Xianggang Zheng is a recent graduate from Carnegie Mellon University with a Master's degree in Information Networking and a Bachelor's degree in Computer Science from the University of Washington. He has experience as a software engineer intern at LinkedIn and has worked on several academic projects related to cloud computing, distributed systems, and computer networks. His technical skills include languages like Java, C, C++, Node.js, SQL, Python, and Perl.
On-demand self-service. Cloud computing resources can be provisioned without human interaction from the service provider. ...
Broad network access. ...
Multi-tenancy and resource pooling. ...
Rapid elasticity and scalability. ...
Measured service.
On-demand self-service. Cloud computing resources can be provisioned without human interaction from the service provider. ...
Broad network access. ...
Multi-tenancy and resource pooling. ...
Rapid elasticity and scalability. ...
Measured service.
Opal: Simple Web Services Wrappers for Scientific ApplicationsSriram Krishnan
The grid-based infrastructure enables large-scale scientific applications to be run on distributed resources and coupled in innovative ways. However, in practice, grid resources are not very easy to use for the end-users who have to learn how to generate security credentials, stage inputs and outputs, access grid-based schedulers, and install complex client software. There is an imminent need to provide transparent access to these resources so that the end-users are shielded from the complicated details, and free to concentrate on their domain science. Scientific applications wrapped as Web services alleviate some of these problems by hiding the complexities of the back-end security and computational infrastructure, only exposing a simple SOAP API that can be accessed programmatically by application-specific user interfaces. However, writing the application services that access grid resources can be quite complicated, especially if it has to be replicated for every application. In this presentation, we present Opal which is a toolkit for wrapping scientific applications as Web services in a matter of hours, providing features such as scheduling, standards-based grid security and data management in an easy-to-use and configurable manner
Dynamic Provisioning of Data Intensive Computing Middleware FrameworksLinh Ngo
Presentation at the 1st Workshop on Science of Cyberinfrastructure: Research, Experience, Applications and Models.
Topic: A Case study by the Big Data Systems Lab group at Clemson University on setting up non-root dynamic provisioning of two big data infrastructures on a shared research computing resource: Hadoop and HPCC Systems.
Opal: Simple Web Services Wrappers for Scientific ApplicationsSriram Krishnan
The grid-based infrastructure enables large-scale scientific applications to be run on distributed resources and coupled in innovative ways. However, in practice, grid resources are not very easy to use for the end-users who have to learn how to generate security credentials, stage inputs and outputs, access grid-based schedulers, and install complex client software. There is an imminent need to provide transparent access to these resources so that the end-users are shielded from the complicated details, and free to concentrate on their domain science. Scientific applications wrapped as Web services alleviate some of these problems by hiding the complexities of the back-end security and computational infrastructure, only exposing a simple SOAP API that can be accessed programmatically by application-specific user interfaces. However, writing the application services that access grid resources can be quite complicated, especially if it has to be replicated for every application. In this presentation, we present Opal which is a toolkit for wrapping scientific applications as Web services in a matter of hours, providing features such as scheduling, standards-based grid security and data management in an easy-to-use and configurable manner
Dynamic Provisioning of Data Intensive Computing Middleware FrameworksLinh Ngo
Presentation at the 1st Workshop on Science of Cyberinfrastructure: Research, Experience, Applications and Models.
Topic: A Case study by the Big Data Systems Lab group at Clemson University on setting up non-root dynamic provisioning of two big data infrastructures on a shared research computing resource: Hadoop and HPCC Systems.
