Open-sourcing Katran, a scalable network load balancer
Open-sourcing Katran, a scalable network load balancer Katran creates a software-based solution to load balancing with a reengineered forwarding plane that takes advantage of recent innovations in kernel engineering. With billions of people around the globe using Facebook services, our infrastructure engineers have created a range of systems to optimize traffic and to enable fast, reliable access
Building a Billion User Load Balancer | USENIX
Want to learn how Facebook scales their load balancing infrastructure to support more than 1.3 billion users? We will be revealing the technologies and methods we use to global route and balance Facebook's traffic. The traffic team at Facebook has built several systems for managing and balancing our site traffic, including both a DNS load balancer and a software load balancer capable of handling s
Facebookã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ - eagle_raptorã®æ—¥è¨˜
ã“ã‚Œã¯http://www.adventar.org/calendars/440ã®ï¼’日目ã®è¨˜äº‹ã§ã™ï¼Ž ã¯ã˜ã‚㫠本エントリã§ã¯ï¼Œæœ€è¿‘ã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿NWã®ãƒˆãƒãƒã‚¸ã«ã¤ã„ã¦ï¼Œãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿NWã®æœ€å…ˆç«¯ã‚’èµ°ã£ã¦ã„ã‚‹facebookã®äº‹ä¾‹ã‚’ベースã«ï¼Œè«–æ–‡/blogç‰äº¤ãˆã¦ç°¡å˜ã«ç´¹ä»‹ã—ãŸã„ã¨æ€ã„ã¾ã™ï¼Ž 一般的ãªãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿NW ã¾ãšï¼Œç¾åœ¨ã®DCå‘ã‘ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ãƒˆãƒãƒã‚¸ã®åŸºæœ¬ã¨ãªã‚‹Fat-Treeã«ã¤ã„ã¦ã€‚ Fat-Treeã¯ï¼Œä¸€èˆ¬çš„ãªãƒ„リーã§ãƒœãƒˆãƒ«ãƒãƒƒã‚¯ã¨ãªã‚‹ä¸Šä½éšŽå±¤ã®ãƒªãƒ³ã‚¯å¸¯åŸŸã‚’太ã/多é‡åŒ–ã—ãŸãƒ„リー構æˆã‚’æŒã¤NWトãƒãƒã‚¸ã§ã™ã€‚ å„段ã®ã‚¹ã‚¤ãƒƒãƒã®ä¸Šæµ/下æµãŒåŒã˜å¸¯åŸŸå¹…ã‚’æä¾›ã™ã‚‹ã‚ˆã†ã«æ§‹æˆã™ã‚‹ã“ã¨ã§ã€Fat-Treeã«æŽ¥ç¶šã™ã‚‹ãƒŽãƒ¼ãƒ‰ãŒå…¨ä½“å…¨ã®é€šä¿¡ã‚’一斉ã«è¡Œã†æœ€æ‚ªã®ã‚±ãƒ¼ã‚¹ã«ãŠã„ã¦ã‚‚ã€ãƒªãƒ³ã‚¯å¸¯åŸŸã®é£Ÿã„åˆã„を発生ã•ã›ã‚‹ã“ã¨ãªãã€ãƒŽãƒ¼ãƒ‰é–“通信ã®å¸¯åŸŸã‚’ä¿è¨¼ã™ã‚‹ã“ã¨ãŒå‡ºæ¥ã¾ã™ã€‚ ã“ã®ã‚ˆã†ã«æ§‹æˆã•ã‚ŒãŸãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®
Semi-Synchronous Replication at Facebook
After intensive testing and hack, we started using Semi-Synchronous MySQL Replication at Facebook production environments. Semi-Synchronous Replication itself was ready since MySQL 5.5 (GA was released 3.5 years ago!), but I'm pretty sure not many people have used in production so far. Here are summary of our objective, enhancements and usage patterns. If you want to hear more in depth, please fee
大è¦æ¨¡ç’°å¢ƒã§MySQLã®GTIDã‚’é©ç”¨ã—ã¦å¾—られãŸæ•™è¨“ | Yakst
MySQL 5.6ã‹ã‚‰ã®æ©Ÿèƒ½ã§ã‚ã‚‹GTIDã‚’ã€Facebookã®ç’°å¢ƒã«é©ç”¨ã—ãŸéš›ã®æµã‚Œã¨ä¸»ãªä¸å…·åˆã€ãã—ã¦ãれらã®ä¿®æ£ç‚¹ã«ã¤ã„ã¦ã€Facebookã®ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã«ã‚ˆã‚‹ã¾ã¨ã‚。 by Evan Elias and Santosh Praneeth Banda Global Transaction ID (GTID)ã¯ã€MySQL 5.6ã®æ–°æ©Ÿèƒ½ã®ä¸ã§ã‚‚最も使ã‚ãšã«ã¯ã„られãªã„機能ã®ä¸€ã¤ã 。ã“ã®ãŠã‹ã’ã§ã€ãƒ•ã‚§ã‚¤ãƒ«ã‚ªãƒ¼ãƒã‚„ãƒã‚¤ãƒ³ãƒˆã‚¤ãƒ³ã‚¿ã‚¤ãƒ リカãƒãƒªã€éšŽå±¤ã‚’æŒã£ãŸãƒ¬ãƒ—リケーションãªã©ã«éžå¸¸ã«æœ‰ç›Šã ã—ã€ã‚¯ãƒ©ãƒƒã‚·ãƒ¥ã‚»ãƒ¼ãƒ•ãªãƒžãƒ«ãƒã‚¹ãƒ¬ãƒƒãƒ‰ãƒ¬ãƒ—リケーションã®å¿…é ˆæ¡ä»¶ã«ã‚‚ãªã£ã¦ã„る。ã“ã®æ•°ãƒ¶æœˆã§ã€æˆ‘々ã¯Facebookã®å…¨ã¦ã®æœ¬ç•ªç”¨MySQLインスタンスã§ã€GTIDを有効ã«ã—ãŸã€‚ãã®ä¸ã§ã€ã“ã®æ©Ÿèƒ½ã®é©ç”¨æ–¹æ³•ã‚„æ“作ã«ã¤ã„ã¦ã€ãŸãã•ã‚“ã®çŸ¥è¦‹ãŒå¾—られãŸã€‚ãŸãã•ã‚“ã®ã‚µãƒ¼ãƒã‚µã‚¤ãƒ‰ã®ä¿®æ£äº‹é …ã«ã¤ã„ã¦ã¯ã€WebScaleS
