隨著科技日新月異的發展，快閃記憶體儲存裝置有取代硬碟成為虛擬儲存的主要裝置的潛力。快閃記憶體儲存產品像是固態硬碟，已逐漸在市場上流行。固態硬碟相對於傳統硬碟來說有更快速的讀寫速度及較高的平行處理能力，使得固態硬碟成為採購儲存裝置時的第一選擇。如果虛擬儲存裝置的供應商們都能夠採用固態硬碟作為虛擬儲存伺服器的儲存裝置，將能大幅提升讀寫效能。儘管固態硬碟有許多好處，但會遭受到垃圾回收問題的影響。垃圾回收問題使得固態硬碟很難在虛擬儲存的環境下達到效能獨立性，因此，固態硬碟不被虛擬儲存供應商們採用。
傳統的虛擬儲存方法無法解決效能獨立性的問題，直到VSSD的出現。VSSD是一個虛擬固態硬碟的框架，具有效能獨立性的保證。VSSD展示出固態硬碟也可以在虛擬儲存的環境下達到效能獨立性。VSSD使用兩種機制來達成效能獨立性， 1. 以塊為單位的分配 : 將使用者資料以塊為單位分開，在同一塊內的資料將屬於同個使用者。2.自我垃圾回收策略 : 這項策略使得使用者只能挑選自己的塊來進行垃圾回收。這兩項機制使得VSSD可以達到良好的效能獨立性，但也導致了低的空間利用率。低的空間利用率將使得VSSD無法充分利用到固態硬碟的效能，呈現出比單獨使用固態硬碟還要更糟糕的效能。
本篇論文著重在VSSD低的空間利用率並提出新的解決辦法。我們實作一套儲存空間借取的方法在VSSD框架上，並表現出比原本的VSSD還要好的效能，同時維持使用者的服務水平目標及效能獨立性的保證。

With updating development of modern technology, the flash-based storage device has a great potential to replace hard-disk-drive (HDD) as next generation primary storage device in virtual storage server. Flash-based storage products such as solid-state-drives (SSDs) are popular in market now. SSDs have faster read/write latency and higher parallelism compare to HDDs, these make SSD to be a first choose when buying storage device. If virtual storage vendors adopt flash-based storage device as primary storage device in virtual storage server, it can easily perform better than before, both on read/write performance. Although flash-based storage device has so many advantages, but it still suffered from garbage-collection (GC) problem. The GC problem makes flash-based storage hard to achieve performance isolation in virtual storage environment, therefore, not be adopted by virtual storage vendors.
The traditional virtual storage method is not able to solve performance isolation problem on flash-based storage, until VSSD has been proposed. VSSD is a virtual SSD framework which with performance isolation guarantee. VSSD demonstrates that SSD can also achieve performance isolation in virtual storage environment. VSSD enable two major methods to achieve performance isolation, 1. Block unit allocation: separate user data in block unit, pages in the same block belong to same user. 2. Local GC policy: This policy only allow user to select self-data block to execute GC. These two methods make VSSD to achieve good performance isolation, but result in low space utilization. Low space utilization makes VSSD not fully use SSD performance, and leads to worse performance than SSD should achieve.
This paper is focusing on the low space utilization problem of VSSD and proposing a new method to solve it. We implement a storage lending method in VSSD framework and perform better performance than origin VSSD, and still achieve user SLO with performance isolation guarantee.

[1] N. Agrawal, V. Prabhakaran, T. Wobber, J. Davis, M. Manasse, and R. Panigrahy, “Design tradeoffs for SSD performance,” in USENIX Annual Technical Conference, 2008, pp. 57–70.
[2] A. Gupta, Y. Kim, and B. Urgaonkar. DFTL: A Flash Translation Layer Employing Demand-based Selective Caching of Page-level Address MUserings. In ASPLOS’09, Washington, DC, USA, March 2009.
[3] Intel Corporation and Seagate Technology. Serial ATA Native Command Queuing. July, 2003
[4] Microsoft, SSD Extension for DiskSim Simulation Environment
http://research.microsoft.com/en-us/downloads/b41019e2-1d2b-44d8-b512-ba35ab814cd4/
[5] M. Mesnier. Intel open storage toolkit.
http://www.sourceforge.org/projects/intel-iscsi.
[6] SPC Trace File Format Specification
http://traces.cs.umass.edu/index.php/Storage/Storage
[7] M. Wachs, M. Abd-El-Malek, E. Thereska, and G. R. Ganger. Argon: performance insulation for shared storage servers. In FAST, 2007.
