Common use of Service Registry Clause in Contracts

Service Registry. a database for raw measures The Service Registry contains all the information about the network performance. The raw measures produced by the SLA Observer can’t be stored in the WSN nodes, that are memory-constrained (they can only keep the current local information). Moreover, the raw measures must be consulted by the other enti- ties of the architecture (SLA Manager, SLA Admit- ter), without occupying the WSN resource. Hence, the Service Registry must be centralized (e.g. a database in the WSN operator core network). We keep the raw measures all along the SLA du- ration. Thus, if the WSN operator changes the algo- rithms of the SLA Manager and SLA Admitter, new policies can be applied taking into account the whole history of the WSN. Given that we monitor various sensor nodes, dur- ing long periods, the measures have to be accurately identified and classified. We propose the following log format for the raw measures: • timestamp: the time of the measure; • measurer (id): the producer of the measure; • entity (ids): which entities have been measured;

Service Registry. a database for raw measures The Service Registry contains all the information about the network performance. The raw measures produced by the SLA Observer can’t be stored in the WSN nodes, that are memory-memory- constrained (they can only keep the current local information). Moreover, the raw measures must be consulted by the other enti- ties entities of the architecture (SLA Manager, SLA Admit- terAdmitter), without occupying the WSN resource. Hence, the Service Registry must be centralized (e.g. a database in the WSN operator core network). We keep the raw measures all along the SLA du- rationduration. Thus, if the WSN operator changes the algo- rithms algorithms of the SLA Manager and SLA Admitter, new policies can be applied taking into account the whole history of the WSN. Given that we monitor various sensor nodes, dur- ing during long periods, the measures have to be accurately identified and classified. We propose the following log format for the raw measures: • timestamp: the time of the measure; • measurer (id): the producer of the measure; • entity (ids): which entities have been measured;