Effective runtime power management requires hardware activity to be tracked at a very fine granularity in both space and time in order to meet diverse workload performance requirements within a tight power budget. As the available instantaneous power budget itself represents a shared resource, this practically translates into finding the optimal allocation of the power budget among active tasks of execution. Covert communication over a previously unexplored class of channels thereby becomes possible, which forms the focus of this paper.
Bibliographical noteFunding Information:
This work is supported in part by the NSF CAREER Award under Grant CCF-1350451, in part by the NSF/SRC Award under Grant CNS-1715286, and in part by the Cisco Systems Research Award.
This work is supported in part by the NSF CAREER Award under Grant CCF-1350451
- Power management vulnerabilities
- covert channels