Page 538 - Cloud computing: From paradigm to operation
P. 538

2                                            Cloud Computing management


            5       Conventions

            In this Recommendation:
            The keywords ''is required to'' indicate a requirement which must be strictly followed and from which no
            deviation is permitted if conformance to this document is to be claimed.

            The keywords ''is recommended'' indicate a requirement which is recommended but which is not absolutely
            required. Thus this requirement need not be present to claim conformance.


            6       Overview of inter-cloud data management
            Nowadays, one of the main challenges for cloud service providers (CSPs) is to ensure credibility of data storage
            and  transport  in  multi-cloud  environments.  Inter-cloud  data  management  functions  need  to  consider  and
            reflect security and governance aspects of data handling uniformity and interoperability across different CSPs.
            Inter-cloud data categorization in aspects of identification qualifiers and dependency, as well as inter-cloud
            data annotation, processing and usage, is necessary to attain the data treatment required among multiple
            CSPs. The appropriate security and access control mechanisms determine access control to inter-cloud data
            under particular conditions (e.g., temporarily, a certain number of times or related to a particular community
            and context).
            With the development and wide usage of big data-related technologies, more and more data and datasets
            are stored in geographically distributed and heterogeneous computing environments, also known as inter-
            cloud  environments  (e.g.,  in  a  high-performance  computing  scenario).  Migrating  an  application  from  a
            centralized or static place to the distributed computing environments may bring the following challenges:
            1)      transmission of intensive data among different CSPs may cause low data access efficiency;
            2)      heterogeneous dataset access and sharing may bring extra computation consumption;
            3)      the  necessary  and  unavoidable  maintenance  of  extensive  tracking  of  metadata,  data  locations,
                    access control policies, etc.
                    NOTE – Intensive data transmission and heterogeneous dataset access and sharing-related technologies lie
                    outside the scope of this Recommendation.
            [ITU-T Y.3601] specifies the big data exchange framework and requirements based on the big data ecosystem
            and capabilities specified in [ITU-T Y.3600]. In [ITU-T Y.3601], data exchange is described as a process of
            receiving source data under a source schema from data source, transforming it into target data under a target
            schema without altering the representation of source data, and delivering the target data to the data target.
            This Recommendation focuses on the inter-cloud data management aspects and does not aim to specify any
            concrete  data  exchange  details.  For  more  information  about  big  data  exchange,  please  refer  to
            [ITU-T Y.3601].


            6.1     Inter-cloud data categorization
            [b-ISO/IEC 19944] identifies the categories of data that flow across cloud service customer (CSC) devices and
            cloud services, and can be captured, processed, used and shared. It extends the definitions of CSC data, cloud
            service-derived data, CSP data and account data. This Recommendation follows the taxonomy of data in
            [b-ISO/IEC  19944]  and  focuses  on  the  inter-cloud  data  aspect.  Similarly  to  [b-ISO/IEC  19944],  this
            Recommendation is not intended to be exhaustive, but is intended to be extensible. It is recommended that
            the  hierarchical  relationship  based  on  the  four  topmost  categories  defined  in  [b-ISO/IEC  19944]  be
            maintained.

            6.2     Data policy language
            The data policy language (DPL) allows CSPs to annotate, manage, process and use CSC or CSP data in an inter-
            cloud environment (who can do what and when) according to data policies in force. The main challenges for
            DPL are related to small storage footprint, human readability, detection of conflicts at time of specification,
            performance and adaptation to new policy statements that come up with emerging inter-cloud services.
            The DPL expresses inter-cloud data policies for different cloud service categories [e.g., NaaS, SaaS, platform
            as a service (PaaS)] provided by the CSPs in an inter-cloud environment.


            530
   533   534   535   536   537   538   539   540   541   542   543