Dai Dong

My research interests are in cloud computing focusing on the backend storage and processing system. I have a particular interest in mobile cloud computing facing next generation computer revolution - Mobile revolutionary.

There are many things to be done to satisfy the massive data storage and processing these data in a short time to provide instance and useful information for mobile device.

• Sedna: A Memory Based Key-Value Storage System for Realtime Processing in Cloud Dong Dai, Xi Li, Chao Wang, Mingming Sun, Xuehai Zhou In The 2012 International Conference on Cluster Computing Workshops (IASDS 2012) in conjuntion with IEEE Cluster'12, 24-28, September 2012
• Cloud Based Short Read Mapping Service Dong Dai, Xi Li, Chao Wang, Xuehai Zhou In IEEE Conference on Cluster Computing, Beijing, China, Sep.24-28, 2012
• Phase Detection for Loop-based Programs on Multicore Architectures Chao Wang, Xi Li, Dong Dai, Gangyong Jia, Xuehai Zhou In IEEE Conference on Cluster Computing, Beijing, China, Semp. 2012
• Credit evaluation in e-business based on inverted reference using mapreduce Dai Dong and Feng Yang In The Third E-business Ecological Workshop, July 2010 Download: pdf
• An auto-configuration tool for heterogeneous hadoop cluster Dai Dong and Xuehai Zhou and Feng Yang and Chao Wang In Journal of the Graduate School of the Chinese Academy of Sciences Download: pdf

• Detecting Associations in Large Dataset on MapReduce
  - In this research project, we propose a MapReduce-based solution for accelerating the complex association detecting algorithm which was firstly introduced by David N.Reshef in Science Journal 2011. Our solution includes dataset storage, data pre-processing, distributed memory cache, and four dependent MapReduce jobs. We implement our solution based on Hadoop and prove our distributed solution can achieve a linear speed up along with server number increasing through experiments.
• Mobile Device Management System
  - Mobile device management system targets the enterprise customers. We have successfully deployed it for several service providers of China. The whole system was designed and implemented by a team i leaded in the past year. In this project, we implemented Android and iOS apps, distributed socket servers, and a J2EE web server.
• Sedna Distributed Storage System
  - Sedna is a key-value distributed storage system based on DRAM. It is designed to run on commodity servers and process data under a realtime requirement. We build Sedna based on Memcached and ZooKeeper. Besides, we use DHT to partition data, keep three consistent copies for each data piece, use fast recovery architecture to keep persistency. Besides these features, Sedna also provides some trigger-based APIs to support realtime processing.
• Hadoop Auto-configuration Tools
  - In this project, we develop an auto-configuration tool for large-scale heterogeneous Hadoop cluster. We propose a fuzzy logic strategy to define a serial of rules instead of providing specific value of each parameter. This way provides a chance to re-use others’ effective rules and simplifies the configuration of large-scale Hadoop cluster.
• FastMR
  - A very simple prototype where MapReduce is streaming and Storage is in memory.
• HProbes
  - HProbes is an instrument tool for Hadoop. It helps users monitor a running cluster, running jobs, even running tasks in a more flexible way. System administrators can submit scripts to monitor their interested events and cancel these monitors any time. Also HProbes provides monitor in MapReduce level to help users identify the problems of their jobs see details in [HProbes]

• Important Blog

Perspectives	James Hamilton's Blog	Last Update On 11/04/2011
Advising From Wierman	Wierman's HomePage (Minghong Lin's PhD. Professor)	Last Update On 16/04/2011
Butler W.Lampson's Blog	Microsoft Research Center in Cambridge.	Last Update On 12/04/2011
Facebook Engineering Blog	Facebook Engineering Blog	Last Update On 11/04/2011
ShangKun Yang's Blog	TaoBao architecture designer. Basic Introduction on TaoBao's Cloud System.	Last Update On 12/04/2011
Chuanhui's Blog	A architecture desinger from TaoBao's Team	Last Update On 12/04/2011
Yun Liu's Blog	Yet another architecture desinger from TaoBao's Team	Last Update On 12/04/2011

• Reading

Fang, LiZhi's Anthology	Fang, LiZhi's Anthology	Last Update On 12/04/2012
Call For Paper Wiki	A Wiki for Calls For Papers	Last Update On 09/04/2011
Network Conferences	Networking Conferences Statistics	Last Update On 09/04/2011
Usenix Best Papers	Best Papers List in USENIX Conferences	Last Update On 12/04/2011
Dynamo: A flawed architecture	A discussion on the Dynamo architecture. It's not the best choice!	Last Update On 11/04/2011
Dynamo: Yet another Flaw discussion	Yet another discussion on Dynamo, From a chinese engineer	Last Update On 12/04/2011
Cassandra Consistency Discussion	Some Great Ideas on design of mastered replication	Last Update On 11/04/2011

• Links

Google Scholar	Google search engine I use mostly.	Last Update On 03/04/2011
PyCon2012US	Python Conference 2012	Last Update On 03/16/2012
Ustc bbs	BBS board of USTC	Last Update On 03/04/2011
Sina Mini-Blog	See me in the popular twitter in China.	Last Update On 03/04/2011
Ren Ren	See me in the popular social network in China.	Last Update On 03/04/2011
Dave Andersen's Page	My websit is based on this page. Thanks here!	Last Update On 03/04/2011

• Music

Douban FM	Douban FM is a great random mp3 player. All the mp3 files are copyrighted. It's also the first mp3 radio service in China.	Last Update On 03/04/2011
Beethoven.Symphony.No.5	Herbert.von.Karajan.-.1973 Version. One of the best version.	Last Update On 06/04/2011

---

• Video