高性能磁记录设备的设计与研究
摘要
近年来以计算机和网络为基础的信息产业获得了空前的发展，对信息的存储技术提出了更高的要求，其主要表现在：海量数据的存储容量；存储系统的可靠性；对不同数据存储类型的适应性；数据访问的高带宽和高保密；以及提供有效的数据存储解决方案。虽然磁盘的制造工艺和磁头技术的改进，磁盘性能已经有了很大的提高，但是磁盘作为机电部件，本身受到机械执行的惯性的制约，其I／O性能始终成为提升的瓶颈。
为了解决这一问题，出现了采用冗余技术构造的RAID(Redundant Array of Independent Disks独立磁盘冗余阵列)阵列。RAID阵列由于采用了数据分块技术，即在多个磁盘上交叉存储使得多盘可平行操作，I／O响应时间得到了改善，同时利用冗余容错技术，极大的提高了磁盘阵列的可靠性和安全性。本文深入研究了RAID设计中的多种关键技术，对Cache的管理，数据分布，数据存储的均衡，
磁盘性能监控，磁盘性能的优化等方面。讨论了各种关键技术对RAID的不同应用中的重要性，在进行RAID系统设计中，须在不同的应用环境中均衡各种RAID关键技术，从而获得最佳的性能。本文设计了一种独立与PC系统的高性能的磁盘阵列的系统解决方案。
采用SOPC(System-on-a-Programmable Chip)的结构，以FPGA(Field Programmable Gate Array可编程逻辑阵列)为核心，在其中建立NIOS的32位RISC指令集的MCU，FLASH接口，SDRAM接口，中断控制器，DMA控制器等模块。在系统硬件设计中，针对数据在RAID中的存储控制流程，建立了读写分离的数据流水线，减少存储系统的资源开销。应用了uCOS实时操作系统，针对RAID应用的不同使用环境，设计了不同的软件控制模式，很大的提高RAID系统的性能和数据的可靠性。本方案可以应用兼容PC的系统方案，这样的系统可以在民用和商用环境中应用；也可以应用自定义系统，采用自定义文件系统，加密／解密方案和文件分割方式，可大大提高数据的可
武汉理工大学硕十学位论文
靠性和保密性，这样的系统可以在金融和军用系统环境中应用。
大规模的信息存储技术方兴未艾，随着我国个行业信息话的发展对磁盘
阵列的要求会越来越大，特别是高科技产业，金融机构和国家安全部门需
要高可靠性和高安全性的磁盘存储系统。因此对研制RAID系统有深远的意
义和广阔的前景。
In recent years information industry based on computer and network gains unprecedented development and proposes a higher requirement for the technology of information storage. It is mainly embodied in the following aspects: memory capacity of mass date, reliability of memory system, adaptability to different types of data storage, high bandwidth and security of data access and providing effective solution to data memory.
Although of enhancing the performance of disk through improving on its manufacture craftwork and the technology of magnetic head, the disk, working as the part of machine electricity, is restrained by the inertia of mechanic performance, whose I/O performance has reached its wit end of promotion. In order to solve such problem, the RAID array that adopt redundancy technology appears. Data cent technology is adopted in RAID array, that means the parallel operation in many disks. The I/O respond time is improved, the same time the reliability and security of disk array is highly improved by the technology of redundancy and permitting mistakes.
The paper studies much key technology of RAID design, including the management of Cache, data distribution, the equalization of data memory, the supervision of disk performance, the optimization of disk performance and so on, discusses the importance of all kinds of key technology in different applications of RAID. In RAID system design, optimum performance is gained through equalizing diversified RAID key technology in different application environments.
The paper designs a kind of solution independent PC system with high performance disk array's system, which adopts the structure of SOPC, the core of FPGA in which the MCU of 32 RISC instruction gather of NIOS, FLASH interface, SDRAM interface, interrupt controller DMA and DMA controller etc are established. In system hardware design, data pipeline of the separation of reading and writing, which aims at data memory control-flow analysis in RAID and reduces the resource's expanse of memory system, is established. UCOS real-time operation system is applied. Different modes of software control, which aim at different usage environment and enhance RAID system performance and data dependability, are designed. The project can be applied in compatible PC system, which can be used for the public and commercial purpose and in self-defined, financial and military system.
