媒體存儲技術(shù)-4存儲基礎(chǔ)iscsi(編輯修改稿)
2025-03-26 18:16 本頁面
【文章內(nèi)容簡介】 N UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future FCSAN vs. iSCSI ? Since the iSCSI appliance attaches to the existing Ether work, NAS and iSCSI are very similar in work architecture – However, the performance would be significantly different. ? Both iSCSI and SAN use Block I/O to transport data, whereas NAS uses File I/O. ? SAN offers better performance (. NAS), but is more expensive and requires a higher skill set to implement. iSCSI and NAS offer better pricing and skills may already be in place to implement them. ? Both SAN and iSCSI offer the performance benefit of Block I/O. C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future iSCSI: Performance Factors ?Workload characteristics – Sequential streaming vs random access – Read/write, large/small transfers ?Network characteristics – Speed (100, 1000, 10000 Mbps) – Distance (LAN, MAN, WAN) – Error rates – Congestion C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future iSCSI: Performance Metrics ?Bandwidth utilization – high is desirable ?CPU utilization – low is desirable ?Latency – low is desirable ?Transaction rate – high is desirable C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future iSCSI: Performance ?Numerous studies done, many more to do ?Many, many tunable parameters at all levels – SCSI – iSCSI – TCP – Ether ?Interactions/tradeoffs within/between levels ?Dynamic parameter adjustment C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future Memory a Critical iSCSI Resource ?Initiator Paging to an iSCSI disk – VM system MUST NOT block for memory – Without care, standard TCP stack will block for memory (buffers and control structures) – Without care, iSCSI data path will block for memory ?Target memory starvation – May get multiple mands at once – Must hold memory until receipt acknowledged – Acknowledgement may be delayed indefinitely C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future iSCSI: CPU Load ?CPU utilization is not negligible – Biggest percent from TCP/IP, not iSCSI or SCSI –Standard TOE offloading helps output – iSCSI HBA offloading helps input and output – Software iSCSI CRC is expensive for large PDUs C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future CPU Overhead without HBA ?Interrupt rate – 1500 byte frame every 12 microsecs on 1 GE – 9000 byte frame every 5 microsecs on 10 GE ?Frequent cache flushing ?Extra copying – TOEs help mainly on output – Input requires intermediate TCP buffers or costly memory mapping C OMMUNICAITON UNIVERSITY of CHINA STORAGY LABORATORY iSCSI: Past, Present, Future iSER: iSCSI Extensions for RDMA ?Interface between iSCSI and RDM
數(shù)學(xué)相關(guān)推薦
鄂ICP備17016276號-1