Linux network stack
This document outlines the process flow for receiving a packet on a network interface, passing it through various networking stacks in the kernel, and delivering it to a socket or application. Key steps include: 1) The packet is received by the network interface driver and passed to netif_receive_skb. 2) It then goes through processing such as checksum verification, filtering by iptables, and defr
TCPã¨ã‚¿ã‚¤ãƒ アウトã¨ç§ - Cybozu Inside Out | サイボウズエンジニアã®ãƒ–ãƒã‚°
本部長や副本部長もプãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã‚’（ãŸã¾ã«ï¼‰ã™ã‚‹ã“ã¨ã§æœ‰åãªã‚µã‚¤ãƒœã‚¦ã‚ºã®é‹ç”¨æœ¬éƒ¨é•·ã€å±±æœ¬æ³°å®‡ã§ã™ã€‚ 有åã˜ã‚ƒãªã„ã‹ã‚‚ã—ã‚Œã¾ã›ã‚“ãŒã€ãƒ–ãƒã‚°ã«æ›¸ã„ãŸã®ã§æœ‰åã«ãªã‚‹ã¨ã„ã†ã“ã¨ã§ã”了承ãã ã•ã„。 今回ã¯ã€å…ˆæ—¥ç™ºç”Ÿã—㟠yrmcds ã«èµ·å› ã™ã‚‹éšœå®³ã®åŽŸå› ã¨å¯¾ç–を解説ã—ã¾ã™ã€‚ yrmcds ã¨ã„ã†ã®ã¯ã€ã‚µã‚¤ãƒœã‚¦ã‚ºãŒé–‹ç™ºã—ã¦ã„ã‚‹ memcached 互æ›ã®ã‚ーãƒãƒªãƒ¥ãƒ¼ã‚¹ãƒˆãƒ¬ãƒ¼ã‚¸ã§ã™ã€‚ å•é¡Œã®ç†è§£ã®ãŸã‚ã€ã¾ãš TCP 通信ã§ã€é€šä¿¡å…ˆã®ç›¸æ‰‹ã®éšœå®³ã«ã©ã†å¯¾å¿œã™ã‚‹ã‹è§£èª¬ã—ã¾ã™ã€‚ データã®é€ä¿¡ä¸ã«ç›¸æ‰‹ãŒè½ã¡ã‚‹ã‚±ãƒ¼ã‚¹ ã“ã®ã‚±ãƒ¼ã‚¹ã¯ã•ã‚‰ã«äºŒã¤ã«åˆ†ã‹ã‚Œã¾ã™ã€‚ 相手㮠OS ã¯ç”Ÿãã¦ã„ã‚‹ãŒã€é€šä¿¡ã—ã¦ã„るプãƒã‚°ãƒ©ãƒ ãŒè½ã¡ã‚‹ã‚±ãƒ¼ã‚¹ 相手㮠OS ã”ã¨ï¼ˆã‚ã‚‹ã„ã¯ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã”ã¨ï¼‰è½ã¡ã‚‹ã‚±ãƒ¼ã‚¹ 1 㨠2 ã®é•ã„ã¯ã€å‰è€…ã®å ´åˆ RST パケットãŒè¿”ã£ã¦ãã‚‹ã®ã«å¯¾ã—ã¦ã€å¾Œè€…ã§ã¯ãªã«ã‚‚è¿”ã£ã¦ã“ãªã„点ã§ã™ã€‚後者ã®å ´åˆã€ack ã•ã‚Œãªã„
Linux ftrace TCP Retransmit Tracing