The technology of large-scale information storage is on the ascendant. THE Requirement for disk array is coming larger and larger with the
development of information popularization in various fields, especially in high-tech industries, financial reorganization and national security department, where the reliable and highly secure disk storage system are required. So the research of RAID system contains profound value and wide future.
为了解决这一问题，出现了采用冗余技术构造的RAID(Redundant Array of Independent Disks独立磁盘冗余阵列)阵列。RAID阵列由于采用了数据分块技术，即在多个磁盘上交叉存储使得多盘可平行操作，I／O响应时间得到了改善，同时利用冗余容错技术，极大的提高了磁盘阵列的可靠性和安全性。本文深入研究了RAID设计中的多种关键技术，对Cache的管理，数据分布，数据存储的均衡，
磁盘性能监控，磁盘性能的优化等方面。讨论了各种关键技术对RAID的不同应用中的重要性，在进行RAID系统设计中，须在不同的应用环境中均衡各种RAID关键技术，从而获得最佳的性能。本文设计了一种独立与PC系统的高性能的磁盘阵列的系统解决方案。
采用SOPC(System-on-a-Programmable Chip)的结构，以FPGA(Field Programmable Gate Array可编程逻辑阵列)为核心，在其中建立NIOS的32位RISC指令集的MCU，FLASH接口，SDRAM接口，中断控制器，DMA控制器等模块。在系统硬件设计中，针对数据在RAID中的存储控制流程，建立了读写分离的数据流水线，减少存储系统的资源开销。应用了uCOS实时操作系统，针对RAID应用的不同使用环境，设计了不同的软件控制模式，很大的提高RAID系统的性能和数据的可靠性。本方案可以应用兼容PC的系统方案，这样的系统可以在民用和商用环境中应用；也可以应用自定义系统，采用自定义文件系统，加密／解密方案和文件分割方式，可大大提高数据的可
武汉理工大学硕十学位论文
靠性和保密性，这样的系统可以在金融和军用系统环境中应用。
大规模的信息存储技术方兴未艾，随着我国个行业信息话的发展对磁盘
阵列的要求会越来越大，特别是高科技产业，金融机构和国家安全部门需
要高可靠性和高安全性的磁盘存储系统。因此对研制RAID系统有深远的意
义和广阔的前景。
In recent years information industry based on computer and network gains unprecedented development and proposes a higher requirement for the technology of information storage. It is mainly embodied in the following aspects: memory capacity of mass date, reliability of memory system, adaptability to different types of data storage, high bandwidth and security of data access and providing effective solution to data memory.
Although of enhancing the performance of disk through improving on its manufacture craftwork and the technology of magnetic head, the disk, working as the part of machine electricity, is restrained by the inertia of mechanic performance, whose I/O performance has reached its wit end of promotion. In order to solve such problem, the RAID array that adopt redundancy technology appears. Data cent technology is adopted in RAID array, that means the parallel operation in many disks. The I/O respond time is improved, the same time the reliability and security of disk array is highly improved by the technology of redundancy and permitting mistakes.
The paper studies much key technology of RAID design, including the management of Cache, data distribution, the equalization of data memory, the supervision of disk performance, the optimization of disk performance and so on, discusses the importance of all kinds of key technology in different applications of RAID. In RAID system design, optimum performance is gained through equalizing diversified RAID key technology in different application environments.
The paper designs a kind of solution independent PC system with high performance disk array's system, which adopts the structure of SOPC, the core of FPGA in which the MCU of 32 RISC instruction gather of NIOS, FLASH interface, SDRAM interface, interrupt controller DMA and DMA controller etc are established. In system hardware design, data pipeline of the separation of reading and writing, which aims at data memory control-flow analysis in RAID and reduces the resource's expanse of memory system, is established. UCOS real-time operation system is applied. Different modes of software control, which aim at different usage environment and enhance RAID system performance and data dependability, are designed. The project can be applied in compatible PC system, which can be used for the public and commercial purpose and in self-defined, financial and military system.
