steve souders souders@google.com even faster web sites disclaimer: this content does not...

Steve Souderssouders@google.com

http://stevesouders.com/docs/ebay-20090416.ppt

Even Faster Web Sites

Disclaimer: This content does not necessarily reflect the opinions of my employer.

17%

83%

iGoogle, primed cache

the importance of frontend performance

9% 91%

iGoogle, empty cache

time spent on the frontend

Empty Cache

Primed Cache

www.aol.com 97% 97%

www.ebay.com 95% 81%

www.facebook.com 95% 81%

www.google.com/search

47% 0%

search.live.com/results 67% 0%

www.msn.com 98% 94%

www.myspace.com 98% 98%

en.wikipedia.org/wiki 94% 91%

www.yahoo.com 97% 96%

www.youtube.com 98% 97%April 2008

The Performance Golden Rule

80-90% of the end-user response time is spent on the frontend. Start there.

greater potential for improvement

simpler

proven to work

14 RULES

1. MAKE FEWER HTTP REQUESTS

2. USE A CDN3. ADD AN EXPIRES HEADER4. GZIP COMPONENTS5. PUT STYLESHEETS AT THE

TOP6. PUT SCRIPTS AT THE

BOTTOM7. AVOID CSS EXPRESSIONS8. MAKE JS AND CSS

EXTERNAL9. REDUCE DNS LOOKUPS10.MINIFY JS11.AVOID REDIRECTS12.REMOVE DUPLICATE

SCRIPTS13.CONFIGURE ETAGS14.MAKE AJAX CACHEABLE

Sept 2007

June 2009

Even Faster Web SitesSplitting the initial payloadLoading scripts without blockingCoupling asynchronous scripts

Positioning inline scriptsSharding dominant domainsFlushing the document early

Using iframes sparingly

Simplifying CSS Selectors

Understanding Ajax performance...Doug Crockford

Creating responsive web apps......Ben Galbraith, Dion Almaer

Writing efficient JavaScript...........Nicholas Zakas

Scaling with Comet......................Dylan Schiemann

Going beyond gzipping...............Tony Gentilcore

Optimizing images.....................Stoyan Stefanov, Nicole Sullivan

AOLeBayFacebookMySpaceWikipediaYahoo!

Why focus on JavaScript?

YouTube

scripts block

<script src="A.js"> blocks parallel downloads and rendering

What's "Cuzillion"?

http://stevesouders.com/cuzillion/?ex=10008

a tool for quickly constructing web pages to see how components interact

Open Source

http://stevesouders.com/cuzillion/

Cuzillion'cuz there are a zillion pages to check

JavaScript

Functions Executed before

onload

www.aol.com 115K 30%

www.ebay.com 183K 44%

www.facebook.com 1088K 9%

www.google.com/search

15K 45%

search.live.com/results

17K 24%

www.msn.com 131K 31%

www.myspace.com 297K 18%

en.wikipedia.org/wiki 114K 32%

www.yahoo.com 321K 13%

www.youtube.com 240K 18%

26% avg252K avg

initial payload and execution

Splitting the initial payload

split your JavaScript between what's needed to render the page and everything else

load "everything else" after the page is rendered

separate manually (Firebug); tools needed to automate this (Doloto from Microsoft)

load scripts without blocking – how?

MSNScripts and other resources downloaded in parallel! How? Secret sauce?!var p= g.getElementsByTagName("HEAD")[0];var c=g.createElement("script");c.type="text/javascript";c.onreadystatechange=n;c.onerror=c.onload=k;c.src=e;p.appendChild(c)

