A method and device for enhancing the introspection security of a virtual machine on an arm platform

Abstract

The invention provides a method and a device for enhancing ARM platform virtual machine introspection security. The device comprises a security enhancing interface front end located in an untrusted environment virtualization software stack, a security enhancing interface at a high privilege level and a core library located in a hardware isolation environment Trust Zone. The method and the device resist information leak and guarantee the accuracy of VMI codes and data in a whole life period, namely guaranteeing the operation security of VMI and accuracy of VMI bottom layer input data, via the security enhancing interfaces independent of Hypervisor or QEMU and the proper hardware isolation environment Trust Zone. The method and the device solve two layers of semantic isolation under an ARMv8 platform, namely Hypervisor semantic isolation and VM semantic isolation.

Description

technical field [0001] The invention relates to the technical field of self-introspection security of a virtual machine in a security vulnerability host machine, in particular to a method and a device for enhancing the self-inspection security of an ARM platform virtual machine. Background technique [0002] As we all know, the general virtual machine introspection (VMI) code is deployed in the user space of the host (Host), accesses the binary data inside the virtual machine (VM) through the virtual machine manager (Hypervisor), and bridges the semantic isolation problem inside the VM These binary data are converted into effective VM internal state information at the operating system level, so that the state information can be used for further security monitoring. Therefore, the normal operation of the existing VMI depends on the premise that both the Host and the Hypervisor are safe and credible. This means that once this premise is broken, normal operation of the VMI can...

AI Technical Summary

Problems solved by technology

[0003] VMI in a vulnerable host environment faces multiple security threats. One is information leakage: VMI technology breaks the isolation between multi-tenants in the cloud environment. The malicious use of VM internal information obtained by VMI will cause serious leakage of VM information.

The second is that VMI is not safe when running: VMI runs in user mode. Once the vulnerability of the host environment is exploited, the code and data of VMI will be exposed to attackers.

The third is that the underlying input data of VMI is incorrect: VMI relies on Hypervisor to obtain VM-related core data, which gives Hypervisor and Host the opportunity to tamper with these core data

The industry has not been involved in the research on VMI runtime insecurity and incorrect underlying input data.

Images