The prototype was built with contributions and expertise from AT&T, IBM and ACS, and the work was performed under the auspices of the U.S. Government's DARPA CORONET programme, which focuses on rapid reconfiguration of terabit networks.
"The programme was visionary in anticipating the convergence of Cloud computing and networking, and in setting aggressive requirements for network performance in support of Cloud services", stated Ann Von Lehmen, the ACS programme lead.
AT&T was responsible for developing the overall networking architecture for this concept, drawing on its industry leadership in bandwidth-on-demand (BoD) technologies and advanced routing concepts. IBM provided the Cloud platform and intelligent Cloud data centre orchestration technologies to support dynamic provisioning of Cloud-to-Cloud communications. ACS contributed its expertise in network management and innovation in optical-layer routing and signaling as part of the overall Cloud networking architecture.
Cloud computing already has widespread impact across how we access todays applications, resources, and data. Today's traditional Cloud computing model is built on the premise of automation and lower operational costs, which requires dynamic provisioning of resources. However, the traditional Cloud-to-Cloud network is static and creating it is labour intensive, expensive and time consuming.
In response to the rapid advent of Cloud-based services and explosion in data centre size and scope, Cloud Service Providers (CSPs) have installed automatic and intelligent resource management systems within their data centres. For example, these systems can load balance both processor and storage resources, as well as perform massive transfers of data among multiple data centres.
"These shifts have driven the need to develop rapid and high rate bandwidth-on-demand in the Wide Area Network (WAN)", stated Robert Doverspike, executive director of Network Evolution Research at AT&T Labs. "By combining software defined networking (SDN) concepts with advanced, cost-efficient network routing in a realistic carrier network environment, we have successfully demonstrated how to address this need."
This prototype was implemented on OpenStack, an open-source Cloud-computing platform for public and private Clouds, elastically provisioning WAN connectivity and placing virtual machines between two Clouds for the purpose of load balancing virtual network functions. The use of flexible, on-demand bandwidth for Cloud applications, such as load balancing, remote data centre back-up operation, and elastic scaling of workload, provides the potential for major cost savings and operational efficiency for both CSPs and carriers.
This technology not only represents a new ability to scale Big Data workloads and Cloud computing resources in a single environment but the elastic bandwidth model removes the inefficiency in consumption versus cost for Cloud-to-Cloud connectivity", stated Douglas Freimuth, IBM Research Senior Technical Staff Member and Master Inventor. "IBM Research brought a unique understanding of both Cloud environments and networking infrastructures which made us an ideal collaborator for this project."
Instigated in 2007, the DARPA CORONET programme seeks to develop the target network architectures and technologies needed to build next-generation bandwidth on demand services. During Phase 3 of the multi-phase project, which was completed at the end of May, the current group of collaborators brought together the necessary cutting-edge research and industry expertise to potentially deliver this technology from a research lab to commercialization.
The DARPA long-term commitment to research was vital to kick-start these ideas and this current proof-of-concept trial is a key enabler along the path to commercialization in the telecommunications industry.
Further reinforcing the industry need for this type of technology, AT&T recently announced its vision for the network of the future - titled the User-Defined Network Cloud. This transformative initiative to move to a Cloud-based architecture will utilize SDN tools in the WAN to create a programmable network that is more flexible, efficient and aware of applications.
Complementary BoD capabilities are already commercially available between AT&T's network enabled Cloud solution, AT&T NetBond - which allows network capacity to scale or contract on demand based on the Cloud workload's needs - and public Cloud services, such as IBM's Cloud Managed Services. When coupled together, these solutions enable highly secure, reliable and dynamic connectivity between a customer's MPLS-VPN network and IBM's public Cloud services.
In the demonstration, the IBM Cloud platform and orchestration technology manages the life cycle of Virtual Machine (VM) network applications on OpenStack software to automatically monitor server load and request both Cloud-to-Cloud network bandwidth from a SDN WAN Orchestrator developed by AT&T and compute resources as needed for VM migration.
The AT&T SDN WAN Orchestrator automatically routes data server connection requests across the appropriate network layer: IP/MPLS, subwavelength or Dense Wavelength Division Multiplexing (DWDM). Rapid, robust provisioning protocols developed by ACS are integrated with commercial transport DWDM network elements to set up and tear down connections as needed.
In the demo, set-up times as short as 40 seconds were achieved, with sub-second provisioning times possible with next generation DWDM equipment, called ROADMs. This approach also takes BoD into a truly dynamic regime, by enabling the high-connection request rates that will be required in future Cloud service environments.
"The DARPA CORONET program has created a truly innovative solution that will enable dynamic Cloud services of the future. By delivering highly efficient bandwidth sharing, this technology will significantly reduce costs for both carriers and cloud providers. It truly is a major step forward for the industry", stated Matt Goodman, DARPA Programme Manager.