enough capacity Availability Turtles all the day way down LB of LBs! DNS load balancing between datacenters …
  • 20. HTTPS: myths and reality Too computationally expensive Only a few percent (imperialviolet.org); is your webserver actually CPU bound? doubt it SSL acceleration cards, GPUs, etc Too much latency Handshaking is 5-7xRTT Session resume False start Snap start Caching breaks
  • 21. My latency is huge in Japan RTT to USA is (or any single DC) can be huge Re-use connections (connection: keep-alive) Send work in parallel (pipelining) Use compression (content-encoding) Lots of tricks for static resources (bundling, CDNs, caching, etc) Pre-fetch data
  • 22. Let’s get crazy: SPDY Don’t limit yourself to HTTP; use a different protocol SPDY developed by Google, supported by Chrome, google.com (and soon facebook.com) Connection re-use w/o head-of-line blocking Headers always compressed Always SSL (but breaks caching)
  • 23. Let’s get crazy: TCP termination Synchronous RTTs: the silent killer Opening new TCP connections is very costly Run proxies close to users and proxy traffic back to core using optimized protocol Low RTT to proxy Do SPDY-like tricks between edge + core Potentially faster network to core than public internet Advertise these proxies via DNS Geo-targetting AS-adjacency Akamai CDN does this, sort of
  • 24. Let’s get crazy: DNS anycast Remember how DNS resolutions were slow? DNS servers could be far away from a user Advertise multiple network routes for the same DNS IP, let the IP stack pick the closest one