Date of Award
Department of Computer Science
The snapshot problem was first proposed over a decade ago and has since been well-studied in the distributed algorithms community. The challenge is to design a data structure consisting of $m$ components, shared by upto $n$ concurrent processes, that supports two operations. The first, $Update(i,v)$, atomically writes $v$ to the $i$th component. The second, $Scan()$, returns an atomic snapshot of all $m$ components. We consider two termination properties: wait-freedom, which requires a process to always terminate in a bounded number of its own steps, and the weaker obstruction-freedom, which requires such termination only for processes that eventually execute uninterrupted. First, we present a simple, time and space optimal, obstruction-free solution to the single-writer, multi-scanner version of the snapshot problem (wherein concurrent Updates never occur on the same component). Second, we assume hardware support for compare&swap (CAS) to give a time-optimal, wait-free solution to the multi-writer, single-scanner snapshot problem (wherein concurrent Scans never occur). This algorithm uses only $O(mn)$ space and has optimal CAS, write and remote-reference complexities. Additionally, it can be augmented to implement a general snapshot object with the same time and space bounds, thus improving the space complexity of $O(mn^2)$ of the only previously known time-optimal solution.
Ba, Khanh Do, "Wait-Free and Obstruction-Free Snapshot" (2006). Dartmouth College Undergraduate Theses. 49.