个人泛读后的理解概括

主要方向

  • 搜索引擎系统关键技术。
    主要研究搜索引擎系统相关的数据索引、数据压缩、计算和优化等关键技术。

  • 云存储系统及可靠性研究
    研究基于数据中心的云存储系统的数据布局、数据编码、数据迁移、重复数据删除和可靠性分析等内容。

  • 面向大数据的GPU计算技术
    该课题针对生物信息学、图形图像等领域的大数据处理问题，研究基于GPU的并行算法。

  • 应用数学、组合数学等领域的并行计算研究
    该课题针对相关领域的大规模计算问题，研究基于多核、众核及集群的并行计算方法。

  • 云游戏系统及其资源管理问题研究
    该课题研究云游戏系统的设计与优化，以及云游戏系统中的请求调度、游戏布局、服务器提供等资源管理问题。

  • 区块链系统和数据可靠性
    该课题面向企业需求，重点研究区块链系统基础技术、机制设计和数据可靠性等问题。

2019

Journal:

1. Communication-Aware 容器在大规模因特网数据中心放置与再分配

Liang Lv, Yuchao Zhang, Yusen Li, Ke Xu, Dan Wang, Wendong Wang, *** Li, Xuan Cao, Qingqing Liang, .Communication-Aware Container Placement and Reassignment in Large-Scale Internet Data Centers IEEE Journal on Selected Areas in Communications 37(3): 540-555 (2019) (SCI Tier 1 Journal).
容器化: 隔离，轻量级，可称量，高度便捷的特点
挑战: 将容器化的思想应用于大规模因特网数据中心里
主要内容: 容器的分配问题、容器的放置问题（采用了一种有效的最差拟合递减算法）、容器的重分配（two-stage 算法）
主要成就：通过百度的数据模拟测试，算法的评估结果比当前普遍采取的策略提升了70%，同时所有服务提升超过90%
关键词: Container communication, multi-resourse load balacne, large-scale data centers, container placement, container reassignment

2. 基于分选变换的普遍压缩策略

Bo Liu, Xi Huang, Xiaoguang Liu, Gang Wang, Ming Xu, A universal compression strategy using sorting transformation. IJCSE 18(3): 211-216 (2019)
打不开

3. 对循环MNOLS的枚举结果

F. Demirkale, D. Donovan, J.Kokkala, T. Marbach,The enumeration of cyclic MNOLS，Journal of Combinatorial Designs (2019) Volume27, Issue5May 2019,Pages 265-276
MNOLS：mutually nearly orthogonal Latin squares 相互近似正交的拉丁方阵
（这篇有点读不懂）

4. Gecko：多云存储的一种可适应分布方案

MENG YAN, JIAQI FENG, TRENT G. MARBACH, REBECCA J.STONES, GANG WANG, XIAOGUANG LIU, Gecko: A Resilient Dispersal Scheme for Multi-Cloud Storage. IEEE Access 7: 77387-77397 ， 2019
Gecko: a multi-cloud dispersal scheme, to meet the rising privacy, reliability, and verifiability needs
(a) key in a Latin-squares
(b) data files and keys are dispersed to multiple clouds
(c) a blockchain-based integrity-check protocol pinpoint faulty data
成果：

  • Enable fast and thorough key renewal
  • Key renewal is further accelerated by the blockchain-based integrity check
  • Gecko admits powerful renewal and identification of damaged data, with a minor transfer overhead.
  • Gecko performs key renewal three to five times faster than AONT_RS hybrid-slice renewal

关键词：Blockchain，data recovery，dispersal scheme，integrity check，Latin square，multi-cloud

5. 在大规模搜索引擎下的高效资源索引碎片复制

Yusen Li， Xueyan Tang，Wentong Cai， Jiancong Tong, Xiaoguang Liu, Gang Wang：Resource-Efficient Index Shard Replication in Large Scale Search Engines. IEEE Trans. Parallel Distrib. Syst. 30(12): 2820-2835 (2019 ，CCF Tier A Journal)
打不开

