Page 302 - Cloud computing: From paradigm to operation
P. 302

1                                    Framework and requirements for cloud computing


            3)      The CSC accesses the primary CSP's catalogue of service offerings and selects one or some of the
                    services in the catalogue;
            4)      The primary CSP arranges the service selected by the CSC. The service may be provided by the
                    primary CSP, by (some of) the secondary CSP(s) or by a combination of these CSPs. The primary CSP
                    intermediates, aggregates and/or arbitrages the services (see clause 7.3). The CSC starts using the
                    service.

            In practice, the actual process will be more complex. There are different ways for the CSC to use the services.
            If the requested service is hosted in the primary CSP, the CSC may access to this service directly (in the case
            of 4a)). If the requested service is offered by a secondary CSP, the CSC may access to the service in the
            secondary CSP via the primary CSP (in the cases of 4b) and 4c)). In the latter case, the network conditions,
            service characteristics and service agreement among the CSC, the primary CSP and the involved secondary
            CSP should be considered in order that the CSC accesses to the service in an appropriate manner.


            9       Functional requirements for inter-cloud
            This clause describes the CSPs' capabilities necessary to support different inter-cloud computing patterns
            described in clause 8.

            The capabilities and the relevant requirements identified in this clause are complementary to the general
            requirements applicable to a CSP involved in inter-cloud as specified in [ITU-T Y.3501].

            9.1     SLA and policy negotiation

            The SLA and policy negotiation capability deals with the matching made by a primary CSP between SLA
            requirements of a CSC and the SLAs of secondary CSPs involved in the considered inter-cloud pattern. The
            subject  includes  the  QoS  related  aspects.  This  capability  also  deals  with  the  negotiation  of  service
            provisioning policies associated with the different CSPs involved in the inter-cloud pattern.
            The SLA requirements (including QoS) of a CSC for a given cloud service are expected to be met by appropriate
            interworking with selected CSPs, even in the event of service performance degradation or a disaster.
            The SLA and policy negotiation capability is required to:

            –       be aware of the SLA information related to the QoS and performance aspects of the CSPs involved
                    in the inter-cloud using standard formats.

            The SLA and policy negotiation capability is recommended to:
            –       allow comparing, negotiating and settling down service provisioning policies between multiple CSPs
                    (for example, based on the settlement, these CSPs may be considered as a trusted group for inter-
                    cloud support).
            NOTE – In this clause, policy refers to a way for a CSP to provide services in terms of presumed reliability, including its
            backup scheme and target service levels. The policy affects SLAs. Policies may be different among CSPs. Policies may be
            negotiated beforehand and settled. This process is referred to as the policy negotiation.
            9.2     Resource monitoring

            The resource monitoring capability deals with the monitoring by the primary CSP of the resources of the
            secondary CSPs and the status attributes of these resources (e.g., usage amount, performance and quality
            aspects). The primary CSP collects and monitors the information from the secondary CSPs in a secure way.
            By monitoring the secondary CSPs' resources status (e.g., availability and dead/alive status of machines) and
            detecting service level performance degradation (in terms of delay and response time), the primary CSP can
            initiate actions to maintain service availability with the help of other secondary CSPs.

            The resource monitoring capability is recommended to:
            –       allow describing and expressing the resource information (e.g., resource type, configuration and
                    status) in a standard manner in order to be able to monitor these resources across multiple CSPs;




            294
   297   298   299   300   301   302   303   304   305   306   307