Memory Efficient Loss Recovery for Hardware-based Transport in Datacenter

被引:24
作者
Lu, Yuanwei [1 ,2 ]
Chen, Guo [2 ]
Ruan, Zhenyuan [1 ,2 ]
Xiao, Wencong [2 ,3 ]
Li, Bojie [1 ,2 ]
Zhang, Jiansong [2 ]
Xiong, Yongqiang [2 ]
Cheng, Peng [2 ]
Chen, Enhong [1 ]
机构
[1] USTC, Hefei, Anhui, Peoples R China
[2] Microsoft Res, Redmond, WA 98052 USA
[3] BUAA, Beijing, Peoples R China
来源
PROCEEDINGS OF THE 2017 ASIA-PACIFIC WORKSHOP ON NETWORKING (APNET '17) | 2017年
关键词
Datacenter networks; Loss recovery; Hardware memory;
D O I
10.1145/3106989.3106993
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Limited by the small on-chip memory, hardware-based transport typically implements go-back-N loss recovery mechanism, which costs very few memory but is well-known to perform inferior even under small packet loss ratio. We present MELO, an efficient selective retransmission mechanism for hardware-based transport, which consumes only a constant small memory regardless of the number of concurrent connections. Specifically, MELO employs an architectural separation between data and meta data storage and uses a shared bits pool allocation mechanism to reduce meta data on-chip memory footprint. By only adding in average 23B extra on-chip states for each connection, MELO achieves up to 14.02x throughput while reduces 99% tail FCT by 3.11x compared with go-back-N under certain loss ratio.
引用
收藏
页码:22 / 28
页数:7
相关论文
共 16 条
[1]   Agile Co-Design for a Reconfigurable Datacenter [J].
Alkalay, Shlomi ;
Angepat, Hari ;
Caulfield, Adrian ;
Chung, Eric ;
Firestein, Oren ;
Haselman, Michael ;
Heil, Stephen ;
Holohan, Kyle ;
Humphrey, Matt ;
Juhasz, Tamas ;
Kaur, Puneet ;
Lanka, Sitaram ;
Lo, Daniel ;
Massengill, Todd ;
Ovtcharov, Kalin ;
Papamichael, Michael ;
Putnam, Andrew ;
Seera, Raja ;
Tadros, Rimon ;
Thong, Jason ;
Woods, Lisa ;
Chiou, Derek ;
Burger, Doug .
PROCEEDINGS OF THE 2016 ACM/SIGDA INTERNATIONAL SYMPOSIUM ON FIELD-PROGRAMMABLE GATE ARRAYS (FPGA'16), 2016, :15-15
[2]  
[Anonymous], 2010, SIGCOMM Comput. Commun. Rev., DOI DOI 10.1145/1851275.1851192
[3]  
[Anonymous], 2016, 2016 USENIX ANN TECH
[4]  
[Anonymous], 2010, SUPPL INF ARCH SPEC, V1
[5]  
[Anonymous], 2008, INFINIBAND ARCH
[6]  
[Anonymous], 2012, SUPPL INF ARCH SPEC, V1
[7]  
Arpaci-Dusseau R.H., 2014, OPERATING SYSTEMS 3, V151
[8]  
Cisco, 2015, PRIOR FLOW CONTR BUI
[9]  
Guo CX, 2015, ACM SIGCOMM COMP COM, V45, P139, DOI 10.1145/2785956.2787496
[10]  
Guo CX, 2016, PROCEEDINGS OF THE 2016 ACM CONFERENCE ON SPECIAL INTEREST GROUP ON DATA COMMUNICATION (SIGCOMM '16), P202