Shouqian Shi

PhD from UCSC (University of California, Santa Cruz) in Computer Networking. Now with Google.

About Me

I finished my 3 years and 3 months' PhD study at the Department of Computer Science and Engineering, UC Santa Cruz. I received the B.Sc. degree in Physics and B.E. degree in Computer Science and Engineering from the University of Science and Technology of China.

My research covers the fields of quantum networks, cloud networks, edge computing, software defined networks, network security, and network verification. I have also worked on Internet-of-Things (IoT), Radio Frequency Identification (RFID), vehicular ad-hoc networks, and compiler toolchain. I am interested in addressing fundamental problems in emerging networks (e.g., quantum networks and data center networks) and providing more advanced security services to resource limited hardware. The unique characteristic and strength of my research is the comprehensive study based on the expertise of a variety of fields including quantum systems, software-defined networking, security protocols, and cloud networks. In particular, I have had rigorous academic training on both computer science/engineering and physics. Compared to other CSE PhDs, this unique strength benefits my current and future research on quantum computing and networking. My research has been published to top conferences of many system fields, including SIGCOMM (networks), SIGMETRICS (performance), SOCC (cloud), ICNP (protocols), INFOCOM (networks), VLDB (databases), ICDCS (distributed systems), and IoTDI (IoT).

I enjoy reasoning. I have high-level analytical skills which have been already applied to every corner of my work and my life.

"Practices always go with concepts" and "Do the best" are my mottoes.

Download My CV

Publications

Education

Computer Engineering, Ph.D. Student View Transcript

University of California, Santa Cruz
2017-now

GPA: 4.0/4.0

Computer Science, Postgraduate View Transcript

University of Chinese Academy of Sciences
2014-2017

Weighted Average Grade: 88.4, Major Courses: 90.8, No ranking data available

Computer Science(B.E.) View Diploma View Transcript

University of Science and Technology of China
2010-2014

Weighted Average Grade: 89.2, No ranking data available

Applied Physics (B.S.) View Diploma View Transcript

University of Science and Technology of China
2010-2014

Weighted Average Grade: 91.2, Ranking: Top 5%, the 12th of 272 students

National College Entrance Exam

Shandong Province, China
July.2010

Grade: 675 / 750, Ranking: Top 0.1%, the 369th of 356,446 students

Research & Work Experience

Software Engineer - Google (Mar.2021 - Present)


Ph.D. Student - UCSC (Sep.2017 - Jan.2021)

Computer Networking, advised by Chen Qian

My research covers the fields of quantum networks, cloud networks, edge computing, software defined networks, network security, and network verification. I have also worked on Internet-of-Things (IoT), Radio Frequency Identification (RFID), and compiler toolchain. I am interested in addressing fundamental problems in emerging networks (e.g., quantum networks and data center networks) and providing more advanced security services to resource limited hardware. The unique characteristic and strength of my research is the comprehensive study based on the expertise of a variety of fields including quantum systems, software-defined networking, security protocols, and cloud networks. In particular, I have had rigorous academic training on both computer science/engineering and physics. Compared to other CSE PhDs, this unique strength benefits my current and future research on quantum computing and networking. My research has been published to top conferences of many system fields, including SIGCOMM (networks), SIGMETRICS (performance), SOCC (cloud), ICNP (protocols), INFOCOM (networks), VLDB (databases), ICDCS (distributed systems), and IoTDI (IoT).


Intern at Google (Aug.2020 - Nov.2020)

Declarative Route Derivation Engine, advised by Hongkun Yang

Google Cloud Platform features a set of virtual components to facilitate the network deployment, including virtualized cloud router, firewall, routing table, network peering, VPN gateway and VPN tunnel, as well as other interconnections such as direct connections and public Internet access. Though even simpler in configuration and deployment compared with SDN, virtualized components still exhibit subtle behavior features which once encountered with, may be hard to analysis and reinstate. Existing efforts present in network intent verifications where a series of user-specified checks can be launched against the current network control plane or data plane to find out whether the current network contains a forwarding loop, drops traffic which should be forwarded, etc.

This work fills the gap between the configuration and verification by providing a derivation engine to figure out the new data plane based on the current one and a proposed set of configuration changes. The derivation focuses on the routing table because most, if not all, configuration changes that may cause "chain reactions" on the data plane are ones that influencing the routes.

