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
Concurrency for distributed Web crawlers
Search
Takumi Sueda
March 01, 2017
Technology
2
3.4k
Concurrency for distributed Web crawlers
golang.tokyo #4 @ eureka, Inc.
Takumi Sueda
March 01, 2017
Tweet
Share
More Decks by Takumi Sueda
See All by Takumi Sueda
組み込みLinuxの時系列
puhitaku
4
1.6k
電子辞書にステータスバーを実装する
puhitaku
0
230
知育おもちゃの遊び方 Lv. 999
puhitaku
5
2.9k
ルーターでプレゼンする
puhitaku
1
5.4k
AST as Black Magic
puhitaku
0
130
次世代の足下を舗装するコミュニティ
puhitaku
5
1.8k
電子辞書で学ぶ Linux のサスペンド
puhitaku
0
650
USB PD で迎える AC アダプター大統一時代
puhitaku
2
2.6k
電子辞書のアイデンティティを消す方法
puhitaku
3
4k
Other Decks in Technology
See All in Technology
【詳説】コンテンツ配信 システムの複数機能 基盤への拡張
hatena
0
210
AWSではじめる Web APIテスト実践ガイド / A practical guide to testing Web APIs on AWS
yokawasa
7
560
EMConf JP 2025 懇親会LT / EMConf JP 2025 social gathering
sugamasao
2
180
データエンジニアリング領域におけるDuckDBのユースケース
chanyou0311
8
2.1k
ExaDB-XSで利用されているExadata Exascaleについて
oracle4engineer
PRO
3
180
1行のコードから社会課題の解決へ: EMの探究、事業・技術・組織を紡ぐ実践知 / EM Conf 2025
9ma3r
8
3.3k
LINE NEWSにおけるバックエンド開発
lycorptech_jp
PRO
0
190
Two Blades, One Journey: Engineering While Managing
ohbarye
4
1.5k
PHPカンファレンス名古屋-テックリードの経験から学んだ設計の教訓
hayatokudou
2
540
ESXi で仮想化した ARM 環境で LLM を動作させてみるぞ
unnowataru
0
160
スキルだけでは満たせない、 “組織全体に”なじむオンボーディング/Onboarding that fits “throughout the organization” and cannot be satisfied by skills alone
bitkey
0
160
手を動かしてレベルアップしよう!
maruto
0
170
Featured
See All Featured
Six Lessons from altMBA
skipperchong
27
3.6k
Designing Experiences People Love
moore
140
23k
Designing on Purpose - Digital PM Summit 2013
jponch
117
7.1k
Thoughts on Productivity
jonyablonski
69
4.5k
No one is an island. Learnings from fostering a developers community.
thoeni
21
3.1k
GraphQLとの向き合い方2022年版
quramy
44
14k
Distributed Sagas: A Protocol for Coordinating Microservices
caitiem20
330
21k
The Straight Up "How To Draw Better" Workshop
denniskardys
232
140k
Art, The Web, and Tiny UX
lynnandtonic
298
20k
Design and Strategy: How to Deal with People Who Don’t "Get" Design
morganepeng
129
19k
Site-Speed That Sticks
csswizardry
4
400
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
