Dissertation (klassisch)

\documentclass[12pt, oneside]{report}
\usepackage[margin=1.25in]{geometry}
\usepackage{setspace}
\usepackage[hidelinks]{hyperref}

\newcommand{\committeeline}[2]{%
  \noindent\makebox[3.4in]{\hrulefill}\\
  {#1, #2}\\[1.6em]}

\begin{document}

% ---------- Title page ----------
\begin{titlepage}
\begin{center}
\vspace*{0.4in}
{\large ALDERMERE UNIVERSITY}\\[0.15in]
{\large Graduate School of Arts and Sciences}\\[0.9in]

{\Large\bfseries PREDICTABLE RECOVERY IN GEO-REPLICATED\\[0.35em]
STORAGE SYSTEMS UNDER PARTIAL\\[0.35em]
NETWORK PARTITIONS}\\[0.9in]

A Dissertation Submitted to the Faculty of the\\
Department of Computer Science\\
in Partial Fulfillment of the Requirements\\
for the Degree of\\[0.4in]
{\large DOCTOR OF PHILOSOPHY}\\[0.6in]

by\\[0.3in]
{\large Elena Marchetti}\\[0.3in]
Aldermere, July 2026
\end{center}

\vspace{0.5in}
\begin{flushleft}
Approved by the examining committee:\\[2.2em]
\committeeline{Priya Nadeau}{Professor of Computer Science, Chair}
\committeeline{Samuel Okoro}{Associate Professor of Computer Science}
\committeeline{Lena Fischer}{Professor of Electrical Engineering}
\end{flushleft}
\end{titlepage}

% ---------- Abstract ----------
\pagenumbering{roman}
\begin{center}
{\large\bfseries ABSTRACT}
\end{center}
\doublespacing
Distributed storage systems replicate data across geographic regions to survive
failures, yet the outages that dominate operational experience are caused by
partial network partitions in which connectivity is asymmetric rather than
absent. This dissertation develops a formal model of recovery latency under
such faults, proves bounds for leader-based and leaderless consensus protocols,
and presents a fault-synthesis framework that derives worst-case partitions
directly from protocol specifications. Evaluation on a 40-node testbed across
three storage systems demonstrates that synthesized faults reproduce production
incidents that randomized testing fails to find, and that partition-aware
reconfiguration reduces tail recovery latency by an order of magnitude.

\end{document}