MSN.com: parallel scripts

Loading Scripts Without Blocking

XHR Eval

XHR Injection

Script in Iframe

Script DOM Element

Script Defer

document.write Script Tag

XHR Eval

script must have same domain as main page

must refactor script

var xhrObj = getXHRObject();xhrObj.onreadystatechange = function() { if ( xhrObj.readyState != 4 ) return; eval(xhrObj.responseText); };xhrObj.open('GET', 'A.js', true);xhrObj.send('');

http://stevesouders.com/cuzillion/?ex=10009

XHR Injectionvar xhrObj = getXHRObject();xhrObj.onreadystatechange = function() { if ( xhrObj.readyState != 4 ) return; var se=document.createElement('script'); document.getElementsByTagName('head') [0].appendChild(se); se.text = xhrObj.responseText; };xhrObj.open('GET', 'A.js', true);xhrObj.send('');

script must have same domain as main pagehttp://stevesouders.com/cuzillion/?ex=10015

Script in Iframe<iframe src='A.html' width=0 height=0 frameborder=0 id=frame1></iframe>

iframe must have same domain as main page

must refactor script:// access iframe from main pagewindow.frames[0].createNewDiv();

// access main page from iframeparent.document.createElement('div');

http://stevesouders.com/cuzillion/?ex=10012

Script DOM Elementvar se = document.createElement('script');se.src = 'http://anydomain.com/A.js';document.getElementsByTagName('head')[0].appendChild(se);

script and main page domains can differ

no need to refactor JavaScript

http://stevesouders.com/cuzillion/?ex=10010

Script Defer<script defer src='A.js'></script>

only supported in IE (just landed in FF 3.1)

script and main page domains can differ

no need to refactor JavaScript

http://stevesouders.com/cuzillion/?ex=10013

document.write Script Tagdocument.write("<script type='text/javascript' src='A.js'> <\/script>");

parallelization only works in IE

parallel downloads for scripts, nothing else

all document.writes must be in same script block

http://stevesouders.com/cuzillion/?ex=10014

browser busy indicators

browser busy indicators

good to show busy indicators when the user needs feedback

bad when downloading in the background

Loading Scripts Without Blocking

*Only other document.write scripts are downloaded in parallel (in the same script block).

and the winner is...XHR EvalXHR InjectionScript in iframeScript DOM ElementScript Defer

Script DOM ElementScript Defer

Script DOM Element

Script DOM Element (FF)Script Defer (IE)

XHR EvalXHR InjectionScript in iframeScript DOM Element (IE)

XHR InjectionXHR EvalScript DOM Element (IE)

Managed XHR InjectionManaged XHR EvalScript DOM Element

Managed XHR InjectionManaged XHR Eval

Script DOM Element (FF)Script Defer (IE)Managed XHR EvalManaged XHR Injection

Script DOM Element (FF)Script Defer (IE)Managed XHR EvalManaged XHR Injection

different domains same domains

no order

preserve order

no order

no busyshow busy

show busyno busy

preserve order

asynchronous JS example: menu.js

<script type="text/javascript">var domscript = document.createElement('script');domscript.src = "menu.js"; document.getElementsByTagName('head')

[0].appendChild(domscript);

var aExamples = [ ['couple-normal.php', 'Normal Script Src'], ['couple-xhr-eval.php', 'XHR Eval'], ... ['managed-xhr.php', 'Managed XHR'] ];

function init() { EFWS.Menu.createMenu('examplesbtn', aExamples);}

init();</script>

script DOM element approach

before

after

Loading Scripts Without Blocking

*Only other document.write scripts are downloaded in parallel (in the same script block).

!IE

what about

inlined code that depends on the script?

coupling techniques

hardcoded callback

window onload

timer

degrading script tags

script onload

technique 5: script onload<script type="text/javascript">var aExamples = [['couple-normal.php', 'Normal Script Src'], ...];

function init() { EFWS.Menu.createMenu('examplesbtn', aExamples);}

var domscript = document.createElement('script');domscript.src = "menu.js";

domscript.onloadDone = false;domscript.onload = function() { if ( ! domscript.onloadDone ) { init(); } domscript.onloadDone = true; };domscript.onreadystatechange = function() { if ( "loaded" === domscript.readyState ) { if ( ! domscript.onloadDone ) { init(); } domscript.onloadDone = true; }}

document.getElementsByTagName('head')[0].appendChild(domscript);</script>

pretty nice, medium complexity

asynchronous loading & coupling

async technique: Script DOM Elementeasy, cross-browserdoesn't ensure script order

coupling technique: script onloadfairly easy, cross-browserensures execution order for external

script and inlined code

multiple interdependent external and inline scripts:much more complex (see hidden slides)concatenate your external scripts into

one!

bad: stylesheet followed by inline script

browsers download stylesheets in parallel with other resources that follow...