44
7k
Transcript
$PODVSSFODZGPSEJTUSJCVUFE8FCDSBXMFST HPMBOHUPLZP!FVSFLB *OD 5BLVNJ4VFEB
)J 5BLVNJ4VFEB !QVIJUBLV 'VMMFS *OD (PMBOH1ZUIPO$ 4PMEFSJOHFNCFEEFETZTUFN 8SJUJOH(PMBOHTJODF4FQUFNCFS
5PEBZˏTUPQJD JOUFSNTPGDPODVSSFODZ (PQIFS DSFBUFECZ3FOFF'SFODI $$#: GPS8FCDSBXMFST
5PEBZˏTUPQJD 8FCDSBXMFSGPSXIBU
5PEBZˏTUPQJD
5PEBZˏTUPQJD ˖ #SJFGFYQMBOBUJPOPGUIFDSBXMFSTZTUFN ˖ )PXXFVTF(PSPVUJOFBOE$IBOOFM $SBXMNJMMJPOBQQTFWFSZEBZ $POUJOVPVTDPOWFSTBUJPOPO4MBDL ˖
1JUGBMMBOETPMVUJPO
5PEBZˏTUPQJD ˖ #SJFGFYQMBOBUJPOPGUIFDSBXMFSTZTUFN ˖ )PXXFVTF(PSPVUJOFBOE$IBOOFM $SBXMNJMMJPOBQQTFWFSZEBZ $POUJOVPVTDPOWFSTBUJPOPO4MBDL ˖
1JUGBMMBOETPMVUJPO
$IBSBDUFSJTUJDTPGDSBXMUBSHFU ˖ 4JOHMFEPNBJO /PUUPNBLF)551(&5PPEPSZPVˏMM CFCBOOFEGSPNBDDFTT ˖ .BOZ L UBSHFUBQQT ˖
7BSJBCMFDPOUFOUTˏTJ[F *ODMVEJOHJNBHFTEPXOMPBEUJNF NBZWBSZ
$IBSBDUFSJTUJDTPGDSBXMUBSHFU ˖ 4JOHMFEPNBJO /PUUPNBLF)551(&5PPEPSZPVˏMM CFCBOOFEGSPNBDDFTT ˖ .BOZ L UBSHFUBQQT ˖
7BSJBCMFDPOUFOUTˏTJ[F *ODMVEJOHJNBHFTEPXOMPBEUJNF NBZWBSZ $SBXMUIFNBMM8JUIJOIST&WFSZEBZ
*OGPSNBOU UIFEJTUSJCVUFEDSBXMFSTZTUFN JOGPSNBOU <せ>䞔㜠䲿⣘罏 鸐濼罏 ˌGSPN薉鳤龤POUIF8&#
*OGPSNBOU UIFEJTUSJCVUFEDSBXMFSTZTUFN *OGPSNBOU $SBXMFS ˖ *ODSFBTFEFDSFBTF TDBMF DSBXMFST ˖ 1VTIMJTUPGUBSHFUBQQTJOUP"84
424 ˖ 3FQPSUTUBUVTWJB"84$MPVE8BUDI BOE4MBDL ˖ %FQMPZFEPO"84&$$POUBJOFS 4FSWJDFDBOCFTDBMFEFBTJMZ ˖ 3FUSJFWFUBSHFUBQQTGSPN424 ˖ 1VUDSBXMFETDSBQFEEBUBPO %ZOBNP%#BOE4 $PNNBOEFS
1IBTF"4DIFEVMF "QQTMJTU 424 "QQTMJTU "QQTMJTU "QQTMJTU 1VTI 1VTI 1VTI $PNNBOEFS
$PNNBOEFSQVTIFTBQQTMJTUT JOUPRVFVF "84424
1IBTF#4DBMFDSBXMFST $PNNBOEFS &$*OTUBODF $SBXMFS &$*OTUBODF $SBXMFS &$*OTUBODF $SBXMFS "QQTMJTU 424
"QQTMJTU "QQTMJTU "QQTMJTU $PNNBOEFSDPVOUTOVNCFSPG BQQTUIFOTDBMFTDSBXMFST
1IBTF$$SBXMBQQT $PNNBOEFS &$*OTUBODF $SBXMFS &$*OTUBODF $SBXMFS &$*OTUBODF $SBXMFS "QQTMJTU 424
"QQTMJTU "QQTMJTU "QQTMJTU 8FC %ZOBNP%# 4 *OWPLFEDSBXMFSTCFHJODSBXMJOH SFUSJFWJOHBQQMJTUGSPNRVFVF
5IVT &BDIEJTUSJCVUFEDSBXMFSTIBWF EJFSFOUHMPCBM*1BOEDBOCF TDBMFEBSCJUSBSJMZ $BOˏUFTUJNBUFDSBXMEVSBUJPO ˖ 4JOHMFEPNBJO /PUUPNBLF)551(&5PPEPSZPVˏMM CFCBOOFEGSPNBDDFTT ˖
.BOZ L UBSHFUBQQT ˖ 7BSJBCMFDPOUFOUTˏTJ[F *ODMVEJOHJNBHFTEPXOMPBEUJNF NBZWBSZ Cleared!