Recent posts: 29 Oct 2024 » AI Flame Graphs 22 Jul 2024 » No More Blue Fridays 24 Mar 2024 » Linux Crisis Tools 17 Mar 2024 » The Return of the Frame Pointers 10 Mar 2024 » eBPF Documentary 28 Apr 2023 » eBPF Observability Tools Are Not Security Tools 01 Mar 2023 » USENIX SREcon APAC 2022: Computing Performance: What's on the Horizon 17 Feb 2023 » USENIX SREcon APAC 2023: CFP 02 May 2022 » Brendan
Netstat -s showing (and growing) "packets pruned from receive queue" and "packets collapsed in receive queue"
we are seeing the following: [root@primary data]# netstat -s | grep buffer ; sleep 10 ; netstat -s | grep buffer 20560 packets pruned from receive queue because of socket buffer overrun 997586 packets collapsed in receive queue due to low socket buffer 20587 packets pruned from receive queue because of socket buffer overrun 998646 packets collapsed in receive queue due to low socket buffer [root@p
8.4. 一般的ãªã‚ュー/フレームæ失å•é¡Œã®è§£æ±º Red Hat Enterprise Linux 6 | Red Hat Customer Portal
パフォーマンスãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã‚¬ã‚¤ãƒ‰ 1. æ¦‚è¦ Expand section "1. 概è¦" Collapse section "1. 概è¦" 1.1. 本書をèªã‚€ã«ã‚ãŸã£ã¦ Expand section "1.1. 本書をèªã‚€ã«ã‚ãŸã£ã¦" Collapse section "1.1. 本書をèªã‚€ã«ã‚ãŸã£ã¦" 1.1.1. 対象èªè€… 1.2. リリースã®æ¦‚è¦ Expand section "1.2. リリースã®æ¦‚è¦" Collapse section "1.2. リリースã®æ¦‚è¦" 1.2.1. Red Hat Enterprise Linux 6 ã«ãŠã‘る新機能 1.2.2. 水平方å‘ã®ã‚¹ã‚±ãƒ¼ãƒ©ãƒ“リティ 1.2.3. 分散システム2. Red Hat Enterprise Linux 6 ã®ãƒ‘フォーマンス機能 Expand section "2. Red Hat Enterprise Li
8.3. パケットå—ä¿¡ã®æ¦‚è¦ | Red Hat Product Documentation
ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®ãƒœãƒˆãƒ«ãƒãƒƒã‚¯ã¨ãƒ‘フォーマンスå•é¡Œã‚’ã•ã‚‰ã«åˆ†æžã™ã‚‹ã«ã¯ã€ãƒ‘ケットå—ä¿¡ã®æ©Ÿèƒ½æ–¹æ³•ã‚’ç†è§£ã™ã‚‹å¿…è¦ãŒã‚ã‚Šã¾ã™ã€‚パケットå—ä¿¡ãŒãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ãƒ‘フォーマンスã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã§é‡è¦ãªã®ã¯ã€å—信パスã§ãƒ•ãƒ¬ãƒ¼ãƒ ãŒå¤±ã‚れるã“ã¨ãŒå¤šã„ã‹ã‚‰ã§ã™ã€‚å—信パスã§ãƒ•ãƒ¬ãƒ¼ãƒ ãŒå¤±ã‚れるã¨ã€ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ãƒ‘フォーマンスã«å¤šå¤§ãªãƒšãƒŠãƒ«ãƒ†ã‚£ãƒ¼ã‚’引ãèµ·ã“ã—ã¾ã™ã€‚
TCP Retransmit・DupACK・Fast Retransmit ～トラブルシューティング時ã®ã‚ーワード～ - troushoo
今回ã¯ã€TCP Retransmit 㨠DupACK 㨠Fast Retransmit ã®ç´¹ä»‹ã‚’è¡Œã„ã¾ã™ã€‚ ã„ãšã‚Œã‚‚ã€ãƒ‘ケットãƒã‚¹ã¨ã„ã£ãŸãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®æ€§èƒ½ãŒå‡ºãªã„時ã«å‡ºç¾ã™ã‚‹ã‚ーワードã§ã™ã€‚ ã€TCP Retransmit】 Retransmit ã¨ã¯â€å†é€â€ã‚’æ„味ã™ã‚‹è‹±å˜èªžã§ã™ã€‚ TCPã§ã¯ã€TCPデータã®é€ä¿¡è€…ãŒã€å—信者ã‹ã‚‰ACKã‚’å—ã‘å–ã‚Œãªã‹ã£ãŸå ´åˆã€TCPデータã®å†é€(=Retransmit)ã‚’è¡Œã„ã¾ã™ã€‚ ã“ã‚ŒãŒTCP Retransmitã§ã™ã€‚ TCP Retrasmitã®ç™ºç”Ÿã¯ã€é€ä¿¡è€…ã¨å—信者ã®é–“ã§ã®ãƒ‘ケットãƒã‚¹ã®å¯èƒ½æ€§ã‚’示唆ã—ã¦ã„ã¾ã™ã€‚ 以下ã®ä¾‹ã¯ã€RetransmitãŒèµ·ã“ã£ã¦ã„ã‚‹ã¨ãã®ãƒ‘ケットã‚ャプãƒãƒ£ã§ã™ã€‚ â‘ ã§ã€192.168.122.129㸠â€HTTP/1.1 200 OK†ã¨ã„ã†ãƒ‡ãƒ¼ã‚¿ã‚’é€ã£ã¦ã„ã¾ã™ã€‚ â‘¡ã®é–“ã€192.168.122.129ã¸ACKãŒå—ä¿¡
SACSIS.TCP - HPCユーザãŒçŸ¥ã£ã¦ãŠããŸã„TCP/IPã®è©±
HPC TCP/IP 2009 5 29 SACSIS2009 HPC TCP/IP 2 • • 1 Gbps RTT 100 • 240 Mbps ☹ • 940 Mbps ☺ (2) HPC TCP/IP • Ethernet PC • TCP/IP 2 10MB 3 0 200 400 600 800 1000 0 100 200 300 400 500 600 Bandwidth (Mbps) Time (msec) 160 0 200 400 600 800 1000 0 100 200 300 400 500 600 Bandwidth (Mbps) Time (msec) x3.5 (2) • TCP/IP • • MPI • TCP/IP TCP • TCP Reno [RFC2851] • TCP CUBIC Compound TCP 4 (3) • TCP/IP • T
