Yao Yue


PhD Student, Computer Science
4161 Upson Hall, Department of Computer Science,
Cornell University, Ithaca, NY
 
yy277 AT cornell DOT edu

::Software

 

NPCryptBench

NPCryptBench (China software copyright registration number 2006SRBJ1227) is a software package for implementing and optimizing cryptographic applications on data plane (fast path) of network processor architecture. It consists of ten cryptographic benchmarks chosen on their popularity, representativeness and availability. With a growing awareness of communication security, cryptographic processing soars at routers, servers or other nodes on fast path. While software-based solutions have an edge in flexibility over their hardware alternatives, the impact of implementing cryptographic applications on network processors is not yet well studied.

NPCryptBench focuses on the inner loops executed by data plane processors, which present the majority of processing time and consumes considerable power. The other functional components, such as initialization and key management, are reckoned as tasks of control plane processors. Nevertheless, all code segments can be extended into full-fledged programs with ease.

So far, nine institutes and universities from four countries have officially requested the package for academic purposes, e.g. power estimation and architecture evaluation. For a complete list of subscribers and package details, please visit NPCryptBench Homepage.

Collaborator: Zhangxi Tan (now at UC Berkeley), Donghua Ruan (MEng 2007, Tsinghua University)

 

BitTorrent Modeling

We have extended the fluid model brought up by Qiu and Srikant[1] to include multiple groups of peers[2]. However, analytical solutions are often not available when more than two groups coexist in the swarm.

Consequently, I wrote this simple program to simulate the dynamics of the differential equations with discrete time slots. The slots are sufficiently small to guarantee accurate results. Several gnuplot scripts illustrate how different groups evolve to their steady states with statistics in result file.

 
 
  1. D. Qiu, R. Srikant, "Modeling and performance analysis of bittorrent-like peer-to-peer networks", Proc.SIGCOMM '04, 2004, pp. 367-378.  [.pdf]
  2. Yao Yue, Chuang Lin, Zhangxi Tan, "Analyzing the Performance and Fairness of BitTorrent-like Networks Using a General Fluid Model", to appear on Elsevier Computer Communications

 Basic

resume (pdf)
Links

 Research

Publication
Projects
Software
Service
Skills

 Personal

Personal
Quotes
Misc