A branch of the US Department of Energy wants to take software-defined network technology and combine it with a high-speed underlying optical system that will be capable of supporting large scientific applications.
+More on Network World: What network technology is going to shake up your WAN?+
Specifically the Office of Advanced Scientific Computing Research in the Office of Science at the U.S. DOE said that networks are becoming too complicated to manage and control, especially as they provide critical support for sophisticated distributed extreme-scale science activities and Big Data-intensive scientific collaborations.
The group says the answer to those challenges lies in emerging technologies such as SDN, Network Function Virtualization (NFV), and Network Service Chaining (NSC), as well as recent advances in related fields such as machine learning and parallel computing.
The focus of this specific funding announcement is to develop a new generation of SDN-enabled High Performance Networks (SDN-HPON) that are agile, open to rapid innovations, and dynamically reconfigurable with automated and intelligent capabilities, the group stated.
An SDN-HPON system has capabilities and requirements that “far outstrip conventional SDN networks optimized for Internet applications” and should include features such as:
a) Multi-domain network capabilities including guaranteed on-demand services optimized for high-end scientific applications;
b) Terabits/sec capacity in the optical backbone network and the ability to deliver at least 40- 100 Gbps end-to-end performance to network-intensive applications;
c) Embedded network intelligence and automation in the design of critical functions and services to reduce complexities for both end users and network operators.
+More on Network World: Hot stuff: The coolest drones+
“The key to realizing the promise of SDN lies in the ability to harness controller programmability, which enables automation and the development of intelligent services to dynamically control the behavior of networks and applications. In extreme-scale science environments, these capabilities will make it possible for network-intensive application workflows to dynamically interact with the network fabric through northbound open APIs, some of which are yet to be defined and standardized,” the group stated.
In the end the DOE office wants intelligent and automated technologies that simplify network usage for scientists, automate complex network control and management functions, and improve the productivity and performance of network-intensive science.
If you want to know more about this project you can download the whole funding request here.
Check out these other hot stories: