Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Performante Web-Apps [de]
Search
Christoph Lühr
February 26, 2014
Programming
1
170
Performante Web-Apps [de]
... mit Varnish & Co - getITtogether Feb. 2014 Berlin - 15min Talk; gekürzte Online Version.
Christoph Lühr
February 26, 2014
Tweet
Share
More Decks by Christoph Lühr
See All by Christoph Lühr
Vektor-Suche & LLMs
chluehr
0
87
Search, Embeddings & Vector-DBs
chluehr
0
110
Reality Check: Automated Content Production at Enterprise Scale with Pimcore
chluehr
0
61
The how and why of getting Freelancers
chluehr
0
67
Content & Master Data Management with Pimcore
chluehr
1
620
Master Data Management with Pimcore
chluehr
0
190
Swoole in 5 Minutes [en]
chluehr
1
12k
PIM & Master Data Management with Pimcore 5 [en]
chluehr
2
300
Digital Transformation & Master Data Management with Pimcore 5
chluehr
0
230
Other Decks in Programming
See All in Programming
未来を拓くAI技術〜エージェント開発とAI駆動開発〜
leveragestech
2
170
エンジニアのための”最低限いい感じ”デザイン入門
shunshobon
0
130
LLMOpsのパフォーマンスを支える技術と現場で実践した改善
po3rin
8
980
decksh - a little language for decks
ajstarks
4
21k
GitHub Copilotの全体像と活用のヒント AI駆動開発の最初の一歩
74th
8
3.1k
一人でAIプロダクトを作るための工夫 〜技術選定・開発プロセス編〜 / I want AI to work harder
rkaga
12
2.8k
Nuances on Kubernetes - RubyConf Taiwan 2025
envek
0
190
Claude Codeで実装以外の開発フロー、どこまで自動化できるか?失敗と成功
ndadayo
2
780
AI OCR API on Lambdaを Datadogで可視化してみた
nealle
0
170
Flutter로 Gemini와 MCP를 활용한 Agentic App 만들기 - 박제창 2025 I/O Extended Seoul
itsmedreamwalker
0
150
技術的負債で信頼性が限界だったWordPress運用をShifterで完全復活させた話
rvirus0817
1
2k
Claude Code と OpenAI o3 で メタデータ情報を作る
laket
0
140
Featured
See All Featured
Facilitating Awesome Meetings
lara
55
6.5k
Docker and Python
trallard
45
3.5k
Building Adaptive Systems
keathley
43
2.7k
How GitHub (no longer) Works
holman
315
140k
Learning to Love Humans: Emotional Interface Design
aarron
273
40k
Why You Should Never Use an ORM
jnunemaker
PRO
59
9.5k
Building Applications with DynamoDB
mza
96
6.6k
RailsConf 2023
tenderlove
30
1.2k
How to train your dragon (web standard)
notwaldorf
96
6.2k
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
10
1k
CSS Pre-Processors: Stylus, Less & Sass
bermonpainter
358
30k
Rebuilding a faster, lazier Slack
samanthasiow
83
9.1k
Transcript
Performante Web Apps mit Varnish & Co. Christoph Lühr /
@chluehr getITtogether - Berlin 2014
Christoph Lühr Head of Development
basilicom SEIT 2000
Image source: http://www.flickr.com/photos/ethanlindsey/321225525/ CC BY 2.0
DEV
None
(#) Internet
None
Unicode Character 'HAPPY PERSON RAISING ONE HAND' (U+1F64B)
1 2 3
Performante Web-Apps
statisch dynamisch
Performance ➊ ➋ ➌
➊
Transfer-Rate
Transfer-Rate
➋
Latenz
RENDER C NETWORK D
➌
X
X
x Faktoren
CPU RAM HDD NETZ
CPU RAM HDD NETZ
t
t
CPU RAM HDD NETZ
R A M
R A M
CPU RAM HDD LAN = IO
IO
IO
RAM SSD WAN HDD LAN 0.05 μs 5 μs x100
x100 50 μs x1.000 x10 5000 μs x100.000 x100 500000 μs x10.000.000 x100
IO
CACHE
None
IO Queue + Worker
t
statisch dynamisch
statisch dynamisch
None
Reverse- Proxy
Varnish
Varnish
Varnish
Varnish
None
None
Multiple Backends
Varnish
Varnish
Varnish
Varnish
Varnish
Varnish
Partielles Caching Edge Side Inc.
Varnish
Varnish
FUBAR? Nachteile
• Content++ = RAM++ Hardwareanforderungen • !SSL Workaround: Pound Offloader
• Purge / Warm-Up Management • Konfiguration VCL
Ratiopharm? Alternativen
None
CDN Content Delivery Network
Fallstudie / Beispiel Adventskalender
None
None
None
None
App Größe 2.6 MB
HTTP Requests 16 Req/App
None
None
js js js js js js js js js js
js js
Ladezeit bei >6 Mbit 1-3 s
None
Transferrate ~300 Mbit/s
Peak Transfer (max.) ( ~4000 Mbit/s )
Transfervolumen / Top-H ~500 GB
HTTP Zugriffe 260 Req/s
System: Eine VM. (Varnish, Pound, Apache/PHP)
None
None
None
CPU RAM HDD NETZ CPU RAM HDD NETZ CPU RAM
HDD NETZ
None
Christoph Lühr email:
[email protected]
twitter: @chluehr web: http://basilicom.de ! ?
basilicom