ã¡ã‚‡ã£ã¨ jot
linux ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚»ã‚ュリティã®ã“ã¨ã¯ã˜ã‚。ã¿ãŸã„ãªãƒšãƒ¼ã‚¸ã« /proc/sys/net/conf/eth#/rpfilter 㯠1 ã«ã—ã¾ã—ょã†ã€‚らã—ã事ãŒæ›¸ã‹ã‚Œã¦ãŠã‚Šã€ç›²ç›®çš„ã«ãーゆーモン。ã¨ä¿¡ã˜ã¦ã„ãŸãŒ LSRR 試ã—始ã‚ã¦ãµã¨æ°—ã«ãªã£ãŸã€‚ 本æ¥ãªã‚‰ã° net/ipv4/fib_frontend.c ã® fib_validate_source() をフムフムã¨èªã¿ãŸã„ã¨ã“ã‚ãªã®ã ãŒ... Understanding Linux Network Internals ã®åŠ›ã‚’借りる始末。曰ãå…¥ã£ã¦ããŸãƒ‡ãƒã‚¤ã‚¹ã®çµŒè·¯ã‚’調ã¹ã¦ã€ãã®ãƒ‡ãƒã‚¤ã‚¹ã‚’経由ã—ãŸåˆ°é”ã®å¯å¦ã‚’åˆ¤æ– --- Reverse Path Filter --- ã¨ã„ã†ã“ã¨ã‚‰ã—ã„。淋ã—ã„自身ã®çµŒè·¯ã‚’示ã™ã¨# ip route ls 172.27.202.0/24 via 172.27.129.1 dev eth0 proto
ルーティングテーブル - Linuxカーãƒãƒ«ãƒ¡ãƒ¢
IPv4ã®ãƒ«ãƒ¼ãƒ†ã‚£ãƒ³ã‚°ãƒ†ãƒ¼ãƒ–ルã«é–¢ã™ã‚‹ãƒ¡ãƒ¢ 1. æ¦‚è¦ Linuxã§ã¯ãƒ«ãƒ¼ãƒ†ã‚£ãƒ³ã‚°ãƒ†ãƒ¼ãƒ–ルã®ã“ã¨ã‚’フォワーディングインフォメーションベース(FIB)ã¨å‘¼ã¶(以下ã§ãƒ«ãƒ¼ãƒ†ã‚£ãƒ³ã‚°ãƒ†ãƒ¼ãƒ–ルã¨å‘¼ã‚“ã å ´åˆã¯FIBã®ã“ã¨ã‚’指ã™)。ã¾ãŸã€ãƒ‘ケットã®é€å—ä¿¡ã§å‚ç…§ã•ã‚ŒãŸãƒ«ãƒ¼ãƒ†ã‚£ãƒ³ã‚°ã‚¨ãƒ³ãƒˆãƒªã¯ãƒ«ãƒ¼ãƒˆã‚ャッシュã«ã‚ャッシュã—ã¦æ¬¡ã‹ã‚‰ç´ æ—©ãルーティングエントリを検索ã§ãるよã†ã«ãªã£ã¦ã„る。 2. フォワーディングインフォメーションベース 2.1 FIBã®æ§‹é€ FIBã®æ§‹é€ を図1ã«ç¤ºã™ã€‚ FIBã¯ãƒãƒƒã‚·ãƒ¥æ§‹é€ ã‚’ã—ã¦ã„ã‚‹(*1)。 ルーティングエントリã¯å®›å…ˆã®ãƒ—レフィクス長(サブãƒãƒƒãƒˆãƒžã‚¹ã‚¯é•·)毎ã«åˆ†ã‘ã¦ç®¡ç†ã•ã‚Œã‚‹ã€‚å„プレフィクス長(0〜32)毎ã«Zone(struct fn_zone)ãŒå˜åœ¨ã—ã€ãƒ—レフィクス長(Zone)毎ã«ãƒãƒƒã‚·ãƒ¥ãƒ†ãƒ¼ãƒ–ル(fz_hash)ã‚’æŒã¤ã€‚宛先ã®ãƒ—レフィクス長ãŒåŒã˜ãƒ«ãƒ¼ãƒ†ã‚£ãƒ³ã‚°ã‚¨ãƒ³ãƒˆãƒªã¯ã€
Linuxã®TCP/IPã®å®Ÿè£…ã«ã¤ã„ã¦èª¿ã¹ã¦ã¿ãŸ - nyantã®ãƒ–ãƒã‚°
