EPA-RIMM-V: Efficient Rootkit Detection for Virtualized Environments
The use of virtualized environments continues to grow for efficient utilization of the available compute resources. Hypervisors virtualize the underlying hardware resources and allow multiple Operating Systems to run simultaneously on the same infrastructure. Since the hypervisor is installed at a h...
| Main Author: | |
|---|---|
| Format: | Text |
| Language: | English |
| Published: |
PDXScholar
2018
|
| Subjects: | |
| Online Access: | https://pdxscholar.library.pdx.edu/open_access_etds/4485 https://pdxscholar.library.pdx.edu/cgi/viewcontent.cgi?article=5556&context=open_access_etds |
| Summary: | The use of virtualized environments continues to grow for efficient utilization of the available compute resources. Hypervisors virtualize the underlying hardware resources and allow multiple Operating Systems to run simultaneously on the same infrastructure. Since the hypervisor is installed at a higher privilege level than the Operating Systems in the software stack it is vulnerable to rootkits that can modify the environment to gain control, crash the system and even steal sensitive information. Thus, runtime integrity measurement of the hypervisor is essential. The currently proposed solutions achieve the goal by relying either partially or entirely on the features of the hypervisor itself, causing them to lack stealth and leaving themselves vulnerable to attack. We have developed a performance sensitive methodology for identifying rootkits in hypervisors from System Management Mode (SMM) while using the features of SMI Transfer Monitor (STM). STM is a recent technology from Intel and it is a virtual machine manager at the firmware level. Our solution extends a research prototype called EPA-RIMM, developed by Delgado and Karavanic at Portland State University. Our solution extends the state of the art in that it stealthily performs measurements of hypervisor memory and critical data structures using firmware features, keeps performance perturbation to acceptable levels and leverages the security features provided by the STM. We describe our approach and include experimental results using a prototype we have developed for Xen hypervisor on Minnowboard Turbot, an open hardware platform. |
|---|