The technology of large-scale information storage is on the ascendant. THE Requirement for disk array is coming larger and larger with the
development of information popularization in various fields, especially in high-tech industries, financial reorganization and national security department, where the reliable and highly secure disk storage system are required. So the research of RAID system contains profound value and wide future.
引文
1. Schmidt R W. Java TM Network Launching Protocol & API(JSR-56) Version 1.0. Java Software, A Division of Sun Microsystems, Inc. 1996
2. World Wide Web Consortium. Extensible Markup Language(XML).http://www.w3.org/TR/REC-xml
3. Simple Object Access Protocol(SOAP) 1_1. http://www.w3.org/TR/SOAP/
4. Web Services Definition Language(WSDL).http://www.w3.org/TR/wsdl
5. David A. Pat terson, Peter Chen, Garth Gib son and Randy Katz. In troduction to Re2dundan t Arrays of Inexpen sire Disks(RAID). Proceeding of COMPCON, 1989: 112-117
6. M. Holland and G. A. Gib son. Parity Declu stering for Con tinuous Operation in Redundant Disk Arrays. Proc. of 5th Asplos, Boston, MA, 1992: 23-25
7. Mark Holland and Garth Gib son. Parity Declu stering for Con tinuous Operation in Redundant Disk Arrays. CarnegieMellon University School of Computer Science Technical Report, CMU-CS-92-130, 1992, AD-A 254052
8. Arif Merchant and PhilipS. Vu. Design and Modeling of Clustered RAID. The 22nd Annual International Symposiumon Fault-To lerant Computing, 1992,8-10: 140-149
9. M. Holland, Garth A. Gibson and D. P. Soewoprek. Fast, On-line Failure Recovery in Redundant Disk Arrays. IEEE 23rd International Symp sium on Fault-To lerant Computing, 1993:422-431
10. JaiMenon and Dick Mattson. Comparison of Sparing Alternatives for Disk Arrays. Annual Intel Symp.on Comp. Architecture, 1992: 318-329
11. JaiMenon and Dick Mattson. Distributed Sparing in Disk Arrays. IEEE COM PCON, 1992: 410-421
12. J. Menon, J. Roche and J. Kasson. Floating Parity and Data Disk Arrays.
J. of Parallel and Distributed Computing, 1993, 17(1): 129-139
13. JaiMenon and Dick Mattson. Performance of Disk Arrays in Transaction Processing Environments.12th Int. Conf. on Distribut. Computing Systems, 1992: 302-309
14. A. L. Narasimha Reddy and D. Banerjee. Gracefully Degradable Disk Arrays. Pro2ceedings of FTCS-21, 1991, 6: 402-408
15. A. L. N. Reddy, J. Chandy and P. Banerjee. Design and Evaluation of Gracefully Degradable Disk Arrays. J. of Parallel and Distributed Computing, 1993, 17(1): 28-40
16. Daniel Stodolsky, G. Gibson and M. Holland. Parity Logging Overcoming the Small Write Problem in Redundant Disk Arrays. Proceedings of the 20th Annual International Symposium on Computer Architecture, San Diego, California, 1993: 64-75
17. JaiMenon and Jim Cortnay. The Architecture of a Fault-To lerant Cached RAID Controller. Proceedings of the 20th Annual International Symposium on Computer Architecture, San Diego, California, 1993: 76-86
18. F. Douglis and John Ousterhout. Log-Structured File Systems. Proceedings of COMPCON, 1989: 124-128
19. M. Rosenblum and J. K. Ousterhout. The Design and Implementation of a Log-Structu red File System. Proceedings of the 13th ACM Sympo slum on Operating System Principle, 1991: 1-15
20. J. K. Ousterhout and Fred Douglis. Beating the I/O Bottleneck: A Case for Log -St ructured File Systems. ACM Operating Systems Review, 1989, 23(1): 11-28
