基于可信虚拟域的云边界安全防护模型与方法研究

Existing security mechanisms fail to scale towards elastic clouds and defocused cloud edges. As a result, the Cloud Edge Security has been a main concern for security industries and researches, with its solutions yet to be found. This project will be the first to propose a unified Cloud Edge Security Protective Model, enlightened by the concept of Trusted Virtual Domains. Grounded in this theory, this project extends its focus towards several aspects of current information security researches: Focusing on threats caused by the share of computing resources between multiple VMs, the project comes up with a Memory Coloring solution to better manage the resource allocation and improve VM isolation. To secure the weak and unstable Domain migration process, this project will research on the methods of Domain migration and policy synchronization, ensuring the confidentiality and integrity of VM migration procedures. To prevent unauthenticated data access between Domains of different security levels, this project will pursue proper access control and secure communication by new Attribute-Based Access Control and Group Key Agreement methods. The project also includes a system prototype of Secure Cloud Edge showing the feasibility of our models and methods. This research will seek breakthroughs in the Cloud Edge Security models and methods. It will surely add new bones into the studies of Cloud Security.

已有的安全机制无法满足云计算环境的动态性和边界泛化等特点，致使其边界安全成为信息安全领域亟待解决的核心问题。本项目借助可信虚拟域的概念，首先提出统一的云边界安全防护模型。并以此为理论依据，针对现有研究未能很好的解决虚拟机之间共享计算资源带来的安全隐患，项目从资源分配的角度，提出一种内存着色方案来改善虚拟机之间的隔离性；针对虚拟机跨域迁移时状态不稳定易受攻击的缺点，研究能够保证虚拟机跨域迁移数据机密性和完整性方法、迁移时安全策略的同步机制和安全监控策略。为了防止不同安全等级虚拟域、虚拟域之间数据的非法访问，研究虚拟域基于属性的访问控制与群密钥协商，实现虚拟域中数据的访问控制和安全通信；最后，搭建云计算虚拟边界原型系统来实证所提模型与方法机制的合理可行性。本项目力求在云环境下虚拟边界安全模型和部分相关机制方法方面取得突破，该研究理论与方法是对云计算安全领域的进一步丰富和创新。

由于云边界动态性和边界泛化的特点，其安全问题是云计算领域面临的核心安全问题。为此，本项目重点研究虚拟边界的安全隔离策略、研究虚拟机安全迁移方法以及研究虚拟域数据安全访问控制机制这三个基础科学问题与关键理论研究。针对当前云边界的动态性和数据资源的高度聚合性，本项目通过研究云计算环境安全威胁的起源，构建虚拟边界安全防护模型；针对云计算环境下虚拟机共享资源带来的安全隐患，提出一系列云环境下跨虚拟机侧信道的检测与防护方案；为了保持租户虚拟机的透明性和安全性，提出了基于虚拟机自省的漏洞检测和恶意行为监控方法；针对虚拟机跨域迁移时状态不稳定易受攻击的劣势，提出保证虚拟机迁移时机密性和完整性方法，和保证安全策略一致方法；针对不同等级虚拟域、虚拟机之间非法访问带来的的安全隐患，提出虚拟域数据安全访问控制策略和群密钥协商与管理机制。通过在VMM体系结构上引入各种监控、检测模块，对现有虚拟化环境进行安全增强，实现了边界安全应用环境。本项目采用的研究方案和关键技术强调对现有云边界状态的隔离与控制，与以往方法不同，是对虚拟化安全领域的丰富与创新。