User Defined Networks
Within GTS, the researcher is able to describe the logical composition of their desired network. The user may incorporate general purpose computational servers, advanced OpenFLow switching/forwarding elements, and/or data transport circuits to connect these components. GTS allows the user to specify desired attributes of each resource - such as the geographic location (a city) for a server or switch, or the desired bandwidth capacity of a transport link.
GTS dynamically allocates and schedules testbeds at the user’s request. From idea to running testbed, a GTS network spanning the entire GEANT footprint can be instantiated in just a few minutes .
Network research must often culminate in studies “at scale” i.e. studying the behavior of a concept as it goes from a small lab experiment to a large globally deployed capability. The base GTS architecture is designed to scale globally – It can support the creation and management of networks containing thousands of network components, spread across a global geographic footprint, and comprised of resource drawn from many different service providers.
Reduced CapEx, OpEx, administrative overhead, and “ramp up” time
The design, engineering, contracting, purchasing, deployment, and operation of a “facilities based” network is an enormous effort. GTS eliminates almost all of these distractions to the actual research. Indeed, with GTS even a small research program can easily be up and running - performing relevant mission oriented research - almost immediately upon award rather than suffering through months of ramp up time and overhead unrelated to the program’s research objectives.
Security and Privacy
Because the GTS testbeds are, by design, insulated and isolated from one another, they are designed to be secure. User access to a project’s testbed(s) is carefully constrained to allow users to see and manipulate only their own testbeds, yet this security is designed to allow authorized users to access their testbed with a minimum of impedance.