TWC: TTP Option: Small: Collaborative: SRN: On Establishing Secure and Resilient Networking Services

几乎每个组织都依赖于基于云的服务。基于云的服务的后端是为多个租户设计的，并且驻留在分布在多个物理位置的数据中心中。在这样一个复杂、共享的环境中，网络安全和安全管理是主要障碍。本研究探讨了采用移动目标防御（MTD）方法来减轻安全挑战。不断调整数据中心的拓扑结构、带宽分配、流量策略等系统资源，使攻击者难以攻破系统。将制定新的评价方法，以确保这些中期发展机制在实践中正常工作。这项研究的结果是拥有更安全、更能抵御攻击的云服务。这项研究是由来自亚利桑那州立大学、杜克大学和密苏里大学堪萨斯城分校三所不同大学的研究人员共同完成的。研究生将接受培训，以满足日益增长的网络安全专业人才教育需求。拟议的研究结果将纳入各院校的几门课程。多位置、多租户数据中心环境中的MTD方法需要复杂的协调级别。本研究研究了基于可编程网络解决方案的数据中心虚拟网络环境中的防御机制，从而可以在考虑系统资源消耗、软件缺陷/漏洞、对策有效性以及对运行应用程序的消费者的影响的情况下部署主动攻击对策。研究成果可用于需要以非常精细的粒度分辨率（从几毫秒到几秒）准确预测安全态势感知变量的应用。这带来了额外的挑战，即为网络、数据收集、支持大数据的安全处理和控制开发新的性能模型。为了应对这些挑战，该项目有两个相互依存的基础研究重点：(a)在网络和软件层面研究动态和适应性防御框架；(b)部署一种自适应安全流量工程方法，通过考虑对策和网络带宽分配的有效性来选择最佳对策，同时最大限度地减少对应用程序的侵入性和部署对策的成本。这个项目的结果将包括一组软件api和工具，用于将度量系统和分析模型集成到实践工作的过渡中。

Almost every organization depends on cloud-based services. The backend of cloud-based services are designed for multiple tenants and reside in data centers spread across multiple physical locations. Network security and security management are major hurdles in such a complex, shared environment. This research investigates mitigating the security challenges by taking a moving target defense (MTD) approach. Continually adjusting the system resources such as the topology of the data center, bandwidth allocation and traffic flow policies makes it difficult for attackers to compromise the system. New evaluations methods will be developed to ensure that these MTD mechanisms work properly in practice. The outcome of this research is to have cloud services that are more secure and resilient to attacks. This research is a collaborative effort conducted by researchers from three different universities, Arizona State University, Duke University, and the University of Missouri-Kansas City. Graduate students will be trained to serve the growing need for educating professionals in cyber security. The results of the proposed research will be incorporated into several courses taught at the respective institutions. The MTD approach in a multi-location, multi-tenant data center environment requires a complex level of coordination. This research investigates defense mechanisms in the data center's virtual networking environment based on programmable networking solutions so that proactive attack countermeasures can be deployed with considerations of the system resource consumption, software bugs/vulnerabilities, effectiveness of countermeasures, and impact on consumers running applications. The research outcomes can be employed for applications that require security situation-awareness variables accurately predicted at a very fine grain resolution, from a few milliseconds to a few seconds. This introduces additional challenges, namely, developing new performance models for networking, data collection, big data-enabled security processing, and control. To address these challenges, this project has two interdependent fundamental research thrusts: (a) investigate a dynamic and adaptive defensive framework at both networking and software levels; and (b) deploy an adaptive security-enabled traffic engineering approach to select optimal countermeasures by considering the effectiveness of countermeasures and network bandwidth allocations while minimizing the intrusiveness to the applications and the cost of deploying the countermeasures. The outcomes of this project will include a set of software APIs and tools to integrate the measurement system and analytical models in a transition to practice effort.