Research

My research lies at the intersection of computer networks and human-computer interaction (HCI). The overarching theme of my research is improving the usability and user friendliness of computer networks, by improving the performance of computer networks. Broadly, I work to understand how people experience computer networks by studying the performance of the applications they use on computer networks, as well as how they react to and interact with these applications.

Current projects

Intuitive tools and terminology for home network maintenance

The computer networks within our home are powerhouses. Our smartphones, our tablets, our laptops and desktops, gaming systems, Smart TVs, streaming services --- they allow us to accomplish an ever-increasing list of life tasks, from work to entertainment, from financial management to health management. As the number of devices grow, the complexity of their connections --- to each other and to the Internet --- also grow.

Such complex computer networks used to only exist in the workplace or at school, with trained tech support staff at the ready to make everything work seamlessly and to fix problems as they arose. Today's complex home computer networks do not come with professional tech support, leaving homeowners and residents to fend for themselves, regardless of their level of technical expertise.

Ideally, the tools homeowners and residents use to set up, maintain, and troubleshoot their home computer networks should match the mental models they hold of home computer networks. Anyone who's provided tech support for a friend or relative knows that this is not currently the case.

This project examines the mental models homeowners and residents hold of their home computer networks, so that we can design home network setup, maintenance, and troubleshooting tools that are

We are currently working on a few different subprojects.

Troubleshooting terminology

Technical terminology is often a barrier to home network users' expressions of desired network performance and of network problems. Text-based feedback is often used by networks to communicate system state to the user. Users seek help in online forums and search engine queries (written text) as well as through family, friends, and technical support staff (verbal). But how user-friendly is this terminology? What technical terms do users understand, and are there non-technical terms for the same concepts that they understand better?

We use card sorting to deduce understanding of troubleshooting terminology. We identify a list of words used in troubleshooting environments to describe home network performance problems, around the themes of connectivity and bandwidth utilization/capacity, extracting these terms from online tech support forums. We classify these terms into three categories: technical (used and understood by technical experts, a.k.a. the "real words" associated with a technical concept), colloquial (used and understood by non-experts), and neutral (used by experts, but understandable to non-experts outside of a technical context).

Study phases:

Troubleshooting behavior in a campus community

How do help-seekers and help-givers negotiate troubleshooting transactions? How do help-seekers request assistance and determine whether or not to trust the advice from a help-giver? How do help-givers modulate their language and advice to more closely match the (perceived) mental model of the help-seeker? By better understanding these interactions, we can better understand what mental models are most common among help-seekers and how these mental models vary with expertise. In turn, this will help us craft better troubleshooting assistance strategies for help-givers, and guide the design of more effective tools and interfaces for troubleshooting.

Using 10 years' worth of data from a campus online trouble ticketing system, we are exploring the ways in which people ask for help: what language they use, what technical terms they do and do not understand, and how they respond to advice. Similarly, we are examining how campus IT workers report on user interactions and provide assistance to help-seekers, what terminology they use in their answers, and how they determine help-seeker expertise.

Project publications

Amy Csizmar Dalal, Jackie Chan, and Kirby Mitchell. 2019. A Preliminary Study of the Role of Language in Home Network Troubleshooting. In Extended Abstracts of the 2019 CHI Conference on Human Factors in Computing Systems (CHI '19). ACM, New York, NY, USA, Pages LBW0272, 6 pages. DOI: https://doi.org/10.1145/3290607.3312856 [PDF] [Poster]

Self-Healing Home Networks

Self-healing networks detect existing or potential pathologies and fix or mitigate them with minimal to no human intervention. For example, a network that detects a drop in the available bandwidth along a particular link, or within a series of links, and then reroutes data around these links is a self-healing network. Because such pathologies are visibly evident in the end-users' quality of experience (QoE), self-healing networks are by definition responsive to the needs of networked applications, such as streaming video. Self-healing networks, then, are fundamentally important to the design, engineering, and development not just of future applications, but of future computer networks. By understanding how video applications are affected by network conditions, we can design network protocols and structure new networks to better support these and future (high quality, high-definition, immersive) video applications. Similarly, understanding how network conditions impact application performance leads to the design of more robust application protocols that can better utilize existing network resources.

Building upon or work on video QoE, we are identifying heuristics for determining when network and/or application conditions signal potential eminent degraded QoE, and constructing system architectures to support QoE measurement, analysis, and feedback into the system.

Project publications

A. Csizmar Dalal. "A Framework for Self-Healing Home Networks." In Proceedings of the Tenth International Conference on Heterogeneous Networking for Quality, Reliability, Security, and Robustnests (Qshine), Rhodes, Greece, August 2014. (Short paper, pdf)

Past Projects

Publications

Full list