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
PyCon2014China-Zhuhai-bpm.py
Search
Zoom.Quiet
November 17, 2014
Technology
0
110
PyCon2014China-Zhuhai-bpm.py
141115 in Zhuhai
http://zoomq.qiniudn.com/CPyUG/PyCon2014China/141115zh-pm1-bpm.py.MP3
Zoom.Quiet
November 17, 2014
Tweet
Share
More Decks by Zoom.Quiet
See All by Zoom.Quiet
PyCon2014China-Zhuhai-high performance
zoomquiet
0
160
PyCon2014China-Zhuhai-meta programming
zoomquiet
1
130
PyCon2014China-Zhuhai-luna kv db
zoomquiet
0
93
PyCon2014China-Zhuhai-seed studio
zoomquiet
0
100
PyCon2014China-Zhuhai-Docker Registry Build By Python
zoomquiet
0
120
PyCon2014China-Zhuhai-jeff
zoomquiet
0
89
PyCon2014China-Zhuhai-pythonic front-end
zoomquiet
0
120
DevFest2014-Zhuhai-Polymer
zoomquiet
0
420
TEDxJLUZH MOMENT future
zoomquiet
0
380
Other Decks in Technology
See All in Technology
Bill One 開発エンジニア 紹介資料
sansan33
PRO
4
14k
Data Hubグループ 紹介資料
sansan33
PRO
0
2.2k
Okta Identity Governanceで実現する最小権限の原則 / Implementing the Principle of Least Privilege with Okta Identity Governance
tatsumin39
0
120
AI時代におけるデータの重要性 ~データマネジメントの第一歩~
ryoichi_ota
0
700
OCIjp_Oracle AI World_Recap
shinpy
1
120
OSSで50の競合と戦うためにやったこと
yamadashy
3
750
現場データから見える、開発生産性の変化コード生成AI導入・運用のリアル〜 / Changes in Development Productivity and Operational Challenges Following the Introduction of Code Generation AI
nttcom
1
400
CREが作る自己解決サイクルSlackワークフローに組み込んだAIによる社内ヘルプデスク改革 #cre_meetup
bengo4com
0
140
研究開発部メンバーの働き⽅ / Sansan R&D Profile
sansan33
PRO
3
20k
Copilot Studio ハンズオン - 生成オーケストレーションモード
tomoyasasakimskk
0
180
[2025年10月版] Databricks Data + AI Boot Camp
databricksjapan
1
230
クラウドとリアルの融合により、製造業はどう変わるのか?〜クラスメソッドの製造業への取組と共に〜
hamadakoji
0
230
Featured
See All Featured
The Web Performance Landscape in 2024 [PerfNow 2024]
tammyeverts
10
880
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
36
6.1k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
46
7.7k
Automating Front-end Workflow
addyosmani
1371
200k
実際に使うSQLの書き方 徹底解説 / pgcon21j-tutorial
soudai
PRO
190
55k
How to Ace a Technical Interview
jacobian
280
24k
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
23
1.5k
A Modern Web Designer's Workflow
chriscoyier
697
190k
Rails Girls Zürich Keynote
gr2m
95
14k
Typedesign – Prime Four
hannesfritz
42
2.8k
The Myth of the Modular Monolith - Day 2 Keynote - Rails World 2024
eileencodes
26
3.1k
A better future with KSS
kneath
239
18k
Transcript
None
芯雲流程引擎 使用Python协程解决阻塞问题
[email protected]
Service Orchestration DNS 系统 接入层 系统 告警 系统 配置管理 系统
Game Server SA 系统 IaaS 系统 官网 系统
Python with @task
阻塞问题 result = call_func() result 无法立即获得,所谓“阻塞”
I/O 阻塞 bytes = tcp_socket.recv(4096) 网络I/O操作需要时间,所谓“I/O 阻塞”
例子 连接 发送 接收
回调 def on_received(bytes): ….. def on_sent(sock): sock.recv(on_received) def on_connect(sock): sock.send(‘xxx’,
on_sent) connect(‘a.b.c.d’, on_connect) 上下文
CALLBACK HELL
Logic Locality 最小化开发人员阅读流程代码的眼球移动距离
协程可以表达I/O阻塞 def some_process(): sock = connect(‘a.b.c.d’, on_connect) sock.send(‘xxx’) bytes =
sock.recv(on_received) …. 上下文 locals()
流程阻塞 ip = res.request_new_server(model=’B6’) 调用“外部系统”需要时间,所谓“流程阻塞”
流程阻塞 yes_no = smcs.wechat_approve(‘reboot?’) 人工审批需要时间,所谓“流程阻塞”
协程也可以表达流程阻塞 def some_process(): ip = res.request_new_server(model=’B6’) yes_no = smcs.wechat_approve(‘deploy %s?’
% ip) if yes_no: ijobs.exec_job(‘deploy’, ip)
流程引擎 申请机器 部署 修改DNS 引擎 流程
各式Event Loop 连接 发送 接收 event loop 协程 gevent tulip
/ asyncio tornado evergreen eventlet
所谓Event Loop def some_process(): sock = connect(‘a.b.c.d’, on_connect) sock.send(‘xxx’) bytes
= sock.recv(on_received) …. 协程 Event Loop fd_map = {} [fd, read] => coroutine 操作系统 select kqueue epoll recv好了告诉我, 我先睡你那了 select(sock, EV_READ) fd 132 可读了 fd 132 对应的协程醒醒 开读了
协程可以存数据库的 def some_process(): ip = res.request_new_server(model=’B6’) yes_no = smcs.wechat_approve(‘deploy %s?’
% ip) if yes_no: ijobs.exec_job(‘deploy’, ip) 协程 引擎 event_map = {} event => coroutine 外部系统 审批好了告诉我, 我先睡你那了 wechat approve 审批结果有了 审批好了,协程醒醒 开工了 mysql import cPickle bytes = cPickle.dumps(coroutine)
Python协程实现 • yield • stackless python / pypy • greenlet
/ python-fibers / stacklet / libgevent
Ultimate Coroutine Scheduler • 语言:Java for true parallelism • 协程:co-routine
with bytecode weaver • I/O阻塞:包装NIO • 流程阻塞:event map,持久化可选 https://github.com/taowen/daili
协程大法好 招聘:
[email protected]