çµè«– tcp_syncookies = 0 ã®å ´åˆ tcp_max_syn_backlog ã® 75% を超ãˆã‚‹ half openãªã‚³ãƒã‚¯ã‚·ãƒ§ãƒ³è¦æ±‚ãŒæ¥ãŸå ´åˆã€dropã•ã‚Œã‚‹(TCP: drop open request from [IP]/[PORT]) somaxconn 㨠tcp_max_syn_backlogã®å°ã•ã„ã»ã†ã®å€¤ã‚’ã€ä»˜è¿‘ã®2ã®ç´¯ä¹—ã®å€¤ã«åˆ‡ã‚Šä¸Šã’ãŸæ•°å€¤ã‚’超ãˆã‚‹ã¨ã€drop (TCP: Possible SYN flooding on port [PORT]. Dropping request.) tcp_syncookies = 1 ã®å ´åˆ somaxconn 㨠tcp_max_syn_backlogã®å°ã•ã„ã»ã†ã®å€¤ã‚’ã€ä»˜è¿‘ã®2ã®ç´¯ä¹—ã®å€¤ã«åˆ‡ã‚Šä¸Šã’ãŸæ•°å€¤ã‚’超ãˆã‚‹ã¨ã€cookieé€ä¿¡ (TCP: Possible SYN flooding on port [PORT]
How TCP backlog works in Linux
When an application puts a socket into LISTEN state using the listen syscall, it needs to specify a backlog for that socket. The backlog is usually described as the limit for the queue of incoming connections. Because of the 3-way handshake used by TCP, an incoming connection goes through an intermediate state SYN RECEIVED before it reaches the ESTABLISHED state and can be returned by the accept s
相手ãŒã„ãªã„ã®ã« ESTABLISHED ã«ãªã£ã¦ã‚‹ TCP ãƒãƒ¼ãƒˆ - tmtms ã®ãƒ¡ãƒ¢
最近 ParallelServer ã¨ã„ã†ãƒ©ã‚¤ãƒ–ラリを作ã£ãŸã®ã§ã™ãŒã€ãã®æœ€ä¸ã«å¥‡å¦™ãªçŠ¶æ…‹ã«ãªã£ã¦ã‚‹ TCP ãƒãƒ¼ãƒˆã‚’見ã¤ã‘ãŸã®ã§ã€ãƒ¡ãƒ¢ã£ã¦ãŠãã¾ã™ã€‚ Ruby ã§ã¯ TCP サーãƒãƒ¼ã¯æ¬¡ã®ã‚ˆã†ãªæ„Ÿã˜ã§ä½œã‚‹ã“ã¨ãŒã§ãã¾ã™ã€‚ãŠæ‰‹è»½ã§ã™ã。 require 'socket' Socket.tcp_server_loop(12345) do |socket, client_addr| socket.puts "Your IP address: #{client_addr.ip_address}" name = socket.gets socket.puts "Hello, #{name}" socket.close end ã“れ㯠12345 ãƒãƒ¼ãƒˆã§ã‚¯ãƒ©ã‚¤ã‚¢ãƒ³ãƒˆã‹ã‚‰ã®æŽ¥ç¶šã‚’å¾…ã¡ã€æŽ¥ç¶šã•ã‚ŒãŸã‚‰ã‚¯ãƒ©ã‚¤ã‚¢ãƒ³ãƒˆã®IPアドレスã¨ã‚¯ãƒ©ã‚¤ã‚¢ãƒ³ãƒˆã‹ã‚‰ã®å…¥åŠ›ã‚’クライアントã«é€ä¿¡ã—ã¦åˆ‡æ–ã™ã‚‹ã ã‘ã®ç°¡å˜ãªãƒ—ãƒã‚°ãƒ©ãƒ
ヘアピンNATルーターã¨ã¯ï¼Ÿ
