Ph.D. Candidate, Computer Science, Case Western Reserve University, Expected Graduation August 2015.
B.S., Computer Science, Ohio University, 2009.
Primary Email: email@example.com
Secondary Email: firstname.lastname@example.org
We leverage a large darknet consisting of five /8 networks to assess the state of TCP port 443 filtering. We describe a methodology that allows us to detect filtering (or lack thereof) of TCP port 443 across a large portion of the network, before examining the limitations of discovering filtering via passive observation.
TCP connections are often modeled as bulk-transfers when simulating network traffic. This work proposes a new, application-agnostic, methodology for capturing the behavior of TCP connections on the network and assesses how often modern applications use TCP for purposes other than bulk transfer.
In this work, we examine the performance of TCP connections on a residential fiber network. After exploring the performance traffic is able to achieve, we study high-volume TCP connections in order to pinpoint what prevents TCP from using the excess available bandwidth in the network.
We perform a long-term study of a residential fiber network. We collect and analyze data over a two-year period in order to understand the way users behave on a network with excess bandwidth.
We examine the performance implications of changing the initial RTO for a TCP connection from 3 seconds to 1 second. We gather traces from four vantage points at different times and study how often connections would be negatively or positively affected by the change.