网络系统可存活性理论的扩展及实证研究

The concept of survivability originated from the study of military survivability and nuclear deterrence strategy in the 1960s, but the currently prevalent survivability analysis and survivable network systems rapidly emerged in the last decade are near a total new scientific theory and engineering discipline, except that they preserve a core idea of survivability. The core idea is that survivable system must survive catastrophic events by either preserving critical functionalities (resistance) or quickly recover from the loss of critical functionalities (resilience). The technology and social background for the paradigms to emerge was that starting in 1990s, computer networks increasingly control a nation's critical infrastructures, whose incapacity or destruction would have a debilitating impact on defense or economic security. It was initiated as a research initiative to implement an Executive Order of the US President George Bush on the protection of critical national infrastructures in 1999, which specified 7 categories of critical infrastructures, including emergency services and continuity of government services. Since 1999, multiple US Presidents' executive orders, related to this agenda, have been issued. In recent years, research on survivability has penetrated into and also draw from other relevant fields in computer science and engineering, most notably, security, reliability, dependability, and safety. Essentially, in the US, survivability is the civilian version of information warfare or defensive information warfare. Nevertheless, there have been few significant advances in quantitative research on survivability due to its necessity in dealing with potentially drastic uncertainties such as human malicious acts. In his 2008 dissertation, Ma (2008) introduced a few novel concepts, mathematical methods, and the strategies for designing and managing survivable network systems, including UUUR (uncertain, latent, unobserved, and unobservable risks), Byzantine generals 'playing' evolutionary game, dynamic hybrid fault models, three-layer survivability, and extended evolutionary game theory. In the last few years, the work has been extended into a general modeling and simulation architecture that can be applied to the study of complex engineering, biological (ecological), and social systems, where the risk analysis and management are a core issue. The overall goal of the proposed project is to extend the results derived from Ma et al's works (2008-2011, about 20 publications in premier platforms) by harnessing the latest advances in network science, and further implement the resulted advances in a cloud computing platform. The research is expected to generate significant theoretical and technological ramifications including survivable, cloud-computing supported emergency response and management systems.

可存活性概念起源于上世纪冷战时期对军事可存活性以及核威慑战略的研究。 现今的可存活性理论虽然仍保留了军事可存活性一些核心概念，本质则是近20年来迅速发展起来的、以计算机网络为核心的、近乎全新的科学理论。 可存活性最核心概念，也是区别于其它相关概念（可靠性、可信性、安全性等）之处在于: 可存活系统必须在遭受毁灭性打击或灾难发生后依然保持或迅速恢复其核心机能。 国际间美国布什当局于1999年颁布了关于保护"应急服务""政府连续性"等7七大国家关键设施的总统政令。可存活性理论和可存活网络系统工程研究则是学术界对该政令的回应。本项目旨在扩展深化网络可存活性理论, 并通过开发必要技术在云计算环境下实证其理论扩展。重心之一是将申请人在可存活性领域的最新研究进展与复杂网络科学相结合。项目预期成果将为未来"国家经济安全关键设施保护"，例如"云计算支撑的可存活应急管理系统"提供坚实的理论基础和核心技术原型。

本项目旨在扩展网络可靠性、可存活性、和稳定性理论, 并通过开发必要技术和软件将其应用于现实复杂系统。重心（也是特色）之一是将主持人在可存活性领域先前的创新研究与复杂网络科学相结合，研发出了一系列全新的网络分析概念、理论和技术，并开展应用示范研究。重心之二是采用了人类基因（宏基因）大数据验证其理论和技术，并探索人体菌群生态系统网络稳定性及其与疾病关系。下面列举部分重要成果：.（1）完成专著《Reliability, Survivability and Resilience: A Unified Theoretical Approach with Applications to Network, Engineering, and Socioeconomic Systems》已经与德国Springer签订出版合同。（2）提出复杂网络“Trio motif”(三角主题）的概念，并提出12个能够代表网络特征的三角主题，开发出了用于搜寻这些三角主题的软件。其发明专利已正式受理。（3）复杂网络“P/N Ratio”正负比率的概念、计算方法和软件， 其发明专利正式受理。（4）提出复杂网络“Dominance”新概念和数学模型，并开发出分析计算软件。 该成果发表在美国生态学会旗舰期刊Ecological Monographs》 SCI影响因子=9.9； 其发明专利正式受理。（5）提出基于云计算、可存活性为中心的应急管理系统设计，其论文 “A Survivability-Centered Research Agenda for Cloud Computing Supported Emergency Response and Management Systems”发表在IEEE-AIAA会议。（6）发表SCI论文20篇，累计影响因子85，单篇平均影响因子4.2。（7）受理发明专利12项。（8）三款大型软件：DBG2OLC, Sparc, HPTree.[1] DBG2OLC https://sites.google.com/site/dbg2olc/ .[2] Sparc https://sourceforge.net/projects/sparc-consensus/ .[3] HPTree http://lab.malab.cn/soft/HPtree/