5PEBZˏTUPQJD ˖ #SJFGFYQMBOBUJPOPGUIFDSBXMFSTZTUFN ˖ )PXXFVTF(PSPVUJOFBOE$IBOOFM $SBXMNJMMJPOBQQTFWFSZEBZ $POUJOVPVTDPOWFSTBUJPOPO4MBDL ˖
1JUGBMMBOETPMVUJPO
6TFDBTF Crawl million apps every day!
$SBXMNJMMJPOBQQT4DIFNBUJD $PNNBOEFS &$*OTUBODF $SBXMFS &$*OTUBODF $SBXMFS &$*OTUBODF $SBXMFS "QQTMJTU 424
"QQTMJTU "QQTMJTU "QQTMJTU 8FC %ZOBNP%# 4
$SBXMNJMMJPOBQQT8JUIPVU(PSPVUJOF )551 (&5 TDSBQF $SBXMFS 5JNF 1VU%# JUFN 1SPDFTT JNBHFT
)551 (&5 TDSBQF 1VU%# JUFN 1SPDFTTJNBHFT )551 (&5 TDSBQF "QQ "QQ 1SPDFTTUJNFPGJNBHFT EPXOMPBE DPOWFSU VQMPBE UBLFTWBSJBCMFUJNF
$SBXMNJMMJPOBQQT8JUIPVU(PSPVUJOF )551 (&5 TDSBQF $SBXMFS 5JNF 1VU%# JUFN 1SPDFTT JNBHFT
$SBXMFS $BOOPUDPOUSPMFTUJNBUFDSBXMEVSBUJPO
$SBXMNJMMJPOBQQT8JUI(PSPVUJOF $SBXMFS 5JNF $SBXMEVSBUJPOJTHVBSBOUFFECZQFSJPEJDBMMZJOWPLFE(PSPVUJOFT $SBXMFS 8BJUEFTJSFEEVSBUJPOTUSJDUMZ *OWPLFOFX(PSPVUJOFFWFSZDSBXM
$SBXMNJMMJPOBQQT$PEF CFGPSF
$SBXMNJMMJPOBQQT$PEF BGUFS
6TFDBTF Continuous conversation on Slack
4MBDLDPOWFSTBUJPO$PNNBOEFS B4MBDLCPU $PNNBOEFSQSPWJEFT$IBU0QTGFBUVSF UPXBUDIDSBXMFST NJHSBUF%#
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT TUBUVTDSBXMFS
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT TUBUVTDSBXMFS $SBXMJOH1MBZ4UPSF+BQBO "QQ4UPSF64 1SPHSFTTQFSDFOUBHFJT 4JNQMFDPOWFSTBUJPOJOSFQMZ
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT )PXBCPVUDPOUJOVPVTXBUDI
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT TUBUVTDSBXMFS $SBXMJOH1MBZ4UPSF+BQBO TUBUVTDSBXMFS $SBXMJOH1MBZ4UPSF64 TUBUVTDSBXMFS $SBXMJOH"QQ4UPSF,PSFB _NJOMBUFS_ _NJOMBUFS_
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT
4MBDLDPOWFSTBUJPO$IBU0QTFYBNQMFT 4MBDLIBTˑFEJUNFTTBHF˒GFBUVSF8IZOPUVTFJU
4MBDLDPOWFSTBUJPO *NQMFNFOUFEDPOWFSTBUJPO NFTTBHF PSJFOUFE4MBDLCPU XJUIHJUIVCDPNOMPQFTTMBDL
4MBDLDPOWFSTBUJPO%FNP
$PNNBOEHPSPVUJOF *OWPLFE $POWFSTBUJPOHPSPVUJOF 4DIFNBUJD EP" 5PLFOJ[FJOWPLFDNE 1SPHSFTT9 %FBMXJUI FWFOUBOE SFBDUJPOT