...unless the stylesheet is followed by an inline script

http://stevesouders.com/cuzillion/?ex=10021

best to move inline scripts above stylesheets or below other resources

use Link, not @import

eBayMSNMySpaceWikipedia

mispositioned inline scripts

Positioning Inline Scripts

remember inline scripts carry a cost

avoid long-executing inline scripts

don't put inline scripts between stylesheets and other resources

Simplifying CSS Selectors

#toc > LI { font-weight: bold; }

combinator

simple selectors

selector

declaration block

rule

types of CSS selectors

ID selectors#toc { margin-left: 20px; } element whose ID attribute has the value "toc"

class selectors.chapter { font-weight: bold; }elements with class=chapter

type selectorsA { text-decoration: none; }all A elements in the document tree

http://www.w3.org/TR/CSS2/selector.html

types of CSS selectors

adjacent sibling selectorsH1 + #toc { margin-top: 40px; } an element with ID=toc that immediately follows

an H1

child selectors#toc > LI { font-weight: bold; }all LI elements whose parent has id="toc"

descendant selectors#toc A { color: #444; }all A elements that have id="toc" as an ancestor

types of CSS selectors

universal selectors* { font-family: Arial; }all elements

attribute selectors[href="#index"] { font-style: italic; }all elements where the href attribute is "#index"

psuedo classes and elementsA:hover { text-decoration: underline; }non-DOM behaviorothers: :visited, :link, :active, :focus, :first-child, :before, :after

writing efficient CSS

https://developer.mozilla.org/en/Writing_Efficient_CSS

"The style system matches a rule by starting with the rightmost selector and moving to the left through the rule's selectors. As long as your little subtree continues to check out, the style system will continue moving to the left until it either matches the rule or bails out because of a mismatch."

#toc > LI { font-weight: bold; }find every LI whose parent is id="toc"

#toc A { color: #444; }find every A and climb its ancestors until id="toc" or DOM root (!) is found

writing efficient CSS

1.avoid universal selectors2.don't qualify ID selectors

bad: DIV #navbar {}good: #navbar {}

3.don't qualify class selectorsbad: LI .tight {}good: .li-tight {}

4.make rules as specific as possiblebad: #navbar A {}good: .a-navbar {}

https://developer.mozilla.org/en/Writing_Efficient_CSS

writing efficient CSS

5.avoid descendant selectorsbad: UL LI A {}better: UL > LI > A {}

6.avoid tag-child selectorsbad: UL > LI > A {}best: .li-anchor {}

7.be wary of child selectors8.rely on inheritance

http://www.w3.org/TR/CSS21/propidx.html

https://developer.mozilla.org/en/Writing_Efficient_CSSDavid Hyatt4/21/2000

testing massive CSS

20K A elementsno style: controltag:

A {}

class: .a00001 {}.a20000 {}

descender: DIV DIV DIV P A.a00001 {}

child: DIV > DIV > DIV > P > A.a00001 {}

http://jon.sykes.me/153/more-css-performance-testing-pt-3

CSS performance isn't linear

IE 7 "cliff" at 18K rules

real world levels of CSS#

Rules#

elementsAvg

DepthAOL 2289 1628 13

eBay 305 588 14

Facebook 2882 1966 17

Google Search 92 552 8

Live Search 376 449 12

MSN.com 1038 886 11

MySpace 932 444 9

Wikipedia 795 1333 10

Yahoo! 800 564 13

YouTube 821 817 9

average 1033 923 12

testing typical CSS

"costly"selectors aren't always costly (at typical levels)

are these selectors "costly"?DIV DIV DIV P A.class0007 { ... }

1K rules (vs. 20K)same amount of CSS

in all test pages30 ms avg delta

http://www.stevesouders.com/blog/2009/03/10/performance-impact-of-css-selectors/

testing expensive selectors

1K rules (vs. 20K)same amount of CSS

in all test pages2126 ms avg delta!