6. 带有Proactive Fault Tolera的云存储系统可靠性方程

Jing Li ; Peng li ; Rebecca Stones ; Gang Wang ; Zhongwei Li ; XaoGuang Liu, Reliability Equations for Cloud Storage Systems with Proactive Fault Tolera. IEEE Transactions on Dependable and Secure Computing, 2019

7. 应用于连接的CPU-GPU结构的加速史密斯-沃特曼算法

Huihui Zou, Shanjiang Tang, Ce Yu, Hao Fu, Yusen Li, Wenjie Tang:ASW: Accelerating Smith-Waterman Algorithm on Coupled CPU-GPU Architecture. International Journal of Parallel Programming 47(3): 388-402 (2019)

Conference:

8. 可持续化内存下的面向虚拟机器图像管理

Jiachen Zhang, Lixiao Cui, Peng Li, Xiaoguang Liu, Gang Wang, Towards Virtual Machine Image Management for Persistent Memory. MSST 2019:116-125. (CCF Tier B Conference)

9. 一种应用于高效数据检索的位置索引有损压缩方法

Shuni Gao, Jipeng Liu, Xiaoguang Liu, Gang Wang:A Lossy Compression Method on Positional Index for Efficient and Effective Retrieval. CIKM 2019: 2317-2320
问题提出：Incorporating proximity brings inevitable storage and computing costs by using positional data in inverted indexs.
新方法:Lossy method for compressing term position data in the case of utilizing term proximity.
成果:The method is competitive with respect to index size, ranking efficiency and effectiveness.
关键词：Query Processing，Term Proximity， Positional Index Compression

10. 云存储的硬件检测错误预测

Dongshi Liu, Bo Wang, Peng Li, Rebecca J. Stones, Trent G. Marbach, Gang Wang, Xiaoguang Liu, Zhongwei Li:Predicting Hard Drive Failures for Cloud Storage Systems. ICA3PP (1) 2019: 373-388

11. 一种基于拉丁方阵的搜索引擎

Wenxiu Fang ,Rebecca J. Stones,Trent G. Marbach,Gang Wang ,Xiaoguang Liu, Towards A Latin-Square Serach Engine, ISPA 2-10.727-735
问题提出: Latin-square has a large number of equivalent Latin squares， we need to canonical label Latin squares(规范标号）
改进方案：Nauty graph partial Latin squares (PLSs) without graph conversion
成果: faster than Nauty， and has reduced space requirements

2018年：

Journal：

12. 一种新的深度神经网络并行训练法

Shizhao Sun,Xiaoguang Liu, EC-DNN: A new method for parallel training of deep neural networks, Neurocomputing 287:118-127 (2018) (SCI Tier 2 Journal).

问题提出：deep neural networks（DNN）is a highly non-convex model（非凸模型）
新框架：EC-DNN
改进方案：

  • parameter averaging the outputs of local models weighted sum(global model)
  • 通过 model compression 减少 global model 的 size same as the local ones

成果：EC-DNN can achieve better speedup than MA-DNN without loss of accuracy，sometimes even accuracy improvement.
关键词：Parallel machine learning，Distributed machine learning，deep learning

13. 云游戏上有成本效益的服务器

Yusen Li, Yunhua Deng, Xueyan Tang, Wentong Cai, Xiaoguang Liu, Gang Wang, Cost-Efficient Server Provisioning for Cloud Gaming. TOMCCAP 14(3s): 55:1-55:22 (2018) (CCF Tier B Journal).
打不开

14. 硬盘错误预测的新指标

Jing Li, Rebecca J. Stones, Gang Wang, Zhongwei Li, Xiaoguang Liu, Jianli Ding, New Metrics for Disk Failure Prediction That Go Beyond Prediction Accuracy, IEEE Access 6: 76627-76639 (2018).

15.Session-Based 多人云游戏的服务器分配问题

Yunhua Deng, Yusen Li, Ronald Seet, Xueyan Tang, Wentong Cai, The Server Allocation Problem for Session-Based Multiplayer Cloud Gaming. IEEE Trans. Multimedia 20(5): 1233-1245 (2018)

16. 更高阶的汉明相关

Ming Su, Arne Winterhof, Hamming correlation of higher order. Adv. in Math. of Comm. 12(3): 505-513 (2018).