[8] A. Povzner et al. Efficient guaranteed disk request scheduling with Fahrrad. SIGOPS Oper. Syst. Rev., 42(4):13–25, 2008
[9] D. Skourtis, S. Kato, and S. Brandt. "QBox: Guaranteeing I/O Performance on Black Box Storage Systems", In Proceedings of the 21st ACM Symposium on High-Performance Parallel and Distributed Computing (HPDC'12), pp. 73-84, 2012.
[10] S. W. Lee, D. J. Park, T. S. Chung, D. H. Lee, S. Park, and H. J. Song, “A Log Buffer-Based Flash Translation Layer Using Fully-Associative Sector Translation,” ACM Trans. Embedded Computing Systems, vol. 6, no. 3, pp. 436-453, 2007.
[11] D. Jung, J.-U. Kang, H. Jo, J.-S. Kim, and J. Lee, “Superblock FTL: A superblock-based flash translation layer with a hybrid address translation scheme,” ACM Trans. Embedded Computing Systems, vol. 9, no. 4, pp. 1-41, 2010.
[12] M.-L. Chiang, P. C. H. Lee, and R. C. Chang, “Cleaning Policies in Mobile Computers Using Flash Memory,” Journal of Systems and Software, vol. 48, no. 3, pp. 213-231, 1999.
[13] M. Chiang, P. Lee, and R. Chang, “Using data clustering to improve cleaning performance for flash memory,” Software-Practice and Experience, vol. 29, no. 3, pp. 267-290, 1999.
[14] Y. Oh, J. Choi, D. Lee, and SH. Noh, “Cache less for better performance: Balancing cache size and update cost of flash memory cache in hybrid storage system”, in 10th USENIX Conference on File and Storage Technologies (FAST’12).
[15] H. Kwon, E. Kim, J. Choi, D. Lee, and SH. Noh, “Janus-FTL: finding the optimal point on spectrum between page and block mapping”, In Proceedings of the 10th ACM international conference on Embedded software (EMSOFT), pp. 169-178
[16] H. Kwon, E. Kim, J. Choi, D. Lee, and SH. Noh, ”Hybrid Solid State Drives for Improved Performance and Enhanced Lifetime”, In Proceedings of the 29th IEEE symposium on Massive Systems and Technologies (MSST2013)
[17] Wei-Jian Su, “VSSD: a Virtual SSD Framework with Performance Isolation Guarantee”
[18] J.R. Douceur, W.J. Bolosky, “ A large-scale study of file-system contents”, In Proceedings of the 1999 ACM SIGMETRICS international conference on Measurement and modeling of computer systems, pp. 59-70
[19] H. Huang, W. Hung, and K.G. shin, “FS2: dynamic data replication in free disk space for improving disk performance and energy consumption”, In Proceedings of 20th ACM symposium on Operating systems principles (SOSP), pp. 263-276
[20] N. Agrawal, W.J. Bolosky, J.R. Douceur, and J.R. Lorch, “A five-years study of file-system metadata”, ACM Transactions on Storage (TOS), Volume 3, Issue 3, No. 9, October 2007
[21] Avoiding the hidden costs of the cloud
http://www.symantec.com/content/en/us/about/media/pdfs/b-state-of-cloud-global-results-2013.en-us.pdf
[22] Real-world data storage utilization
http://searchstorage.techtarget.com/magazineContent/Real-world-data-storage-utilization
[23] Maximizing your enterprise data storage capacity: Improve efficiency and utilization
http://searchstorage.techtarget.com/tutorial/Maximizing-your-enterprise-data-storage-capacity-Improve-efficiency-and-utilization
[24] Study: IT managers tired of unused storage
http://www.computerweekly.com/news/2240082563/Study-IT-managers-tired-of-unused-storage
[25] Dau-Jieu Yang “BLAS: A Block-Level and Adaptive Stripping Policy for Solid-State Drives”
[26] M. L. Chiao and D. W. Chang “ROSE: A novel flash translation layer for NAND flash memory based on hybrid address translation”, IEEE Trans. Comput., vol. 60, no. 6, pp.753 -766 2011
[27] R. MCDOUGALL, “Filebench: A prototype model based workload for file systems, work in progress”, 2004
http://www.nfsv4bat.org/Documents/nasconf/2004/filebench.pdf