馬嘶的網路大道

http://www.sdnap.com/

http://fasterdata.es.net/ Internet Explorer and TCP RST - a reason to dislike http://www.extrahop.com/post/blog/good-reads/tcp-application-monitoring/ http://www.extrahop.com/post/blog/good-reads/tcp-application-monitoring-part2/ http://en.wikipedia.org/wiki/TCP_tuning

http://blog.ioshints.info/2010/08/storage-networking-is-different.html http://www.nanog.org/meetings/nanog55/presentations/Sunday/Hudson_Storage.pdf 其實就是loseless, 不過, 對我這種networking 老賊 , 就還是有笨問題 , 掉就掉, 重傳就好, HD Video這種real time, 都搞得定, IP萬歲! FC到16Gbps到頂 (和SDH到40G一樣), L1/L2已被Ethernet大一統, FC想藉FCOE延續生命, 還是被專家們唱衰(如冬瓜頭) reminder to myself data corruption will occur serious chain reaction http://www.zdnet.com/blog/storage/data-corruption-is-worse-than-you-know/191 SAN needs redundant separate fabric( no connected) + multipath software + redundant HBA SAN connection scheme is p2p SAN connectivity is p2mp ( target-initiator) SAN nodes divide into host(initiator)  and storage( target) SAN no broadcast (heavily rely on  switch intelligence) SAN forwarding is kinda routing (or TRILL)

being network Engineer, you will face tough decisions sometimes that force you to 1) stand still in technical space 2) transform into sales role  Anyway, I choose as engineer and hopefully keep improving.

http://etherealmind.com/the-only-constant-is-change-fundamentals-are-the-exception/ 讀完這篇, 真的深有同感, 一樣身為老賊, 我也花了兩年多搞IBM networking ( SDLC , LLC, DLSW), 掛了, 一整年搞ATM ( PNNI, SVC.....), 掛了, 一整年搞光網路 (SDH, WDM,,,,) , 掛了(SDH),然後花至少五年搞MPLS, Metro Ethernet, Mulitcast, 恩, 還沒掛 , 還沒算花在學modem , lease line ( T1, T3,MUX,ADM...), Frame Relay等的L1/L2的troubleshooting practice 上 還沒完, 其實這些都還有一點同質性, 更難的是另外一個世界, 偏偏搞IP(夾在中間)的一定要能和下(L1/L2, metro, mobile, optical), 和上(Application, cloud)互動, 還有, Software來了, 再把programming請回來吧