Facebookã«ãƒã‚°ã‚¤ãƒ³ | Facebook
Facebookã«ãƒã‚°ã‚¤ãƒ³ã—ã¦ã€å‹é”や家æ—ã¨å†™çœŸã‚„è¿‘æ³ã‚’シェアã—ã¾ã—ょã†ã€‚
Facebookã«ãƒã‚°ã‚¤ãƒ³
Facebookã«ãƒã‚°ã‚¤ãƒ³ã—ã¦ã€å‹é”や家æ—ã¨å†™çœŸã‚„è¿‘æ³ã‚’シェアã—ã¾ã—ょã†ã€‚
ã“ã®ãƒšãƒ¼ã‚¸ã‚’見るã«ã¯ã€ãƒã‚°ã‚¤ãƒ³ã¾ãŸã¯ç™»éŒ²ã—ã¦ãã ã•ã„
Facebookã§æŠ•ç¨¿ã‚„写真ãªã©ã‚’ãƒã‚§ãƒƒã‚¯ã§ãã¾ã™ã€‚
Facebookã«ãƒã‚°ã‚¤ãƒ³
Facebookã«ãƒã‚°ã‚¤ãƒ³ã—ã¦ã€å‹é”や家æ—ã¨å†™çœŸã‚„è¿‘æ³ã‚’シェアã—ã¾ã—ょã†ã€‚
FlashCacheã§I/Oã®ãƒœãƒˆãƒ«ãƒãƒƒã‚¯ã‚’解消ã•ã›ã‚ˆã†
ãƒ¤ãƒ•ãƒ¼æ ªå¼ä¼šç¤¾ã¯ã€2023å¹´10月1æ—¥ã«LINEãƒ¤ãƒ•ãƒ¼æ ªå¼ä¼šç¤¾ã«ãªã‚Šã¾ã—ãŸã€‚LINEãƒ¤ãƒ•ãƒ¼æ ªå¼ä¼šç¤¾ã®æ–°ã—ã„ブãƒã‚°ã¯ã“ã¡ã‚‰ã§ã™ã€‚LINEヤフー Tech Blog ã¯ã˜ã‚㫠皆様ã“ã‚“ã«ã¡ã¯ã€‚ストレージアーã‚テクトã®çŸ¢æ¾¤ç¥å¸ã§ã™ã€‚ Yahoo! JAPANã«ã¦åˆ©ç”¨ã•ã‚Œã¦ã„るマルãƒãƒšã‚¿ãƒã‚¤ãƒˆãªã‚¹ãƒˆãƒ¬ãƒ¼ã‚¸é”ã¨ä»²è‰¯ãéŽã”ã—ã¦ã„ã‚‹ç§ãªã®ã§ã™ãŒã€ä»Šå›žã¯FlashCacheを用ã„ãŸI/Oã®æœ€é©åŒ–ã«ã¤ã„ã¦æ¤œè¨¼ã—ãŸçµæžœã®ä¸€éƒ¨ã‚’皆様ã«ã”紹介ã•ã›ã¦ã„ãŸã ãã¾ã™ã€‚ FlashCacheã¨ã¯ SSD製å“ãŒå®‰ä¾¡ã«èª¿é”ã§ãるよã†ã«ãªã£ãŸæ˜¨ä»Šã§ã‚ã£ã¦ã‚‚大容é‡ãªã‚¹ãƒˆãƒ¬ãƒ¼ã‚¸ã‚’構築ã™ã‚‹éš›ã¯HDDã«é ¼ã‚‰ã–ã‚‹ã‚’å¾—ãšã€ã•ã™ãŒã«ãƒ•ãƒ«SSDã§å¤§å®¹é‡ãªã‚¹ãƒˆãƒ¬ãƒ¼ã‚¸ã‚’構築ã—ã¦ã—ã¾ã†ã¨ã‚³ã‚¹ãƒˆéŽå¤šã§ç¾å®Ÿçš„ã§ã¯ãªã„ã¨åˆ¤æ–ã•ã‚Œã‚‹æ–¹ãŒå¤§åŠã‹ã¨æ€ã„ã¾ã™ã€‚確ã‹ã«SSDを利用ã™ã‚‹ã“ã¨ã«ã‚ˆã£ã¦I/O性能ãŒèª²é¡Œã ã£ãŸã‚·ã‚¹ãƒ†ãƒ を改善ã•ã‚ŒãŸäº‹ä¾‹ã¯ãŸãã•ã‚“ã‚ã‚‹ã‚‚ã®ã®ã€å¤§
FacebookãŒå¤§è¦æ¨¡ã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティã¸ã®æŒ‘戦ã§å¦ã‚“ã ã“ã¨ï¼ˆå¾Œç·¨ï¼‰ï½žã‚ャッシュãŒæŠ±ãˆã‚‹ã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティã®å•é¡Œã¨ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿ãƒ¼ã«ã¾ãŸãŒã‚‹ä¸€è²«æ€§