17. 部分拉丁矩阵双线图，离散数学上的电子记录

Eiran Danan, Raúl M. Falcón, Dani Kotlar, Trent G. Marbach, Rebecca J. Stones, Two-line graphs of partial Latin rectangles. Electronic Notes in Discrete Mathematics 68: 53-58 (2018).

Conference：

18. BitFunnel 搜索进程的索引压缩

Xinyu Liu, Zhaohua Zhang, Boran Hou,Rebecca J. Stones,Gang Wang, Xiaohguang Liu, Index Compression for BitFunnel Query Processing, SIG IR 2018, (CCF Tire A Conference)
前置知识：大规模搜索引擎下使用反转索引而反向链接需要运转存储代价
新方法：dictionary-data-structure BitFunnel with Bloom filter； docID reordering strategy to improve compression
成果：Inprovement in compression by 27% to 30%，at the expense of increasing the query processing time by 16% to 48% and increasing the false positive rate by around 7.6 to 10.7 percentage points.
关键词：BitFunnel; Bloom filter; compression; query processing

19. 关于反向索引的有效语法生成

Yan Fan, Xinyu Liu, Shuni Gao, Zhaohua Zhang, Xiaoguang Liu, Gang Wang, Efficient Grammar Generation for Inverted Indexes, WAIM-APWEB 2018 (CCF Tier C Conference)
新方法：A parallel generation algorithm for context-free grammar generation；A greedy dictionary pruning algorithm to reduce cache misses in query processing；Implement encoding, list,intersection and WAND querying on the grammar index.
成果：

  • Achieves a super-linear speedup with minor data overhead and nearly identical query efficiency compared to single-threaded algorithm.
  • With 10 threads to process the data set, abount 75 times faters with only 4.3% data overhead.
  • Parallel grammar generatrion incurs negligible（忽略不计） impact on query processing efficiency

关键词：inverted index compression，context-free grammar， parallel grammar generation algorithm

20. 基于拓扑链的一种新的增强可翻转数据隐藏

Bing Yan, Ming Su, Gang Wang, Xiaoguang Liu, Mingming Ren, A New Enhanced Reversible Data Hiding Using Topology Preserved Chains, ICME workshop 2018

21. 基于数据库服务的数据中心负载预测

Rui Cao, Zhaoyang Yu, Trent G. Marbach, Jing Li, Gang Wang, Xiaoguang Liu, Load Prediction for Data Centers Based on Database Service, COMPSAC 2018(CCF Tier C Conference)

22. 关于方向延伸搜索引擎平台的效率研究

Zhaohua Zhang, Naiyong Ao, Gang Wang, Xiaoguang Liu, An Extensible Search Engine Platform for Efficiency Research. ISPA 2018: 676-683(CCF Tier C Conference).

23. 在大规模搜索引擎里通过索引碎片复制策略改善资源利用

Yusen Li, Xueyan Tang, Wentong Cai, Jiancong Tong, Xiaoguang Liu, Gang Wang, Chuansong Gao, Xuan Cao, Guanhui Geng, *** Li, Index Shard Replication Strategies for Improving Resource Utilization in Large Scale Search Engines. ICPP 2018: 84:1-84:10(CCF Tier B Conference)
打不开

24. 来源于P2P的区块链共识机制和激励机制

Zhaoyang Yu, Xiaoguang Liu, Gang Wang, A Survey of Consensus and Incentive Mechanism in Blockchain Derived from P2P. ICPADS 2018: 1010-1015(CCF Tier C Conference)

25. 分布式内存缓存系统上的一种有效容错框架

Shuaibing Zhao, Lu Shen, Yusen Li, Rebecca J. Stones, Gang Wang, Xiaoguang Liu, An Efficient Fault Tolerance Framework for Distributed In-Memory Caching Systems. ICPADS 2018: 553-560 (CCF Tier C Conference)

26. 3D XPoint SSDs 在虚拟和非虚拟环境下的性能测试

Performance Analysis of 3D XPoint SSDs in Virtualized and Non-Virtualized Environments. ICPADS 2018: 51-60(CCF Tier C Conference)