280 likes | 381 Vues
This senior thesis explores the design and implementation of a trusted computing system using OpenSolaris, TPM, and DTrace tools. The solution focuses on virtual container attestation for enhanced security in computing environments. The study delves into the challenges, tools, and results of testing applications. Future work includes improving security checks and communication among zones.
E N D
Virtual Container Attestation: Customized trusted containers for on-demand computing. • Katelin Bailey • Senior Thesis 2010 • Dartmouth College • Department of Computer Science
Where are we going? • Introduction • The Problem of Trusted Computing • Tools: OpenSolaris, TPM, DTrace • Design & Implementation • Motivation for the Testing Applications • Testing Applications. • Results & Conclusions
The Problem of Trusted Computing • Why do we need to trust computers? • How can we develop that trust?
Previous Approaches • Attestation • Property-based attestation • Compartmented attestation • Virtualization • Trusted Computing on Demand
OpenSolaris • Zones (containers) • DTrace • Open-source
Zones • OS-level virtualization is lightweight • Global zone’s window into the containers • Zone cloning • Easy configuration • More complete virtualization, not just process isolation
TPM • Cryptographic Capabilities • Platform Control Registers • Trusted Root • Trusted Boot • In relation to Trusted Computing
Virtual Container AttestationThe Goals • Uses client-requested containers • Interface to local and remote machines • Remain usable to client applications • Employs property-attributed certificates • Monitors attributes of each container • Halts zones which do not comply • Ensures that revoked zones remain inactive
In summary... • Flexibility of policy • Containers on demand • Isolation • Policy enforcement • Simple property attestation
Open source software as the basis for the testing applications Unfortunately, we had to create our own...
Power Grid Software • Input comes from device measurements • Format the incoming data • Process in any (possibly multiple) way • Export for large-scale processing • Format/prepare the outgoing data
Hurdles • Zone startup times • TSS stack
Future Work • Fix the hurdles! • Varied revocation scheme • Additional security checks • Negotiation of security • Better zone communication