Software Provenance: Track the Reality Not the Virtual Machine

The growing use of computers and massive storage by individuals is driving interest in digital preservation. The scientific method demands accountability through digital reproducibility, adding another strong motivation for preservation. However, data alone can become obsolete if the interactivity of software required to interpret the data is lost. Virtual machines (VMs) may preserve interactivity however do so at the cost of obscuring the nature of what lies within. Occam, instead, builds VMs on-the-fly while storing and distributing well-described software packages. Thus, the system can track the exact components inside VMs without storing the machines themselves, allowing software to be repeatably built and executed. For Occam to recreate VMs, it needs to know exactly what software was used within. Through this tracking, such software can even be modified and rebuilt. Occam keeps track of all such components in manifests, allowing anybody to know exactly what is in each VM, and the origins of each component.