自宅サーãƒãƒ¼ã‚’構築ã—ãŸã¨ãã®ç½ 自宅サーãƒãƒ¼ã‚’構築ã—ãŸã¨ãã€å½“然ãªãŒã‚‰ãƒ«ãƒ¼ã‚¿ãƒ¼ã®å†…å´ã€ã¤ã¾ã‚Šè‡ªå®…LAN上ã«ã‚µãƒ¼ãƒãƒ¼ãŒã¶ã‚‰ä¸‹ãŒã‚‹ã“ã¨ã«ãªã‚Šã¾ã™ã€‚ ãªã®ã§ã€ã“ã®ã‚µãƒ¼ãƒãƒ¼ã‚’外部ã«å…¬é–‹ã™ã‚‹ãŸã‚ã«ã¯ãƒ«ãƒ¼ã‚¿ãƒ¼ã§é™çš„NATè¨å®šã‚’è¡Œã„ã€å¤–部ã‹ã‚‰ã®ã‚¢ã‚¯ã‚»ã‚¹ã‚’サーãƒãƒ¼ã«èª˜å°Žã™ã‚‹ã‚ˆã†ã«ã—ã¾ã™ã€‚ ã“ã‚Œã¯å½“然ã€ã©ã‚“ãªãƒ«ãƒ¼ã‚¿ã§ã‚‚è¨å®šã§ãã¾ã™ã€‚ ã—ã‹ã—ãªãŒã‚‰ã€ã“ã®è‡ªå®…サーãƒãƒ¼ã«ã€ã‚µãƒ¼ãƒãƒ¼ã¨åŒã˜LAN上ã®ã‚¯ãƒ©ã‚¤ã‚¢ãƒ³ãƒˆPCã‹ã‚‰ã‚¢ã‚¯ã‚»ã‚¹ã™ã‚‹ã¨ã€å•é¡ŒãŒç™ºç”Ÿã™ã‚‹ã“ã¨ãŒã€Œå¤šã€…ã€ã‚ã‚Šã¾ã™ã€‚ ãã‚Œã¯ã€ä¾‹ãˆã°ã€è‡ªå®…サーãƒãƒ¼ã«ç‹¬è‡ªãƒ‰ãƒ¡ã‚¤ãƒ³ã‚’è¨å®šã—ホームページを開è¨ã—ãŸã¨ã—ã¾ã™ã€‚ ã“ã®ãƒ›ãƒ¼ãƒ ページã«LAN内ã®ãƒ‘ソコンã‹ã‚‰ãƒ‰ãƒ¡ã‚¤ãƒ³åã§ã‚¢ã‚¯ã‚»ã‚¹ã™ã‚‹ã¨ã€è‹¦åŠ´ã—ã¦ä½œã£ãŸãƒ›ãƒ¼ãƒ ページã§ã¯ãªãã€ãƒ«ãƒ¼ã‚¿ã®è¨å®šç”»é¢ãŒé–‹ã„ãŸã‚Šã€ç´ ã£æ°—ãªã„エラー画é¢ãŒè¡¨ç¤ºã•ã‚ŒãŸã‚Šã—ã¾ã™ã€‚ ãã®ç†ç”±ã¯ã€ãã®ãƒ«ãƒ¼ã‚¿ãƒ¼ãŒãƒ˜ã‚¢ãƒ”ンNAT（NATループãƒãƒƒã‚¯ï¼‰æ©Ÿèƒ½ã‚’æŒã£ã¦ã„ãªã„ã‹ã‚‰ãªã‚“
マルãƒã‚³ã‚¢ã¨ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚¹ã‚¿ãƒƒã‚¯ã®é«˜é€ŸåŒ–技法
7. Network stack in 4.3BSD Process(User) Process(Kernel) HW Intr Handler SW Intr Handler パケットå—ä¿¡ プãƒãƒˆã‚³ãƒ«å‡¦ç† ソケット å—ä¿¡å‡¦ç† ãƒ¦ãƒ¼ã‚¶ プãƒã‚°ãƒ©ãƒ user buffer input queue socket queue パケット システムコール プãƒã‚»ã‚¹èµ·åºŠ ソフトウェア割り込ã¿ã‚¹ã‚±ã‚¸ãƒ¥ãƒ¼ãƒ« ãƒãƒ¼ãƒ‰ã‚¦ã‚§ã‚¢å‰²ã‚Šè¾¼ã¿ ユーザ空間ã¸ã‚³ãƒ”ー 13å¹´4月23æ—¥ç«æ›œæ—¥
PF_PACKETã®ä»•çµ„ã¿ - ã‚ã‚‚ã‚ã‚‚
Software Design 2012å¹´6月å·ã®ã€Žã€Œç”Ÿã€ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã€ãŒå¦™ã«ãƒ„ボã«ãƒãƒžã£ã¦é¢ç™½ã‹ã£ãŸã®ã§ã™ãŒã€ã•ã™ãŒã«ã‚ã®ãƒšãƒ¼ã‚¸æ•°ã§PF_PACKETを語るã¨ã€ã¤ã„ã¦ã„ã‘ãªã„èªè€…も多ãã¦æ®‹å¿µã‹ã‚‚。 