基于可信包装的可信软件构造模型-计算机系统结构专业论文.docxVIP

I I 摘 要 摘 要 传统的软件开发方法在软件可信性保障方面考虑较少，致使所开发软件的可信性难 以保证，而加载有TPM的可信计算机只能在系统启动的初始阶段获得暂时的安全，一旦 软件在运行过程中发生故障或失效，尤其是遭到恶意攻击时，将给用户和社会带来巨大 的冲击和损害。软件的可信性作为软件质量的一种特殊表现形式受到了高度关注，可信 软件能够保证软件的行为总是与预期相一致，使得软件的功能需求得到正确的实现，并 且在很大程度上可信软件不会因为系统中存在的错误、环境的异常或恶意的攻击而导致 软件的故障或失效。因此，如何保证软件的可信性以及如何构造可信软件将成为软件技 术发展和应用的必然趋势，具有十分重要的理论意义与研究价值。 结合Agent技术和TPM提供的可信计算功能，提出了基于可信包装的可信软件构造 模型（Trusted Software Constitution Model based on Trust Shell，简称TSCMTS），在逻辑 上可信包装负责保证软件的可信性。提出可信引擎的概念，将TCG信任链延伸到了应用 实体，对软件实体进行完整性度量，保证其静态可信性；同时定义了可信视图的概念， 表征软件行为特征信息。基于此，可信引擎通过验证软件运行时其行为是否满足可信性 约束，实现对软件行为的动态可信性的保障。为了提高可信性约束的精度，提出了基于 信息熵的检查点场景和时间偏量多属性权重分配策略。仿真实验表明该模型所设计的软 件能够在降低较小性能的前提下有效地提高可信性。 关键词 可信软件构造 可信包装 可信引擎 可信视图 软件行为轨迹 II II Abstract Abstract Because the guaranty of trustiness wasn’t considered enough in traditional software development methods, the software developed in that methods lack effective measures for ensuring its trustiness. However, trusted computer based on TPM can be of temporary trustiness in the initial phase, once software failures and malfunctions occur, especially when software is attacked maliciously, it will bring tremendous loss. The trustiness as a special representation of software quality has drawn more and more people’s attentions, trusted software can ensure that the software is always works in the intended way and goes towards the intended direction and will not result in malfunction or failure largely even if caused by malicious attacks, or system errors. How to ensure the trustiness of software will be an inexorable trend of software development and application. Therefore, it is of great theory significance and practical value to conduct research on the constitution of trusted software. Combining agent technique with the support of trusted computing provided by TPM, a trusted software constitution model based on Trust Shell （TSCMTS） is demonstrated in this paper, Trust Shell ensures the trustiness of software logically. The concept of Trust Engine is proposed, which extends the “chai