For easy extension with future features, the derivation engine is designed and implemented in a generalized configuration-based style. With a fixed derivation pipeline as a generalized derivation framework, some pluggable configurations (being ProtoBufs or Python snippets) are loaded as a complete description of the current behavior of the real GCP under a configuration update. The configurations are designed in a declarative style instead of imperative style. This way decouples the configuration description and the execution of the derivation engine and is thus considered as much easier for reasoning and debugging.

More details can be found on Intern Project - Declarative Route Derivation Engine.


Postgraduate - CASIA (Sep.2014 - Jun.2017)

Heterogeneous Computing, advised by Lei Wang

The first year of my postgraduate education is spent on all kinds of courses, and as a Direct Ph.D. student in CAS, I must take much more credits than those Master students to directly gain a Ph.D. degree.

I made a driver for MaPU (a not well-known coprocessor) and its corresponding user library during this period.

The ultimate aim is to port OpenCL to ARM-MaPU heterogeneous system while providing a user-friendly interface programming on MaPU.

More details can be found on MaPU-gem5 and MaPU-driver.


B.S. Thesis at CASIA (Apr.2014 - Aug.2014) Download Thesis

Design and Implementation of An Entropy Decoding Module of High-Definition Videos, supervised by Ling Li

I made an RTL description of an H.264 entropy decoder cooperating with MaPU, and it has passed the full series of official test cases and was patented ( A zero-order exponential Columbo code decoder and decoding method and CAVLC entropy decoder and entropy decoding method) by my lab.

Note: "石守谦" is my Chinese name, which you can find at the inventor lists after my professor and supervisor.

Sorry for the absence of the translation of the whole thesis and patents.

B.E. Thesis at USTC (Oct.2013 - Jan.2014) View Source Download Thesis

The Primary Exploration of A Configurable Intelligent Game System, supervised by Guiquan Liu

I independently designed and implemented a framework of chess-like games and an artificial intelligent engine that can be adapted to many popular chess types including Chinese Chess, Gobang and Chinese Draughts, with very little efforts (assumed at the level of 200 lines of code).

My original work is highly appraised by my supervisor Mr. Liu and he graded me 98 out of 100.

Because of the complexity of the chess "Go", I didn't manage to define an appropriate evaluation function of it.

Sorry for the absence of the translation of the whole thesis and absence of the readme file.

Co-Founder & Chief Developer (Part-time) - Jining Little Dream Electronic Technology Ltd. (Mar.2016 - Now)

We are committed to the development of automotive electronics research and tools in China and to boost the growth of Chinese automotive industry by providing tools and full-life-cycle solutions for functional safety development, hardware in the loop test and AUTOSAR software development under V model.


Research Assistant - SoC lab in USTC (2012 - 2014)

I participated in the independent development of a graphics card in the SoC lab, which included porting OpenGL and RTL design. Now this work is applied to the production environment, and my old fellows are working hard in their newly-founded company -- Suzhou Suxianwei Electronic Technology Ltd.

Representative Projects


MaPU-gem5

MaPU-gem5 is a simulator of MaPU with cycle-level precision. MaPU-gem5 is adapted from the open-source project gem5. While inheriting the bus system, memory model and debug system from gem5, MaPU-gem5 extended the support for the MaPU specific features such as microcode, private DMA, VLIW and multi-core.

Find out more


MaPU-driver

A Linux driver to provide methods to transfer data between MaPU private memory and DDR memory under the control of Linux. Our lab has suffered long from the absence of an OS running on the SoC, and have to manage every corner of the process of loading data, parsing ELF files and loading segments to their designated places, control the life cycle of the MaPU program launch. The driver is fully designed and implemented by myself, providing a reliable way to exchange data and control the MaPU programs' life cycle.

Find out more


Quantitative-Analysis - Designed for the Stock Dealers View Source

Having a spider and an LALR(1) front-end embedded, this tool helps stock dealers update stock data and get a list of stocks that meet his or her need in seconds. Well integrated with Chinese prevailing browse tools such as TongDaXin and EastMoney, it is easy to adopt it in your deal.

Though ugly-looking, it has played an important role in the stock deal of one of my richest friends who has earn a profit of at least RMB 3000,000, which is about $440,794.

Find out more


Two Radar Control Tools

I was once introduced to an aerospace research institute and help them to build two control and visualization tools, for a ground penetrating radar to locate resources or detect oil pipeline damage under the ground and an indoor radar to locate the position of a working device or people.

Find out more