CC* Integration: Regional Embedded Cloud for As-a-Service Transformation (RECAST)

点击翻译按钮获取中文摘要

Science DMZs (SDMZ) have been widely deployed in the Research and Education networking community and recognizedas a global best practice for supporting cyberinfrastructure and big data applications. Despite innovationsin performance, security, and management, the SDMZs remain focused on thesupport of big data projects and cyberinfrastructure. It remains difficult to create physical "diskto disk" or virtual "scientist to scientist" connections. Researchers are increasingly interestedin an emerging class of hybrid services focused on turnkey, on-demand integration of private/publicclouds, cyberinfrastructure, scientific resources, and high performance networks.The RECAST project will develop and deploy a regional based Software Defined ScienceDMZ (SD-SDMZ) as a mechanism to provide flexible edge services that include traditional datatransfer functions. RECAST leverages state-of-the-art Cloud and Software-defined networking technologies to transforman SDMZ -- that was historically focused on high-speed data transfer -- into a Software Defined Infrastructure that offerscompute and network services in a fully virtualized, programmable and orchestrated fashion. Built upon technologies similar to thoseused in modern data centers, this "cloudified SDMZ" will provide a scalable, multi-tenant environment, whereeach user or user group can allocate dedicated resources and use the resources in an isolated,Independent, and elastic fashion. Embedding this resource in the University of Maryland Mid-AtlanticCrossroads regional network will allow the SD-SDMZ to leverage rich connectivity to advancecyberinfrastructure to provide these new and advanced services.