&WFOU$IBO 3FBDUJPO$IBO 3FQMZTFOEFNPKJ WJBSFBDUJPODIBO 1SPHSFTT: ̕&EJU $PNNBOEOTIFE 'JOJTIFE ̕&EJU $MPTF3FBDUJPO$IBO %FTUSPZVOTVCTDSJCF
$NEHPSPVUJOF $POOFDUCFUXFFODPOWFSTBUJPOBOEHPSPVUJOF UT 1SPHSFTT9 CPEZ 1SPHSFTT9 'JSTUSFQMZ 3FBDUJPODIBO &WFOU#SPBEDBTUFS
&WFOU$IBO UJNFTUBNQ DIBO 4FOEFNPKJ 3FBDUJPODIBO 1SPHSFTT: 4FDPOESFQMZ 3FBDUJPODIBO
4MBDLDPOWFSTBUJPO$PEF 35.FWFOUTJODPNJOHGSPN4MBDL MUFSFECZ&WFOU#SPBEDBTUFS 3FQMZ FNPKJGSPNDPNNBOE *OWPLFDPNNBOE 3FBDUJPOTGSPNDPNNBOE
4MBDLDPOWFSTBUJPO$PEF *GOP54JTSFHJTUFSFE QPTUSTUUIFOBEE54 "EE54UP&WFOU#SPBEDBTUFS &EJUNFTTBHF DPOWFSTBUJPO TJODF54SFHJTUFSFE
5PEBZˏTUPQJD ˖ #SJFGFYQMBOBUJPOPGUIFDSBXMFSTZTUFN ˖ )PXXFVTF(PSPVUJOFBOE$IBOOFM $SBXMNJMMJPOBQQTFWFSZEBZ $POUJOVPVTDPOWFSTBUJPOPO4MBDL ˖
1JUGBMMBOETPMVUJPO
1JUGBMMBOETPMVUJPO Over-consumption of TCP connection in parallelized DB operation
1JUGBMMDPOTVNF5$1DPOOFDUJPO "QQTMJTU 424 "QQTMJTU "QQTMJTU "QQTMJTU $PNNBOEFS %ZOBNP%# ,OPXOBQQUBCMF
<ˏKQGPPˏ ˏKQCBSˏ> (".&@"3$"%& (".& <ˏKQIPHFˏ ˏKQQJZPˏ> <ˏKQNJLVˏ ˏKQEBZPˏ> (".&@"%7&/563& (".&@"$5*0/ <ˏKQSJOˏ ˏKQMVLBˏ> $PNNBOEFSHFOFSBUFTBQQTMJTUGSPNBMPUPGJUFNTJOUBCMF 1BSBMMFMJ[FEXJUIHPSPVUJOFT
1JUGBMMDPOTVNF5$1DPOOFDUJPO $PNNBOEFS 26&3: 26&3: 26&3: )FBMUIZTUBUF 26&3:
1JUGBMMDPOTVNF5$1DPOOFDUJPO $PNNBOEFS 26&3: 26&3: 26&3: /PUIFBMUIZTUBUF TMPXJOUFSOFU TMPXRVFSZ FUD 26&3:
1JUGBMMDPOTVNF5$1DPOOFDUJPO $PNNBOEFS 26&3: 26&3: 26&3: 26&3: $POTVNFTBMMBWBJMBCMFDPOOFDUJPOJOXPSTUDBTF /PUIFBMUIZTUBUF TMPXJOUFSOFU TMPXRVFSZ
FUD
1JUGBMMDPOTVNF5$1DPOOFDUJPO GPS@ LFZSBOHFUBSHFU,FZT\ HPGVOD \ SFT FSS2VFSZ*UFN LFZ ^
UJNF4MFFQ EVSBUJPO $BMDVMBUFECZBOPUIFSHPSPVUJOF ^ 2VFSZ*UFN XJMMCFDBMMFEXJUINJOJNBMTQFFEDPOUSPM DPODVSSFOUDBMMTBSFOPUDPOTJEFSFE
4PMVUJPONBLFBMJNJUXJUIDIBOOFM MJNJU$IBONBLF DIBOJOU GPS@ LFZSBOHFUBSHFU,FZT\ MJNJU$IBO HPGVOD \
SFT FSS2VFSZ*UFN LFZ MJNJU$IBO ^ UJNF4MFFQ EVSBUJPO $BMDVMBUFECZBOPUIFSHPSPVUJOF ^ MJNJU$IBOCMPDLTJGMJNJU$IBOJTGVMM