openflow (3), internet architecture (2), experimental (2)
Public
comments
#1
posted on Sep 08 2009, 13:48
OpenFlow allows researchers to deploy new protocols on commodity hardware. This is an exciting new prospect that is necessary for adapting the Internet to new objectives. To achieve this capability, the system is constrained to use a controller to direct flows. How will this controller-centric architecture impact the structure of the experimental protocols being tested? What is the best way to achieve robustness with these controllers?
OpenFlow allows researchers to deploy new protocols on commodity hardware. This is an exciting new prospect that is necessary for adapting the Internet to new objectives. To achieve this capability, the system is constrained to use a controller to direct flows. How will this controller-centric architecture impact the structure of the experimental protocols being tested? What is the best way to achieve robustness with these controllers?