Collaborative Research: CC-NIE Integration: Developing Applications with Networking Capabilities via End-to-End SDN (DANCES)

由网络工程师、应用程序开发人员和研究科学家组成的DANCES项目团队正在实施一个支持软件定义网络（SDN）的端到端环境，以优化对科学数据传输的支持。 DANCES通过将高性能计算作业调度、SDN沿着提供的网络控制功能与端到端网络基础设施中的数据移动应用集成在一起来实现这种优化。 这种集成提供了对用于管理网络带宽的控制机制的访问。SDN对网络资源的控制增强了应用的稳定性、可预测性和性能，从而提高了整体网络利用率。DANCES项目的动机是将先进网络服务的优势应用于拥挤的城域网和校园网问题。 DANCES使用XSEDENet在Internet 2上与安装在协作站点的支持OpenFlow的网络交换机一起作为端到端的硬件和软件基础。通过DANCES获得的知识通过参与机构提供的教育计划以及现有的社区研讨会，会议和会议进行传播。 通过DANCES获得的见解和经验将促进更好地理解在广域和园区网络基础设施中支持端到端SDN的技术要求。 由此产生的支持SDN的应用程序将使最终用户能够轻松访问高带宽连接的请求和配置，并提高网络性能和可预测性，以支持广泛的应用程序。

The DANCES project team of network engineers, application developers, and research scientists is implementing a software-defined networking (SDN)-enabled end-to-end environment to optimize support for scientific data transfer. DANCES accomplishes this optimization by integrating high performance computing job scheduling, network control capabilities offered by SDN along with data movement applications in an end-to-end network infrastructure. This integration provides access to control mechanisms for managing network bandwidth. The control of network resources enabled by SDN enhances application stability, predictability and performance, thereby improving overall network utilization. Motivation for the DANCES project is to apply the advantages of advanced network services to the problem of congested metropolitan and campus networks. DANCES uses XSEDENet across Internet2 in conjunction with OpenFlow-enabled network switches installed at the collaborating sites as the end-to-end hardware and software substrate.Knowledge gained through DANCES is being disseminated through educational programs offered by the participating institutions and at existing community workshops, meetings, and conferences. The insights and experience obtained through DANCES will promote a better understanding of the technical requirements for supporting end-to-end SDN across wide area and campus cyberinfrastructure. The resulting SDN-enabled applications will make the request and configuration of high bandwidth connections easily accessible to end users and improve network performance and predictability for supporting a wide range of applications.