ã¨ã„ã†ã“ã¨ã§ã€å‚考ã¾ã§ã«æ¦‚è¦å›³ã‚’書ã„ã¦ã¿ã¾ã—ãŸã€‚ C言語ã®ã‚½ã‚±ãƒƒãƒˆãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã‚’ãªã‚“ã¨ãªã勉強ã—ãŸæ–¹ã¯å¤šã„ã¨æ€ã„ã¾ã™ã€‚ソケットã¯ã€ã‚«ãƒ¼ãƒãƒ«å†…部ã®ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ãƒ¬ã‚¤ãƒ¤ãƒ¼ã¨ãƒ¦ãƒ¼ã‚¶ç©ºé–“ã‚’çµã³ã¤ã‘るインターフェース（API）ãªã‚ã‘ã§ã™ãŒã€ã‚«ãƒ¼ãƒãƒ«å†…部ã§ã¯ã€é€šå¸¸ã€ä¸Šå›³ï¼ˆå³å´ï¼‰ã®æµã‚Œã§å—信パケットãŒå‡¦ç†ã•ã‚Œã¦ã„ãã¾ã™ã€‚ 物ç†NICãŒå—ã‘å–ã£ãŸãƒ‘ケットã¯ã€ãƒ‡ãƒã‚¤ã‚¹ãƒ‰ãƒ©ã‚¤ãƒãŒL2プãƒãƒˆã‚³ãƒ«ã®å‡¦ç†ã‚’ã—ã¦ã€ã•ã‚‰ã«L3プãƒãƒˆã‚³ãƒ«ã®ç¨®é¡žï¼ˆIPv4/IPv6ãªã©ã®é•ã„）ã®åˆ¤åˆ¥ã¾ã§è¡Œã„ã¾ã™ã€‚ 判別ã—ãŸçµæžœã«åŸºã¥ã„ã¦ã€é©åˆ‡ãªL3プãƒãƒˆã‚³ãƒ«ã®ãƒãƒ³ãƒ‰ãƒ©ãƒ¼ãŒå‘¼ã³å‡ºã•ã‚Œã¦ã€L3プãƒãƒˆã‚³ãƒ«ã®å‡¦ç†ã‚’ã—ã¦ã€ã•ã‚‰ã«L4プãƒãƒˆ
10GbEã®NICを挿ã—ã¦ã¿ãŸ
日本最強ã®çœ‹æ¿ã‚’下ã‚ã—ãŸãƒŸãƒ©ãƒ¼ã‚µãƒ¼ãƒftp.jaist.ac.jpã®ç®¡ç†è€…ã®ä¸€äººãŒã€ ã“ã®ã‚µãƒ¼ãƒã«ã¾ã¤ã‚るよã—ãªã—ã”ã¨ã‚’語りã¾ã™ã€‚ English versions of some posts on another blog. Ubuntu 12.04ã®å‰æ—¥ã«Firefox 12ãŒãƒªãƒªãƒ¼ã‚¹ã•ã‚Œã‚‹ã“ã¨ã«ã€Ubuntu 12.04ã®ãƒªãƒªãƒ¼ã‚¹ã®æ•°æ—¥å‰ã«æ°—ãŒä»˜ãã¾ã—ãŸã€‚Firefox 12ã®ãƒªãƒªãƒ¼ã‚¹ãŒ4月25æ—¥ã®æœã§ã€Ubuntu 12.04ã®ãƒªãƒªãƒ¼ã‚¹ã¯4月26æ—¥ã®22時ãらã„ã§ã™ã€‚Firefoxã®æ›´æ–°ã®åˆæ—¥ã¯ãƒ”ークã§3.7Gbpsãらã„帯域を使用ã—ã€ç¿Œæ—¥ã‚‚3Gbpsを超ãˆã‚‹ã“ã¨ãŒã‚ã‚Šã¾ã™ã€‚特ã«22時以é™ã¯ã‚¢ãƒ¡ãƒªã‚«ã®æ±æµ·å²¸ã®Firefoxã®æ›´æ–°ãŒå¢—ãˆã‚‹ã®ã§ã€ã“ã‚ŒãŒUbuntu 12.04ã®ãƒªãƒªãƒ¼ã‚¹ã«ã¶ã¤ã‹ã‚‹ã¨ftp.jaist.ac.jpã®å¸¯åŸŸã®ä¸Šé™ã®4Gbpsを超ãˆã‚‹å¯èƒ½æ€§ãŒã‚ã‚Šã¾ã™ã€‚ãã“ã§æ€¥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
collectors TCPCollector
https://nuclearcat.com/mediawiki/index.php/Intel_Gigabit_Performance
What does ECONNRESET mean in the context of an AF_LOCAL socket?
{{#tags}}- {{label}}
{{/tags}}