21. CaoPei: Application-Controlled File Caching and Prefetching January, 1998
22. Carsten Trinitis and wolfgang Karl: Electrical Phenomena dring Hot Swap
Event
23. D. Pattersom, G. gibson, and R. Katz: A Case for Reduldant Array of Inexpensive Disks Proceedings of the Conference on Management of Data, 1998, pp. 109-116
24. D. stodolsky, G. Gibsom, M. holland: Parity logging overcoming the small write problem in redundant disk arrays Proceedings of the 20th annual International Symposium on Computer Architecture, San Diego, CA, May 1993, pp. 64-75
25. D. Stodolsky, M. Holland, W. V. Courtright, G.A. Gibson: Parity-logging disk arrays ACM Transactions on computer Systems 12(3)(1994)206-235
26. E. Lee: Software and-Performance Issuses in the Implementation of RAID Prototype university of California Technical Report UCB/CSD90/573,1990
27. Garth Alan Gibson: Reliable, Parrel Secondary Storage University of California at Berkeley, December 1991
28. G. Gibson and D. Patterson: Desiging Disk Arrays For High Data Reliablity Journal of Rarallel and Distributed Computing, January, 1993
29. Garth A. Gibson, David F. Nagle: A Cost-effective, High-Bandwidth Storage Architecture Proceedings of the 8th Conference on Architectural Support for Programming Laguages and Operating System, 1998
30. P. Cao. Lim, etal. The Ticker Taip parallel Raid architecture Proc. 20th Annual Int'l Symposium on Computer Architecture, May 16-19 1993
31. P.M. Chen, D.A. Patterson, Maximizing Performance in a Striped Disk Array Proc, 17th Ann, Int'l Symp Computer Architecture, 1990, pp. 322-331
32. G.A. Gibson, L. Hellerstein, etal, Failure Correction Techniques for Large Disk Arrays 3rd, int'l cont on Architectural Support for Programing Languages and Operating Systems, Apr, 1999, pp. 123-132
33. R. H. Katz, G.A. Gibson: Disk System Architectures for High Performance Computing Proc. IEEE. Vol. 77, NO. 12. Dec, 1989, pp, 1842-1858
34. Garth A. Gibson, Diniel Stoolsky, Fay W. Chang etc: The Scotch Parallel Storage Systems Proceedings of the IEEE Computer Conference, March5-8,1995, San Francise
35. Hai Jin, Kai, Hwang: Stripped mirroring RAID Architecture Journal of Systems Architecture 46(2000)543-550
36. Jai Menon, Dick Mattson, and Spencer: Ng. Distributed Sparing for Improved Performance of Disk Arrays Technical Report RJ 7943, IBM, January 1991
37. J. Menon and J. Kasson: Methods for Improved Update Persormance of Disk Array Proceedings of the Hawaii International Conference on System Sciences 1992, pp. 74-83
38．祝夭龙。磁盘阵列性能和可靠性研究：[学位论文]武汉：华中理工大学，1995，
39. Mark Holland Carth A. Gibson: Parity Declustering for Continuous Operation in Redundant Disk Arrays Perceeding of the 5th Conference on Architectural Support for Programming Languages and Operating Systems
40. Peter M. Chen, Garth Gibson, Randy H. katz, and David A. Patterson: An Evaluation of Redundant Arrays of Disks Using Amdahl 5890 1993 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems
41. Y. Hu, Q. Yand, DCD-disk Caching disk: a new approach for boosting I/O performance Proceedings of the 23rd International Symposium on Computer Architecutre, 1996. pp. 169-177
42. D. Patterson, G. Gibson and R. Katz. A case for redundant arrays of
inexpensive disks, in Proc. (J) ACM S IGMOD Conf. 1988, 109～116
