NFS sensitivity to high performance networks

This paper examines NFS sensitivity to performance characteristics of emerging networks. We adopt an unusual method of inserting controlled delays into live systems to measure sensitivity to basic network parameters. We develop a simple queuing model of an NFS server and show that it reasonably characterizes our two live systems running the SPECsfs benchmark. Using the techniques in this work, we can infer the structure of servers from published SPEC results. Our results show that NFS servers are most sensitive to processor overhead; it can be the limiting factor with even a modest number of disks. Continued reductions in processor overhead will be necessary to realize performance gains from future multigigabit networks. NFS can tolerate network latency in the regime of newer LANs and IP switches. Due to NFS’s historic high mix of small metadata operations, NFS is quite insensitive to network bandwidth. Finally, we find that the protocol enhancements in NFS version 3 tolerate high latencies better than version 2 of the protocol.

[1]  Chris J. Scheiman,et al.  LogGP: incorporating long messages into the LogP model—one step closer towards a realistic model for parallel computation , 1995, SPAA '95.

[2]  Derek McAuley,et al.  Experiences of building an ATM switch for the local area , 1994, SIGCOMM 1994.

[3]  Jeffrey C. Mogul.  Network Locality at the Scale of Processes , 1992, ACM Trans. Comput. Syst..

[4]  V. Jacobson,et al.  Congestion avoidance and control , 1988, SIGCOMM '88.

[5]  J. Mogul.  Network locality at the scale of processes , 1991, TOCS.

[6]  Richard B. Gillett.  Memory Channel Network for PCI , 1996, IEEE Micro.

[7]  Ray Jain,et al.  The art of computer systems performance analysis - techniques for experimental design, measurement, simulation, and modeling , 1991, Wiley professional computing.

[8]  Edward D. Lazowska,et al.  Quantitative system performance - computer system analysis using queueing network models , 1983, Int. CMG Conference.

[9]  Brian Wong.  Configuration and Capacity Planning for Solaris Servers , 1997 .

[10]  Joseph Pasquale,et al.  The importance of non-data touching processing overheads in TCP/IP , 1993, SIGCOMM 1993.

[11]  Bruce E. Keith,et al.  LADDIS: The Next Generation in NFS File Server Benchmarking , 1993, USENIX Summer.

[12]  James Lau,et al.  File System Design for an NFS File Server Appliance , 1994, USENIX Winter.

[13]  Murad S. Taqqu,et al.  On the Self-Similar Nature of Ethernet Traffic , 1993, SIGCOMM.

[14]  Charles L. Seitz,et al.  Myrinet: A Gigabit-per-Second Local Area Network , 1995, IEEE Micro.

[15]  Walter Willinger,et al.  On the self-similar nature of Ethernet traffic , 1993, SIGCOMM '93.

[16]  Walter Willinger,et al.  On the Self-Similar Nature of Ethernet Traffic ( extended version ) , 1995 .

[17]  James Hall,et al.  Counting the cycles: a comparative study of NFS performance over high speed networks , 1997, Proceedings of 22nd Annual Conference on Local Computer Networks.

[18]  D.E. Culler,et al.  Effects Of Communication Latency, Overhead, And Bandwidth In A Cluster Architecture , 1997, Conference Proceedings. The 24th Annual International Symposium on Computer Architecture.

[19]  Robert Tappan Morris,et al.  NFS dynamics over flow-controlled wide area networks , 1997, Proceedings of INFOCOM '97.

[20]  Carl Smith,et al.  NFS Version 3: Design and Implementation , 1994, USENIX Summer.

[21]  Riccardo Gusella,et al.  A measurement study of diskless workstation traffic on an Ethernet , 1990, IEEE Trans. Commun..

[22]  Ramesh Subramonian,et al.  LogP: towards a realistic model of parallel computation , 1993, PPOPP '93.

[23]  Robert W. Horst.  TNet: A Reliable System Area Network , 1995, IEEE Micro.