To design, build and operate a multi-site cloud platform to enable the execution of large-scale experiments in IaaS modelTo become the entry point to FIRE for research mainly coming from the Services research community, allowing them to experiment considering also networking aspects (enabling novel and innovative scenarios)To stimulate the use of the facility running test experiments selected in 2 open callsTo analyse the business models for sustainability of the facility in order to ensure its continuity beyond the project lifetimeTo complement and expand existing EU Experimental Facilities collaborating on the construction of FIRE facility by means of federating BonFIRE with other facility projects---Scenario 1 - Extended CloudFederation of cloud sites using normal InternetScenario 2 - Cloud with emulated and controlled networkNetwork emulation environment for cross-cutting testsScenario 3 - Federation with other external facilities Potential use of Federica network slices (network interactions / experimental internet)Connection to Open Cirrus or other cloud infrastructuresUse of Panlab facility (tools and test descriptors)
Transcript of "Michael Boniface: Building Service Testbeds on FIRE "
BonFIRE Project Overview<br />Virtual Cloud FederatorVM managers and repositoriesService Deployment DescriptorsService Life-cycle ManagerService SLAsQoS Monitoring, etc<br />Experimentally driven research for Internet of Services selected through open calls<br />Heterogeneous cloud resources<br />Advanced control and monitoring<br />Experiment lifecycle management<br />Large scale experimentation<br />Service and networkconvergence<br />Permanent and on-request infrastructure facilities<br />
Need for common activities in service testbeds?<br />Testbedshave roles throughout the service lifecycle<br />Tests are broad with the trend towards *aaS<br />Generally three classes of testbedsfunded (commercial public, private, public/private)<br />Need to promote connected as well as common activities<br />
General challenges for sustainable testbeds<br />Users/customers need control<br />what services they consume, how much is used, and by whom<br />Service providers need independence<br />maintain control of their own resources<br />Accountability must be addressed at many levels (QoS, SLAs, usage and cost)<br />best effort not really good enough for anyone<br />centralised brokers have inherent difficulties<br />Security must to be commercial standards<br />especially for near to market experiments<br />Maintenance must be cost effective<br />academia<br />industry<br />
A particular slide catching your eye?
Clipping is a handy way to collect important slides you want to go back to later.