43. P.M. Chen, E K. Lee, G.A. Gibson, R.H. KatzandD. A. Patterson. RAID: High performance, reliable secondary storage. ACM Computing Surverys, Vo 1. 26, No. 2, 1994,145～185
44. R.H. Katz, G.A. Gibson, and D.A. Patterson. Disk system architectures for high performance computing. Proceedings of the IEEE, Vo 1.77 No. 12, Dec. 1989,1842～1858
2. World Wide Web Consortium. Extensible Markup Language(XML).http://www.w3.org/TR/REC-xml
3. Simple Object Access Protocol(SOAP) 1_1. http://www.w3.org/TR/SOAP/
4. Web Services Definition Language(WSDL).http://www.w3.org/TR/wsdl
5. David A. Pat terson, Peter Chen, Garth Gib son and Randy Katz. In troduction to Re2dundan t Arrays of Inexpen sire Disks(RAID). Proceeding of COMPCON, 1989: 112-117
6. M. Holland and G. A. Gib son. Parity Declu stering for Con tinuous Operation in Redundant Disk Arrays. Proc. of 5th Asplos, Boston, MA, 1992: 23-25
7. Mark Holland and Garth Gib son. Parity Declu stering for Con tinuous Operation in Redundant Disk Arrays. CarnegieMellon University School of Computer Science Technical Report, CMU-CS-92-130, 1992, AD-A 254052
8. Arif Merchant and PhilipS. Vu. Design and Modeling of Clustered RAID. The 22nd Annual International Symposiumon Fault-To lerant Computing, 1992,8-10: 140-149
9. M. Holland, Garth A. Gibson and D. P. Soewoprek. Fast, On-line Failure Recovery in Redundant Disk Arrays. IEEE 23rd International Symp sium on Fault-To lerant Computing, 1993:422-431
10. JaiMenon and Dick Mattson. Comparison of Sparing Alternatives for Disk Arrays. Annual Intel Symp.on Comp. Architecture, 1992: 318-329
11. JaiMenon and Dick Mattson. Distributed Sparing in Disk Arrays. IEEE COM PCON, 1992: 410-421
12. J. Menon, J. Roche and J. Kasson. Floating Parity and Data Disk Arrays.
J. of Parallel and Distributed Computing, 1993, 17(1): 129-139
13. JaiMenon and Dick Mattson. Performance of Disk Arrays in Transaction Processing Environments.12th Int. Conf. on Distribut. Computing Systems, 1992: 302-309
14. A. L. Narasimha Reddy and D. Banerjee. Gracefully Degradable Disk Arrays. Pro2ceedings of FTCS-21, 1991, 6: 402-408
15. A. L. N. Reddy, J. Chandy and P. Banerjee. Design and Evaluation of Gracefully Degradable Disk Arrays. J. of Parallel and Distributed Computing, 1993, 17(1): 28-40
16. Daniel Stodolsky, G. Gibson and M. Holland. Parity Logging Overcoming the Small Write Problem in Redundant Disk Arrays. Proceedings of the 20th Annual International Symposium on Computer Architecture, San Diego, California, 1993: 64-75
17. JaiMenon and Jim Cortnay. The Architecture of a Fault-To lerant Cached RAID Controller. Proceedings of the 20th Annual International Symposium on Computer Architecture, San Diego, California, 1993: 76-86
18. F. Douglis and John Ousterhout. Log-Structured File Systems. Proceedings of COMPCON, 1989: 124-128
19. M. Rosenblum and J. K. Ousterhout. The Design and Implementation of a Log-Structu red File System. Proceedings of the 13th ACM Sympo slum on Operating System Principle, 1991: 1-15
20. J. K. Ousterhout and Fred Douglis. Beating the I/O Bottleneck: A Case for Log -St ructured File Systems. ACM Operating Systems Review, 1989, 23(1): 11-28
21. CaoPei: Application-Controlled File Caching and Prefetching January, 1998
22. Carsten Trinitis and wolfgang Karl: Electrical Phenomena dring Hot Swap
Event
23. D. Pattersom, G. gibson, and R. Katz: A Case for Reduldant Array of Inexpensive Disks Proceedings of the Conference on Management of Data, 1998, pp. 109-116
