Optimization of Data Storage Performance in Medical PACS Imaging System

doi:10.16183/j.cnki.jsjtu.2024.099

Abstract

Abstract:

Medical picture archiving and communication system (PACS) is a typical application scenario with massive small files, which faces two challenges in data storage, i.e., efficient metadata management and effective performance reduction caused by fragmentization. By analyzing various key components of the full IO (input/output) path in the medical PACS imaging system, this paper optimizes the design to achieve a significant improvement in the retrieval performance of the PACS from four dimensions, PACS software retrieval algorithm, storage protocol gateway high-concurrency design, small file aggregation, and data storage service concurrency model. The results of test show that the retrieval performance after optimizion can reach 300 images per second, which is more than three times that of traditional storage, and resolves effectively the performance problem of PACS image data retrieval.

Key words: massive small files, picture archiving and communication system (PACS) image, distributed storage, performance optimization

CLC Number:

TP391

YOU Lijue, JIAO Shengpin, LI Xiaoyong. Optimization of Data Storage Performance in Medical PACS Imaging System[J]. Journal of Shanghai Jiao Tong University, 2024, 58(10): 1629-1636.

Figures/Tables 10

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Tab.1

Tab.2

Fig.7

Fig.8

References 15

[1]	ALDOSARI H, SADDIK B, KADI A K. Impact of picture archiving and communication system (PACS) on radiology staff[J]. Informatics in Medicine Unlocked, 2018, 10: 1-16.
[2]	HAAK D, PAGE C E, REINARTZ S, et al. DICOM for clinical research: PACS-integrated electronic data capture in multi-center trials[J]. Journal of Digital Imaging, 2015, 28(5): 558-566. doi: 10.1007/s10278-015-9802-8 pmid: 26001521
[3]	SHVACHKO K, KUANG H R, RADIA S, et al. The hadoop distributed file system[C]// IEEE 26th Symposium on Mass Storage Systems and Technologies. Incline Village, USA: IEEE, 2010: 1-10.
[4]	曾梦, 邹北骥, 张文生, 等. 多模态医疗数据中海量小文件存储优化方法[J]. 软件学报, 2023, 34(3): 1451-1469.
	ZENG Meng, ZOU Beiji, ZHANG Wensheng, et al. Optimization method for storing massive small files in multi-modal medical data[J]. Journal of Software, 2023, 34(3): 1451-1469.
[5]	VORA M N. Hadoop-HBase for large-scale data[C]// Proceedings of 2011 International Conference on Computer Science and Network Technology. Harbin, China: IEEE, 2011: 601-605.
[6]	郑通, 郭卫斌, 范贵生. HDFS中海量小文件合并与预取优化方法的研究[J]. 计算机科学, 2017, 44(11A): 516-519.
	ZHENG Tong, GUO Weibin, FAN Guisheng. Research on optimization method of merging and prefetching for massive small files in HDFS[J]. Computer Science, 2017, 44(11A): 516-519.
[7]	BENDE S, SHEDGE R. Dealing with small files problem in Hadoop distributed file system[J]. Procedia Computer Science, 2016, 79: 1001-1012.
[8]	PATEL A, MEHTA M A. A novel approach for efficient handling of small files in HDFS[C]// IEEE International Advance Computing Conference. Banglore, India: IEEE, 2015: 1258-1262.
[9]	AGGARWAL R, VERMA J, SIWACH M. Small files’ problem in Hadoop: A systematic literature review[J]. Journal of King Saud University-Computer and Information Sciences, 2022, 34(10): 8658-8674.
[10]	TAO W J, ZHAI Y L, TCHAYE-KONDI J. LHF: A new archive based approach to accelerate massive small files access performance in HDFS[C]// IEEE Fifth International Conference on Big Data Computing Service and Applications. Newark, USA: IEEE, 2019: 40-48.
[11]	GAO Z P, QIN Y H, NIU K. An effective merge strategy based hierarchy for improving small file problem on HDFS[C]// 4th International Conference on Cloud Computing and Intelligence Systems. Beijing, China: IEEE, 2016: 327-331.
[12]	陈法河, 柴小丽. 基于Ceph存储系统的小文件存储优化方案[J]. 计算机系统应用, 2022, 31(2): 108-113.
	CHEN Fahe, CHAI Xiaoli. Optimization of small file storage algorithm based on Ceph storage system[J]. Computer Systems & Applications, 2022, 31(2): 108-113.
[13]	SHI A W, TIAN Z C, CHEN G, et al. Research and optimization of massive small file processing performance based on Ceph[C]// International Conference on Electronic Information Engineering and Data Processing (EIEDP 2023). Nanchang, China: SPIE, 2023: 311-315.
[14]	BEAVER D, KUMAR S, LI H C, et al. Finding a needle in Haystack: Facebook’s photo storage[C]// Proceedings of the 9th USENIX Symposium on Operating Systems Design and Implementation. Vancouver, BC, Canada: USENIX Association, 2010: 47-60.
[15]	FU S L, HE L G, HUANG C L, et al. Performance optimization for managing massive numbers of small files in distributed file systems[J]. IEEE Transactions on Parallel and Distributed Systems, 2015, 26(12): 3433-3448.

序号	CT图像信息			平均调阅时间/s
序号	数量/张	大小/MB	分辨率/ (像素×像素)	优化前	优化后
1	311	63.9	512×512	3.83	1.11
2	629	141	512×512	7.96	2.29
3	909	191	512×512	11.35	3.28
4	1539	328	512×512	18.76	5.29
5	2132	491	512×512	25.68	7.55
6	4597	939	512×512	53.45	15.47

序号	MR图像信息		平均调阅时间/s
序号	数量/张	大小/MB	优化前	优化后
1	1010	161	12.46	3.38
2	2106	222	25.37	5.57
3	3064	347	35.62	8.19