Smiley face


Usama Naseer

Email: usama_naseer[at]brown[dot]edu
Office # 545, CIT, Brown University

Graduate Student in Computer Science Department at Brown University


INTERESTS

Hi. I am a PhD student at Brown University. My primary interests are in networks, internet measurments and web performance. Currently, I am working with Dr. Theo Benson. Before joining Brown, I spent an year at Duke University as a PhD student.
I am a huge football (the real footbal) fan, both playing and watching. I have been supporting Chelsea FC for the past 8 years.

SOCIAL LINKS



EDUCATION

PhD in Computer Science
Brown University
RI, USA

IN PROGRESS

PhD in Computer Science
Duke University
NC, USA

2016-17

Bachelors in Computer Science
Lahore University of Management Sciences (LUMS), Pakistan
Pakistan

2012-16



RESUME

Click HERE for resume.



WORK AND RESEARCH

Configtron: Tackling network diversity with heterogeneous configurations.
Paper accepted at USENIX HotCloud 17

The web serving protocol stack is constantly changing and evolving to tackle technological shifts in network- ing infrastructure and website complexity. For example, Cubic to tackle high throughput, SPDY to tackle loss and QUIC to tackle security issues and lower connection setup time. Accordingly, there are a plethora of protocols and configuration parameters that enable the web serv- ing protocol stack to address a variety of realistic con- ditions. Yet, despite the diversity in end-user networks and devices, today, most content providers have adopted a “one-size-fits-all” approach to configuring user facing web stacks (CDN servers). In this paper, we illustrate the drawbacks through empirical evidence that this “one-size-fits-all” approach re- sults in sub-optimal performance and argue for a novel framework that extends existing CDN architectures to provide programmatic control over the configuration op- tions of the CDN serving stack.
Conference link
Presentation slides


InspectorGadget: Inferring Network Protocol Configuration for Web Services.
Poster accepted at IMC 17

Over the last decade, the community has proposed a multitude of changes to the configuration of the network stack to improve the performance of web-services. These changes range, from initial congestion window size and initial retransmit timeout value to congestion control and HTTP protocol. While extensive work has studied the performance of each of these configuration optimizations, there is no holistic and general tool to infer, analyze, and understand the actual configuration choices used by different web services and content providers. In this poster, we present InspectorGadget, a tool to characterize the configuration of web servers. Inspector-Gadget incorporates some domain specific heuristics for reverse engineering configuration parameters and protocol versions. The goal of our work is to develop a framework that allows us to conduct a census of the configuration parameters used by content providers and analyze the impact of parameter-tuning on performance.


All Your Access Tokens 'Are' Belong To Us.
Submission in progress.

The study aims to expose the activity of malicious entities that generate fake activity on online social networks using a system of willing users. We call such entities collusion networks. These networks leverage the existing application infrastructure of the social networks. Collusion networks exploit application’s read/write permissions to collaborate user activity. Users install an app, which is authorized to do activities on their behalf without requiring account login credentials. The process is quid pro quo, users who enroll these services get likes/comments on their pages/posts, and their accounts are used to like/comment on other users’ pages/posts. 80% of these collusion networks use a common applications due to two vulnerabilities in Facebook applications. We tested top 1000 ranked applications and found that nearly 20% of authentic applications are vulnerable. To understand the population dynamics, we infiltrated collusion networks by creating several honeypot Facebook profiles and tracked millions of involved users.

Teacher's Assistant
LUMS

Worked as a T.A in three semesters at LUMS for the courses Algorithms, Databases and Calculus-1.



CONTACT

Email
usama_naseer[at]brown[dot]edu

Address
Office # 545, CIT Building
Brown university
RI, USA.

Created by BLACKTIE.CO