Yang Zhang


715 Broadway Room 705
New York, NY 10003


2012/09 — Now
PhD Student in Computer Science, New York University
Research focus: distributed systems, transactional systems
2009/09 — 2012/07
M.Sc. in Computer Science & Technology, Tsinghua University, Beijing, China
Research focus: virtualization, distributed storage systems
Thesis: "Research on Distributed Deduplication Storage for Virtual Machine Disk Images"
2005/09 — 2009/07
B.Sc. in Computer Science & Technology, Tsinghua University, Beijing, China


2014/06 — 2014/09
Software Engineering Intern, Google Inc., New York
  • Built an indexing service for Google+ ACL (Access Control List). It is used in private search and Google Cloud Platform.
2011/07 — 2012/06
Graduate Intern Technical, Intel Corporation, Beijing, China
  • Joined SoC (Software on Chip) development group, worked on smart TV devices.
  • Built a unit-testing framework for C/C++, which supports testing Linux kernel modules. It is used to test device drivers on Android system. This framework is also ported to Windows, and used in testing device drivers for Windows 8 system.
  • Built a web-based log analyzer in Python, powered by the tornado web server, and visualizing results using jQuery.


Extract more concurrency from distributed transactions.
Shuai Mu, Yang Cui, Yang Zhang, Wyatt Lloyd, and Jinyang Li.
11th USENIX Symposium on Operating Systems Design and Implementation (OSDI '14)
Transaction Chains: Achieving Serializability with Low Latency in Geo-Distributed Storage Systems.
Yang Zhang, Russell Power, Siyuan Zhou, Yair Sovran, Marcos K. Aguilera, and Jinyang Li.
24th ACM Symposium on Operating Systems Principles (SOSP '13)
Paper      Poster      Slides      Talk

Past Projects

Liquid: Distributed Deduplication Storage for Virtual Machine Disk Images
  • When creating a new Virtual Machine, we need to copy a new disk image from a set of templates. Distributing template disk image takes time, and copying new disk image consumes a lot of network bandwidth.
  • We observed that most virtual machine images contain similar data, thus deduplication techniques can tremendously reduce storage and network consumption. Also, P2P techniques can significantly speedup template image distribution.
  • We built a filesystem called Liquid with those ideas, and demonstrated its usefulness. Liquid is a POSIX-compliant filesystem. It was implanted in C, using FUSE library.
  • Part of my master degree thesis.
Nova: Management System for Large-scale Virtual Machine Deployment
  • A distributed system to manage life cycle of many virtual machines in a large cluster.
  • Implemented using Ruby on Rails.
  • Part of my master degree thesis.


C, C++, Python, Ruby, Java in Linux environment. Some of my projects opensourced at http://github.com/santazhang.
Chinese (Native), English (Proficient)