System Management

Xing Lin

*** ********** *******, **** **** City, Utah, 84108

*******@**.****.*** or 801-***-****

OBJECTIVES Understand computer system behaviors: performance analysis and workload charac-

terization

Justify design decisions: prototype validation

INTERESTED Storage and le system, especially in deduplicating and cloud storage systems

AREAS Operating systems and computer architecture, especially in multicore related designs

EDUCATION Doctor of Philosophy Computer Science

University of Utah, Salt Lake City, Utah

Fall 2009 Summer 2014 (expected)

Advisor: Robert Ricci

Bachelor of Engineering Software Engineering

Nankai University, Tianjin, P.R.China

Fall 2005 June 2009

INTERNSHIP Software Developer Intern May 12 Sep. 12

EMC/Data Domain, Backup and Recovery Systems Division (BRS), Advanced De-

velopment Team, Princeton, NJ

Worked on improving the read performance for the Data Domain deduplicating storage

system; Implemented chunk-level LRU and chunk-level access-aware caching prototypes

in the Data Domain File System.

PROJECTS Research Assistant Fall 09 present

University of Utah, School of Computing, Flux Research Group, Salt Lake City, UT

FAST: Fair Assignment for Storage Tenants Jan. 12 present

Workloads using cloud storage systems su er from interference from co-located work-

loads. Our goal is to design a cloud storage system with predictable performance and

provide fairness among tenants.

Study on the read throughput in deduplicating storage systems Apr. 11

present

Data fragmentation is fundamental in deduplicating storage systems. That results

in unpredictable read performances. We did some analysis of the read throughput

and propose to design a complementary prefetching and adaptive block-reorganization

framework to improve the disk I/O performance.

High-Performance Disk Imaging with Deduplicated Storage Apr. 11 Jan.

12

Enhanced the Emulab s disk imaging system (Frisbee) by storing operating system disk

images with a deduplicating storage system (Venti), which achieves e cient dedupli-

cation and high-performance disk imaging performance simultaneously.

Re ning the Utility Metric for Utility-Based Cache Partitioning Feb. 11

June. 11

the Advanced Computer Architecture course, taught by Rajeev Balasubramonian

Cache misses have varying impacts on IPC across programs because of varying levels

of latency tolerance in each program. In this work, we discover a non-trivial number of

cases where the use of MPKI is sub-optimal. We then propose a simple mechanism that

uses two IPC samples and a miss rate curve to compute an IPC curve. This allows us

to better quantify marginal utility in terms of IPC and arrive at performance-optimal

cache partitions.

Research Assistant Oct. 08 Sep. 09

Nankai University, Tianjin, China

Research on Freeze-drying and Resuming Processes

2008 Microsoft Windows Core Theme Project

Implemented the funtion to freeze-dry portable executables(PEs) and dynamic linked

libraries(DLLs); Succeeded to freeze-dry and resume a simple loop program; Inves-

tigated how virtual memory is managed and implemented the funtion to freeze-dry

stacks, heaps and other virtual memory areas for a speci c process; Implemented ve

polices to select a waiting process to run for semaphores. Investigated the handle mech-

anism; Published four papers about freezing and unfreezing for process with multi- le

handles, semaphore, handle and virtual memory management mechanisms.

COURSE Graduate Courses:

PROJECTS Database Kernels and Large Data Management from Feifei Li

Implemented B+ tree dynamic insertion, an external sorting algorithm and bu er

management with love-hate page replacement policy.

Operating System from Matthew Flatt

User-level Thread

Implemented a timer, semaphore mechanism, a proportional share thread scheduler

and a non-blocking socket calls, based on a user-level thread

Page Replacement Algorithms

Implemented FIFO, second chance, LRU and optimal page replacement algorithms

Thread-safe File System

Implemented operations such as open, close, unlink, read, write, getpos, truncate and

lseek for multi-thread processes

Network Security from Sneha Kasera

Implemented an expanded Needham Schroeder Mediated-Authentication Scheme; a

simpli ed SSL protocal; a network discrete event route simulation

Undergraduate Courses:

Basic Memory Management

Allocate a huge memory space from operating system and then manage this memory

pool for memory allocation and deallocation operations for test applications. Performed

boundary and checksum check for reads and writes.

UNIX File System

Used a le as a virtual disk and layout the traditional UNIX le system on it; Imple-

mented operations to change/create/delete directories and create/delete/read/write

les. Only 3 out of more than 120 students implemented a le system during that

semester.

PUBLICATIONS Xing Lin, Yun Mao, Feifei Li, Robert Ricci, Towards Fair Sharing of Block Storage in

a Multi-tenant Cloud, Appeared in 4th USENIX Workshop on Hot Topics in Cloud

Computing (HotCloud 12), Boston, MA, June 12-13, 2012

Xing Lin, Rajeev Balasubramonian, Re ning the Utility Metric for Utility-Based

Cache Partitioning., The 9th Annual Workshop on Duplicating, Deconstructing, and

Debunking(WDDD 2011), held in conjunction with the 38th International Symposium

on Computer Architecture (ISCA-38). San Jose, CA, June 5, 2011.

Chunxia Zhang,Xing Lin( rst student author),Shuguang Lin,Xudong Li, Study of

Handles Mechanism in WRK. The 2nd International Workshop on Education Tech-

nology and Computer Science (ETCS 2010). Wuhan, China. 6-7 March, 2010.

Xudong Lee, Chunxia Zhang, Xing Lin( rst student author), Shuguang Lin, Introduce

Satisfy policies into Semaphore in WRK., International Conference on Computational

Intelligence and Software Engineering (CiSE 2009). Wuhan, China. December, 2009.

Xudong Li,Chunxia Zhang,Xing Lin( rst student author),Shuguang Lin, VAD: A De-

tail Investigation into Process s Memory. 2009 International Conference on Compu-

tational Intelligence and Security, (CIS 2009). Beijing, China. 11-14 Dec, 2009.

POSTERS Raghuveer Pullakandam, Xing Lin, Mike Hibler, Eric Eide, and Robert Ricci, High-

Performance Disk Imaging With Deduplicated Storage 23rd ACM Symposium on

Operating Systems Principles(SOSP 11), Cascais, Portugal, Oct. 23, 2011.

HONOURS AND Student grants from OSDI 10, FAST 11, ISCA 11, FAST 12 and OSDI 12

AWARDS National Encouragement Scholarship for excellent students, Nankai University, 2007-

2008

Three Goods Student of Nankai University, Nankai University, 2006-2009

Second-level Schorlarship of Nankai University, Nankai University, 2006

IBM Certi cate of On Demand Business - Solution Designer V2, Nankai University,

2007

SKILLS Languages: C, Perl, Java, C++, SQL, MPI, OpenMP

Platforms: Linux, Windows Research Kernel, Emulab

Contact this candidate