Senior Position

graduated from Harbin Institute of Technology (the top 10 university in China) in 2003 and my major is computer. I am an expert in the fields of C#, VB.net, asp, asp.net, sql, JavaScript, Ajax, xml, web service, win service, socket.
2003/7--2008/12：hexunI am also good at improving the performance of web server (windows server 2000/2003 and IIS and .net deployed websites) and database (SQL Server 2000/2005), and I have a strong understanding of object-oriented concepts, design patterns, an
Industry:IT Service/System Integration | Company Type: | Company Size:Company Profile:2003-7 - Now devLocation:beijing | Job Type:Full-time | Department: | Report to: | Number of Subordinates:employees | monthly salary:20000RMBchi hair straightener(Non-managerial)Responsibilities and Achievements:More than 7 years experience of MS products (asp.net,c#,sqlserver) .Rich experience in asp.net,c# and windows platform and be in charge of more than 50 web servers;Rich experience in database (sql server) performance and indexing tuning, and experienc of the search engine (including inverted index, Word segmentation, and lucene etc) as well; be good at asp.net(c#) and have architected and developed high performance distributed file system, and multithreaded spider and stress test tool with socket; Rich experience of troubleshooting with windbg for dump analysis; Experience in client side JavaScript (including AJAX), And Rich experience in http and tcp etc. And contributed to several open source projects on codeplex.
Good ability of English reading and writing.More than 7 years of development experience, and at least 3 years of team leader for more than 10 teammates and project management experience.
The latest work(in hexun.com):0.Lead the team to develop an portal for e-learning,which includes video,photo,flash etc.
0.Lead the team to develop the restriction of frequency of each user's action, such as the frequency of junk comments,which can response the requests for more than 100 times per second.
0.Windbg for the deadlock of the several online system.0.Lead the team to develop the stock portal which includes cross domain transfering some data.
1. Implement Sharding in the blog's database (more than 30 million data in total),I design the sharding scheme and readjust the specific code based on that scheme.chi straightener
2. Architect and develop an extensible and scalable distributed file system engine for blog system, and made the performance improve more than 300%; and a single server can efficiently uphold 5 million requests per day. I just had a reference to the Berkeley DB and the search engine, and with the algorithm of self developed high efficiency hash table and double-linked list (both implemented with c# array) instead of using generic provided by .net framework.
3. Reconstruction memcache client, make performance improve at least 200%. And I implement the new client with high performance stack to avoid the deadlock in the former edition.
4. Architect and develop high concurrency interactive products (the latest reader, the new point system, the latest visitor), makes a single server can uphold 10 million requests per day for frequently changed situation. I implement it with a high performance hash table with c# array.
5. Architect and develop high interactive friend event tracking system, and the system can deal with more than 30 million friend events, and make the performance improve dozens of times. The former one stored all the data in database, which made the disc io queue length more than 10 with low performance. And the new edition is mainly manipulated in memory with the data structure of hash table and list, the sort algorithm in this is quick sort.
6. As the leader, tune register system and make stability and performance considerably boost.
7. Architect and develop a module of “the users who've seen this news also have seen” and can deal with millions of related data per day. Though all the data are stored in database and all the executions are all in database, the application is with high performance. For I had a good experience in index tuning, and in this I just try to avoid bookmark lookup and join, and the table columns are nearly all with the type of int or other short length type to avoid another disk seek for the text or ntext column.chi straightener
8. RSS crawler optimization (with socket). And I just add the "head" request before the real "get", and if the "304" which is "Not Modified" return, then the "get" request is not needed any longer.
9. Implement sql server 2k5 partition for online sharing photo system and the amount of data is more than 50 million. And I design the scheme and the partition column and readjust the code.