CT-ISG: Designing Next-Generation, Reliable Internet Servers

计算机系统，尤其是关键服务器上的软件，正不断地变得越来越复杂。这些特性的增加增加了复杂性，并增加了重大的、持续的漏洞。在某种程度上，这可以追溯到设计决策，这些决策依赖于为新功能添加软件，而不是添加额外的硬件来支持更大的分离和增强的安全性。IT硬件成本的下降表明后一种方法对于安全系统开发可能更具成本效益。本研究的重点是研究从简单的、独立的计算机系统构建一个复杂的计算机服务器系统。这项工作将探索如何将众所周知的（但很少使用的）安全工程原理与较新的设计特性相结合，以生产高度安全的组件。该研究还将开发度量标准，以允许对结果系统与更传统的体系结构的脆弱性和收益进行客观比较。这个项目的预期结果将是：1)验证或否定一些安全工程技术的价值；2)一个服务器的演示原型，它应该比传统系统更安全，不受攻击；3)设计客观度量服务器漏洞的方法；4)能够在教育环境中使用原型的材料。

NSF 0523243 CT-ISG: Designing Next-Generation, Reliable Internet Servers Eugene H. SpaffordComputing systems, and especially software on critical servers, are constantly becoming more complex. This addition of features has contributed to greater complexity and to the addition of significant, ongoing vulnerabilities. In part, this may be traced to design decisions that have depended upon adding software for new functions rather than adding additional hardware to support greater separation and enhanced security. The falling cost of IT hardware suggests that this latter approach may be more cost-effective for secure system development. The focus of this research is to investigate the construction of a complex computer server system from simpler, separate computer systems. The effort will explore how to apply well-known (but seldom used) security engineering principles coupled with newer design features to produce highly-secured components. The research will also develop metrics that will allow objective comparison of the vulnerabilities and benefits of the resultant system against more conventional architectures. The expected outcomes of this project will be: 1) validation or disproval of the value of a number of security engineering techniques; 2) a demonstration prototype of a server that should be more secure against attack than are more conventional systems; 3) an approach to designing objective metrics for measuring server vulnerabilities; 4) materials to enable use of the prototype in educational settings.