SDCI Data New: Trust Management for Open Collaborative Information Repositories: The CalSWIM Cyberinfrastructure

加州大学欧文分校的克里斯蒂娜·维迪拉·洛佩斯(PI)、迈克尔·古德里奇(共同PI)和斯坦利·格兰特(共同PI)与布朗大学的Roberto Tamassia(分包商)加州可持续流域信息经理(CalSWIM：http://calswim.org))是一个支持南加州分水岭信息传播和分析的网络基础设施。CalSWIM为分水岭利益相关者提供了什么？科学家、监管者和普通民众？访问河流监测数据和相关的地理信息系统地理空间数据，以及使用各种复杂的建模工具分析数据的能力。与许多其他科学网络基础设施一样，CalSWIM网站一直是一个封闭的基础设施，即其更新由少数人控制，对其进行的任何更改都需要通晓网络知识的专业人员的干预。这种模式不符合CalSWIM项目的雄心壮志。为了使这一网络基础设施可持续、可扩展、向外延伸并不断更新，我们目前正在将CalSWIM转变为开放协作信息库，其形式为维基，具有高级内容管理功能，包括数据库、地理信息系统层和其他形式的数字数据。通过以维基方式打开CalSWIM，我们使领域专家能够直接更新站点，而不需要他们具有Web编程知识。我们还提供了一个框架，让普通公众参与这一进程，从而受益于公众对内容的严格监督，并受益于在封闭的基础设施中无法发生的宝贵的志愿者贡献。目前的工作旨在创建一本可公开更新的百科全书？一切事物的分水岭？这包括加州的所有分水岭。开放的信息基础设施，如我们正在开发的，有时会受到用户的怀疑。维基百科的混乱本质？最显眼的协作信息库之一？将不适用于支持科学和监管信息协作的开放网络基础设施。只有在设计中考虑到信任问题，CalSWIM维基才会成功。信任是一个基于个人和社会因素的复杂概念，它包括关于数据安全的保证、关于数据来源的知识、关于作者声誉的知识，以及关于作者在不受控制的情况下的归属(引文和引文)的保证。该项目将在开放的信息存储库(如CalSWIM wiki)上设计、开发和部署信任层，该信任层将最大限度地减少劣质信息的存在，并最大限度地减少恶意信息操纵的影响。我们的方法基于访问控制系统、信誉管理系统和数据完整性算法的混合。更具体地说，该项目的成果将是：(1)专门设计的新的信任管理模式，以满足协作的科学和监管信息库的需求，其中添加数据和创建数据分析对科学家、监管机构、利益集团和公众开放。这种信任模式将作为一个软件包实现，并将作为开放源码公开提供。(2)确保所用数据(即？报价？)的完整性的新方法和可视化系统在基于加密散列函数和校对的不受控制的上下文中。(3)前两项在一个具体的网络基础设施CalSWIM中的具体部署，以及对它们在管理信任方面的有效性的评估。虽然CalSWIM wiki是将部署该信任管理组件的具体网络基础设施，但本项目中开发的信任管理组件可以应用于共享CalSWIM属性的各种协作信息库？也就是说，科学家、监管机构和公众之间基于数据的合作。我们将传播这一项目的成果，并通过使用开放源码模式并向那些可能从采用这一模式中受益的组织进行外联努力，将其广泛提供给其他群体。

SDCI Data New: Trust Management for Open Collaborative Information Repositories: The CalSWIM CyberinfrastructureCristina Videira Lopes (PI), Michael Goodrich (co-PI), and Stanley Grant (co-PI) of University of California, Irvine, with Roberto Tamassia (subcontractor) of Brown UniversityThe California Sustainable Watershed Information Manager (CalSWIM: http://calswim.org) is a cyberinfrastructure supporting the dissemination and analysis of watershed information in Southern California. CalSWIM provides watershed stakeholders ? scientists, regulators, and the general public ? access to stream monitoring data and relevant GIS geospatial data, and the ability to analyze data with modeling tools of varying sophistication. Like so many other scientific cyberinfrastructures, the CalSWIM web site has been a closed infrastructure, i.e. its update has been controlled by a small number of individuals, and any changes to it require the intervention of web-knowledgeable professionals. This model does not scale to the ambition of the CalSWIM project. In order to make this cyberinfrastructure sustainable, expandable, outreaching, and continually up-to-date, we are currently transforming CalSWIM into an Open Collaborative Information Repository, in the form of a wiki with advanced content management features, including databases, GIS layers, and other forms of digital data. By opening CalSWIM in a wiki manner, we empower the domain experts to update the site directly, without requiring them to have web programming knowledge. We also provide a framework for engaging the general public in the process, therefore benefiting from the public's critical supervision of the content, and from valuable volunteer contributions that cannot occur in a closed infrastructure. The current work aims at creating a publicly updatable encyclopedia of ?all things watershed? that includes all watersheds in California. Open information infrastructures such as the one we are developing are sometimes met with skepticism from their users. The chaotic nature of Wikipedia ? one the most visible collaborative information repositories ? will not be appropriate for open cyberinfrastructures supporting collaboration over scientific and regulatory information. The CalSWIM wiki will only succeed if issues of trust are taken into consideration in its design. Trust is a complex concept based on personal and social factors, that includes assurances about data security, knowledge about data provenance, knowledge about authors' reputations, and assurances about author attribution (citations and quotations) in uncontrolled contexts. This project will design, develop, and deploy a trust layer on top of open information repositories, such as the CalSWIM wiki, that will minimize the existence of poor-quality information and also minimize the effect of malicious information manipulation. Our approach is based on a mixture of access control systems, reputation management systems, and data integrity algorithms. More specifically, the outputs of this project will be: (1) A new trust management model specifically designed to address the needs of collaborative scientific and regulatory information repositories, where the addition of data and the creation of data analyses is open to scientists, regulatory agencies, interest groups, and the general public. This trust model will be implemented as a software package, and it will be made publicly available as open source. (2) A new method and visualization system for ensuring the integrity of data used (i.e., ?quoted?) in uncontrolled contexts based on cryptographic hash functions and prooflets. (3) A concrete deployment of the previous two items in one specific cyberinfrastructure, CalSWIM, along with an assessment of how effective they are in managing trust.Although the CalSWIM wiki is the concrete cyberinfrastructure on which this trust management component will be deployed, the trust management component developed in this project can be applied to a variety of Collaborative Information Repositories that share the properties of CalSWIM ? namely, data-based collaboration between scientists, regulatory agencies, and the general public. We will disseminate the results of this project and make it widely available to other groups by using the open source model and by making an outreach effort to those organizations that could benefit from adopting this model.