Abstract—Dynamic software updating (DSU), the practice of updating software while it executes, is a lively area of research. The DSU approach most prominent in both commercial and research systems is in-place updating, in which patches containing program modifications are loaded into a running process. However, in-place updating suffers from several problems: it requires complex tool support, it may adversely affect the performance of normal execution, it requires challenging reasoning to understand the behavior of an updated program, and it requires extra effort to modify program state to be compatible with an update. This paper presents preliminary work investigating the potential for state transfer updating to address these problems. State transfer updates work by launching a new process running the updated program version and transferring program state from the running process to the updated version. In this paper, we describe the use and implementation of Ekiden, a new state transfer updating library for C/C++ programs. Ekiden seeks to redress the difficulties of in-place updating, and we report on our experience updating VSFTPD using Ekiden. This initial experience suggests that state transfer provides the availability benefits of in-place DSU approaches while addressing many of their shortcomings. I.

Recent years have seen significant advances in dynamic software updating (DSU) systems, which allow programs to be patched on the fly. However, a significant challenge remains: How can we ensure the act of applying a patch does not itself introduce errors? In this paper, we address this problem by presenting a new systematic testing methodology for updatable programs. Our idea is to transform standard system tests into update tests that execute as before, but each transformed test applies a patch at a different update point during execution. To mitigate the increase in the number of tests, we developed an algorithm for test suite minimization that finds a subset of update points that, if fully tested, yields the equivalent to full update point coverage. We implemented our approach and evaluated it on OpenSSH and vsftpd, two widely used server applications. We found that minimization is highly effective, reducing the number of update tests required for full coverage by 93%. 1.