truly expensive selectorA.class0007 * { ... }

compare to:DIV DIV DIV P A.class0007 { ... }

the key is the key selector – the rightmost argument

selectors to avoid

A.class0007 DIV { ... }

#id0007 > A { ... }

.class0007 [href] { ... }

DIV:first-child { ... }

reflow time vs. load time

reflow – time to apply CSS, re-layout elements, and repaint

triggered by DHTML:elem.className = "newclass";elem.style.cssText = "color: red";elem.style.padding = "8px";elem.style.display = "";

reflow can happen multiple times for long-lasting Web 2.0 apps

reflow time by browser

reflow performance varies by browser and action

"1x" is 1-6 seconds depending on browser (1K rules)

DHTML action Chr1 Chr2 FF2 FF3IE6,7 IE 8 Op Saf3 Saf4

className 1x 1x 1x 1x 1x 1x 1x 1x 1x

display none - - - - 1x - - - -

display default

1x 1x 1x 2x 1x 1x - 1x 1x

visibility hidden

1x 1x 1x 1x 1x 1x - 1x 1x

visibility visible

1x 1x 1x 1x 1x 1x - 1x 1x

padding - - 1x 2x 4x 4x - - -

width length - - 1x 2x 1x 1x - 1x -

width percent - - 1x 2x 1x 1x - 1x -

width default 1x - 1x 2x 1x 1x - 1x -

background - - 1x 1x 1x - - - -

font-size 1x 1x 1x 2x 1x 1x - 1x 1x

Simplifying CSS Selectors

efficient CSS comes at a cost – page weight

focus optimization on selectors where the key selector matches many elements

reduce the number of selectors

1. long HTML doc response

2. flush (good)

3. inline script blocks .jsvar pageName='HomePagePortal';

4. scripts block

5. ads .js non-blocking (good)

6. 26 bg images – no sprites

7. sharded domains – pics & rtm (good)

8. compress images by 20%

9. thumbs load slowly – HTTP/1.0?

10. remove ETags (?)

11. ~40 inefficient CSS selectors.playgrnd * {}

1

9

76

54

32

4

4 4

8

www.ebay.com

1. long HTML doc response

2. flush (good)

3. inline script blocks .js

4. scripts block

5. 44 bg images – use sprites

6. sharded domains

7. compress images by 20%

8. thumbs load slowly – HTTP/1.0?

9. remove ETags (?)

10. ~40 inefficient CSS selectors.playgrnd * {}

1

8

65

432

7

shop.ebay.com/ items/Nike

http://cgi.ebay.com/ws/eBayISAPI.dll?

ViewItemNext&Item=2603809159111. long HTML doc response

2. flush (good)

3. put stylesheets above scripts

4. inline script blocks .js

5. scripts block, concatenate scripts, minify

6. sharded domains – pics & rtm (good)

7. remove ETags (?)

8. ~15 inefficient CSS selectors.overlay .contentbox * {}

1

6

5

324

5

3

takeaways

focus on the frontend

run YSlow: http://developer.yahoo.com/yslow

speed matters

impact on revenue

Google:

Yahoo:

Amazon:

1 http://home.blarg.net/~glinden/StanfordDataMining.2006-11-29.ppt2 http://www.slideshare.net/stoyan/yslow-20-presentation

+500 ms -20% traffic1

+400 ms -5-9% full-page traffic2

+100 ms -1% sales1

cost savings

hardware – reduced load

bandwidth – reduced response size

http://billwscott.com/share/presentations/2008/stanford/HPWP-RealWorld.pdf

if you want better user experience more revenue reduced operating expenses

the strategy is clear

Even Faster Web Sites

Steve Souderssouders@google.com

http://stevesouders.com/docs/ebay-20090416.ppt