1. XIANGGUNAG ZHENG
4742 Centre AVE APT#402, Pittsburgh, PA, 15213 206-617-4233 zxg.zheng@gmail.com
EDUCATION:
Carnegie Mellon University – Pittsburgh, Master of Information Networking May 2017
Cumulative GPA: 3.95
University of Washington – Seattle, Bachelor of Science: Computer Science June 2015
Cumulative GPA: 3.84 (Dean’s list 7 quarters)
TA of Advanced Cloud Computing, CMU January 2016
TA of Computer Graphics, UW March 2015
WORK EXPERIENCE:
LinkedIn Software Engineer Intern 06/2016 – 08/2016
---- Leverage spare/unused resources in the online clusters by dynamically add them to offline Yarn cluster to run Hadoop
Jobs with low SLA
• Designed the architecture from scratch
• Demonstrated real Hadoop jobs running on online cluster using offline cluster setup
• Implemented job selection mechanism in order to only run selective jobs on added dynamic resources
• Implemented service to dynamically manage the resources added to Yarn cluster by monitoring online cluster usage.
RELATED COURSES:
Hardware/Software Interface Data Structure Software Design & Implementation
System Programming Database Management Operating System
Computer Network Distributed System Advanced Cloud Computing
TECHNICAL SKILLS:
Languages: JAVA, C, C++, Node.js, SQL, Python, Perl, PHP
ACADEMIC PROJECTS:
Advanced Cloud Computing, CMU -- Prof. Garth Gibson/ Prof. Majd F. Sakr 01/2016 – 05/2016
----Implemented change or feature in a state of the art cloud computing framework
• Built cloud using OpenStack, including load balancer and auto scaling service
• Parallelized Matrix Factorization with Spark
• Implemented Heterogeneous scheduling on Yarn
Cloud Computing, CMU -- Prof. Majd F. Sakr 09/2015 – 12/2015
----Enable technologies and experiences through projects utilizing public cloud infrastructure
• Utilized multiple Data Analytical Engines to handle big data problems, including Haddop, Spark, Kafka and Samza
• Built distributed Key-Value Store with different consistency level
• Utilized AWS Load Balancer and Auto Scaling group to handle scaling problems, including improving the Load
Balancing strategy
• Design and built an OLAP online tweet data service from scratch, including frontend, backend, database and
processing over 1TB data. In the end, our service is able to handle 6 different queries at the same time with certain
performance (request per second) as requirement.
Emulab Cluster Enhancement, CMU -- Prof. Garth Gibson 01/2016 – 05/2016
----Improved resource sharing in Emulab, the cluster management used by CMU to allocate shared resource to graduate
students for research purposes
• Designed and implemented cool down mechanism to prevent greedy user from taking over the majority of the
resources
• Designed and implemented notification mechanism to notify users who registered themselves in subscription list when
free resource is enough to fulfill their requests
Distributed System, CMU -- Prof. Mahadev Satyanarayanan / Padmanabhan Pillai 01/2015 – 03/2015
• Built abstraction over system call to redirect request to remote server using RPC
• Built file cache proxy, including defining the protocol used between the proxy and the server and how the cache is
managed
• Built scalable service to auto scale out/in based on the current load of user requests
• Built failure tolerant two phase commit mechanism to coordinate the decision making process
2. ACADEMIC PROJECTS(UNDERGRADUATE):
Distributed System, UW -- Prof. Tom Anderson 01/2015 – 03/2015
• Built Primary/backup Service with state fault tolerant coordinated by master server
• Built Paxos Based Key/value Service with fault tolerant without the coordination of master
• Built Shared Key/value Service, a distributed system that handles all faults, where different shards handle different
categories of requests and each shard consists of a replica group with Paxos protocol
Computer Network, UW -- Prof. John Zahorjan 01/2015 – 03/2015
• Built network proxy, capable of both replaying HTTP requests and HTTP CONNECT tunneling
• Built Tor Routing Network that routes traffic from the browser through a number of Tor nodes before sending it to the
server, using asynchronous event loop provided in Node.js
Operating System, UW -- Prof. Tom, Anderson 03/2014 – 06/2014
This project is to build basic functionalities for os161 operating system
• Implemented synchronization primitives for os161 and use them to design a multi-threaded bounded network queue
that stores k items and provides fair allocation among connections that share the queue
• Implemented the interface between user and kernel (system calls), including open, read, write, lseek, close,dup2, fork,
execv, waitpid, _exit
• Implemented virtual memory management, including page swapping and TLB miss