FacebookãŒå¤§è¦æ¨¡ã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティã¸ã®æŒ‘戦ã§å¦ã‚“ã ã“ã¨ï¼ˆå¾Œç·¨ï¼‰ï½žã‚ャッシュãŒæŠ±ãˆã‚‹ã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティã®å•é¡Œã¨ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿ãƒ¼ã«ã¾ãŸãŒã‚‹ä¸€è²«æ€§ 全世界ã§3億人を超ãˆã‚‹ä¼šå“¡ã‚’抱ãˆã€ä¸–界最大ã®SNSã¨ãªã£ãŸFacebook。åŒç¤¾ã®æŠ€è¡“担当ãƒã‚¤ã‚¹ãƒ—レジデント Jeff Rothschildæ°ãŒã€10月8æ—¥ã«ç±³ã‚«ãƒªãƒ•ã‚©ãƒ«ãƒ‹ã‚¢å¤§å¦ã‚µãƒ³ãƒ‡ã‚£ã‚¨ã‚´æ ¡ã§è¡Œã£ãŸã‚»ãƒŸãƒŠãƒ¼ã€ŒHigh Performance at Massive Scale-Lessons learned at Facebookã€ã®å†…容をå†æ§‹æˆã—ã¦ç´¹ä»‹ã—ã¾ã™ã€‚ （ã“ã®è¨˜äº‹ã¯ã€ŒFacebookãŒå¤§è¦æ¨¡ãªã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティã¸ã®æŒ‘戦ã§å¦ã‚“ã ã“ã¨ï¼ˆå‰ç·¨ï¼‰ï½ž800å„„æžšã®å†™çœŸãƒ‡ãƒ¼ã‚¿ã¨PHPã®ã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティå•é¡Œã€ã®ç¶šãã§ã™ï¼‰ ã‚ャッシュãŒã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティã«å¤§ããªå½¹å‰²ã‚’æžœãŸã—ã¦ã„ã‚‹ Facebookã®ä¸»ãªå½¹å‰²ã¯ã€ãƒ¦ãƒ¼ã‚¶ãƒ¼ãŒç°¡å˜ã«ï¼ˆå‹äººãŸã¡ã®ï¼‰æƒ…å ±ã‚’é›†ã‚ã‚‹ã“ã¨ãŒã§
Facebookã€memcachedã«300TB以上ã®ãƒ©ã‚¤ãƒ–データを置ã大è¦æ¨¡é‹ç”¨ã®å†…å´
クラウドã®ã‚ˆã†ã«å¤§è¦æ¨¡ãªã‚·ã‚¹ãƒ†ãƒ ã§ã¯ã€ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢ã®é–‹ç™ºã¨åŒç‰ä»¥ä¸Šã«ã€å¤§è¦æ¨¡é‹ç”¨ã®å·§æ‹™ãŒã€ã‚·ã‚¹ãƒ†ãƒ 全体ã®æˆåŠŸã‚’大ããå·¦å³ã—ã¾ã™ã€‚ 6月22æ—¥ã‹ã‚‰ã€ç±³ã‚µãƒ³ã‚¿ã‚¯ãƒ©ãƒ©ã§è¡Œã‚ã‚Œã¦ã„ãŸWebサイトã®ãƒ‘フォーマンスã¨é‹ç”¨ã«é–¢ã™ã‚‹ã‚ªãƒ©ã‚¤ãƒªãƒ¼ã®ã‚¤ãƒ™ãƒ³ãƒˆã€ŒVelocity 2010ã€ã§ã€Facebookã®Technical Operations teamを担当ã™ã‚‹Tom Cookæ°ãŒã€ŒA Day in the Life of Facebook Operationsã€ï¼ˆFacebooké‹ç”¨ã®ã‚ã‚‹1日）ã¨é¡Œã—ãŸã‚»ãƒƒã‚·ãƒ§ãƒ³ã§ã€FacebookãŒãµã ã‚“ã©ã®ã‚ˆã†ãªé‹ç”¨ã‚’è¡Œã£ã¦ã„ã‚‹ã‹ã€ç´¹ä»‹ã—ã¦ã„ã¾ã™ã€‚ 世界ã§ãƒˆãƒƒãƒ—クラスã®å¤§è¦æ¨¡ã‚µã‚¤ãƒˆãŒã€æ™®æ®µã©ã®ã‚ˆã†ãªãƒ„ールを用ã„ã€ã©ã®ã‚ˆã†ãªæ–¹æ³•ã§é‹ç”¨ã—ã¦ã„ã‚‹ã®ã‹ã€ã‚»ãƒƒã‚·ãƒ§ãƒ³ã®å†…容を紹介ã—ã¾ã—ょã†ã€‚ 6å¹´ã§4億アクティブユーザーã€3カ所ã®ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿ãƒ¼ Tom Cookæ°ã€‚Facebo
