Desktop Grid Computing, chapter 10: Security and Result Certification
Authors
Abstract
As in many other areas of computing, security and dependability are of paramount importance in the desktop grid ecosystem. In such environments, threats exist for both the donated resources and the submitters. On one hand volunteers need protection against code that inadvertently or maliciously might harm their resources. Indeed, a volunteer project disrupting the regular working of the volunteer machines would have serious public relation consequences seriously hindering the public perception of volunteer computing. On the other hand, the behavior and results produced by volunteers need to be properly scrutinized to detect inappropriate actions and/or erroneous results, and thus ensure the dependability of the performed computations.Therefore, security and dependability are two major goals that need to be carefully considered for a successful desktop grid experience. However, as we shall see in this chapter, security and dependability are complex issues, with many desktop grid frameworks addressing both issues in a less than perfect way. Thus, the volunteering of resources and submission of tasks still require a fairly amount of implicit trust between donors and harvesters.
This chapter aims to present the state of the art regarding both security and dependability for desktop grids. The first part presents an overview of the existing issues regarding security and dependability of desktop grid computing. The second part provides a formal approach on several techniques oriented toward the efficient detection of malicious behavior, namely collusion.