24. D. stodolsky, G. Gibsom, M. holland: Parity logging overcoming the small write problem in redundant disk arrays Proceedings of the 20th annual International Symposium on Computer Architecture, San Diego, CA, May 1993, pp. 64-75
25. D. Stodolsky, M. Holland, W. V. Courtright, G.A. Gibson: Parity-logging disk arrays ACM Transactions on computer Systems 12(3)(1994)206-235
26. E. Lee: Software and-Performance Issuses in the Implementation of RAID Prototype university of California Technical Report UCB/CSD90/573,1990
27. Garth Alan Gibson: Reliable, Parrel Secondary Storage University of California at Berkeley, December 1991
28. G. Gibson and D. Patterson: Desiging Disk Arrays For High Data Reliablity Journal of Rarallel and Distributed Computing, January, 1993
29. Garth A. Gibson, David F. Nagle: A Cost-effective, High-Bandwidth Storage Architecture Proceedings of the 8th Conference on Architectural Support for Programming Laguages and Operating System, 1998
30. P. Cao. Lim, etal. The Ticker Taip parallel Raid architecture Proc. 20th Annual Int'l Symposium on Computer Architecture, May 16-19 1993
31. P.M. Chen, D.A. Patterson, Maximizing Performance in a Striped Disk Array Proc, 17th Ann, Int'l Symp Computer Architecture, 1990, pp. 322-331
32. G.A. Gibson, L. Hellerstein, etal, Failure Correction Techniques for Large Disk Arrays 3rd, int'l cont on Architectural Support for Programing Languages and Operating Systems, Apr, 1999, pp. 123-132
33. R. H. Katz, G.A. Gibson: Disk System Architectures for High Performance Computing Proc. IEEE. Vol. 77, NO. 12. Dec, 1989, pp, 1842-1858
34. Garth A. Gibson, Diniel Stoolsky, Fay W. Chang etc: The Scotch Parallel Storage Systems Proceedings of the IEEE Computer Conference, March5-8,1995, San Francise
35. Hai Jin, Kai, Hwang: Stripped mirroring RAID Architecture Journal of Systems Architecture 46(2000)543-550
36. Jai Menon, Dick Mattson, and Spencer: Ng. Distributed Sparing for Improved Performance of Disk Arrays Technical Report RJ 7943, IBM, January 1991
37. J. Menon and J. Kasson: Methods for Improved Update Persormance of Disk Array Proceedings of the Hawaii International Conference on System Sciences 1992, pp. 74-83
38．祝夭龙。磁盘阵列性能和可靠性研究：[学位论文]武汉：华中理工大学，1995，
39. Mark Holland Carth A. Gibson: Parity Declustering for Continuous Operation in Redundant Disk Arrays Perceeding of the 5th Conference on Architectural Support for Programming Languages and Operating Systems
40. Peter M. Chen, Garth Gibson, Randy H. katz, and David A. Patterson: An Evaluation of Redundant Arrays of Disks Using Amdahl 5890 1993 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems
41. Y. Hu, Q. Yand, DCD-disk Caching disk: a new approach for boosting I/O performance Proceedings of the 23rd International Symposium on Computer Architecutre, 1996. pp. 169-177
42. D. Patterson, G. Gibson and R. Katz. A case for redundant arrays of
inexpensive disks, in Proc. (J) ACM S IGMOD Conf. 1988, 109～116
43. P.M. Chen, E K. Lee, G.A. Gibson, R.H. KatzandD. A. Patterson. RAID: High performance, reliable secondary storage. ACM Computing Surverys, Vo 1. 26, No. 2, 1994,145～185
44. R.H. Katz, G.A. Gibson, and D.A. Patterson. Disk system architectures for high performance computing. Proceedings of the IEEE, Vo 1.77 No. 12, Dec. 1989,1842～1858