Facebookã«ãƒã‚°ã‚¤ãƒ³
Facebookã«ãƒã‚°ã‚¤ãƒ³ã—ã¦ã€å‹é”や家æ—ã¨å†™çœŸã‚„è¿‘æ³ã‚’シェアã—ã¾ã—ょã†ã€‚
Facebookã€ãƒ¦ãƒ¼ã‚¶ãƒ¼ç´„70万人ã®ãƒ‹ãƒ¥ãƒ¼ã‚¹ãƒ•ã‚£ãƒ¼ãƒ‰ã‚’æ“作ã—ãŸå®Ÿé¨“çµæžœè«–文を発表
ç±³FacebookãŒç´„70万人ã®ãƒ¦ãƒ¼ã‚¶ãƒ¼ã®ãƒ‹ãƒ¥ãƒ¼ã‚¹ãƒ•ã‚£ãƒ¼ãƒ‰ã«è¡¨ç¤ºã™ã‚‹æŠ•ç¨¿ã‚’実験ã®ãŸã‚ã«æ“作ã—ãŸã“ã¨ãŒã€ç±³å›½ç§‘å¦ã‚¢ã‚«ãƒ‡ãƒŸãƒ¼ã®æ©Ÿé–¢èªŒPNASã§å…¬é–‹ã•ã‚ŒãŸè«–æ–‡ã§æ˜Žã‚‰ã‹ã«ãªã£ãŸã€‚ Facebookã®ãƒ‡ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ãƒ†ã‚£ã‚¹ãƒˆã€ã‚¢ãƒ€ãƒ ・クレイマーæ°ã‚‰3人ã«ã‚ˆã‚‹ã“ã®è«–æ–‡ã¯ã€Œã‚½ãƒ¼ã‚·ãƒ£ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã«ãŠã‘る大è¦æ¨¡æƒ…動感染ã«é–¢ã™ã‚‹å®Ÿé¨“çš„è¨¼æ‹ ï¼ˆExperimental evidence of massive-scale emotional contagion through social networks）ã€ã¨é¡Œã•ã‚Œã€Facebookã®ãƒ‹ãƒ¥ãƒ¼ã‚¹ãƒ•ã‚£ãƒ¼ãƒ‰ã«è¡¨ç¤ºã•ã‚Œã‚‹æŠ•ç¨¿ã®æ„Ÿæƒ…ãŒãƒ¦ãƒ¼ã‚¶ãƒ¼ã®æŠ•ç¨¿ã®æ„Ÿæƒ…ã«å½±éŸ¿ã™ã‚‹ã‹ã©ã†ã‹ã‚’実験ã—ãŸçµæžœãŒã¾ã¨ã‚られã¦ã„る。 ã“ã®å®Ÿé¨“ã§ã€Facebookã¯68万9003人ã®è‹±èªžã‚’使ã†ãƒ¦ãƒ¼ã‚¶ãƒ¼ã‚’抽出ã—ã€ãƒ‹ãƒ¥ãƒ¼ã‚¹ãƒ•ã‚£ãƒ¼ãƒ‰ã®ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ ã‚’æ“作ã—ã¦ã€ãƒã‚¸ãƒ†ã‚£ãƒ–ãªè¨€è‘‰ã‚’å«ã‚€æŠ•ç¨¿ã®è¡¨ç¤ºã‚’減らã—ãŸå ´åˆã®ãƒ¦
ã“ã®ãƒšãƒ¼ã‚¸ã‚’見るã«ã¯ã€ãƒã‚°ã‚¤ãƒ³ã¾ãŸã¯ç™»éŒ²ã—ã¦ãã ã•ã„
Facebookã§æŠ•ç¨¿ã‚„写真ãªã©ã‚’ãƒã‚§ãƒƒã‚¯ã§ãã¾ã™ã€‚
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
Publications - Meta Research
GitHub - jserv/facebooc: Yet another Facebook clone written in C
Connecting the World: A look inside Facebook’s Networking Infrastructure
Scaling Userspace @ Facebook
GitHub - facebookarchive/flashcache: A general purpose, write-back block cache for Linux.
{{#tags}}- {{label}}
{{/tags}}