This thesis provides a novel architecture for the creation and management of virtual multimedia adaptation services offered by a multimedia-enabled cloud. The aim of the proposed scheme is to provide an optimal yet a transparent user access to adapted media contents while isolating them from the heterogeneity of the utilized devices, diversity of media formats, as well as the details of the adaptation services and performance variations of the underlying network. This goal is achieved through the development of service virtualization models that provide various levels of abstraction of the actual physical services and their performance parameters. Such virtual models offer adaptation functions by comprising adaptation services with accordance to their parameters. Additionally, parameters describing the functional specifics of the adaptation functions, as well as multimedia content features, are organized into a hierarchical structure that facilitates extraction of the virtual models capable of satisfying the conditions expressed by the user requests. At the same time the paramter/feature organization structure itself is flexible enough to allow users to specify media delivery requests at various levels of request details (e.g., summarize video vs. drop specific frames). As a result, in response to a user request for a multimedia content, an optimal virtual service adaptation path is calculated, describing the needed media adaptation operations as well as the appropriate mapping to the physical resources capable of executing such functions. The selection of the adaptation path is done with the use of a novel performance-history based selection mechanism that takes into account the performance variations and relations of the services in a dynamically changing environment of multimedia clouds. A number of experiments are conducted to demonstrate the potential of the proposed work in terms of the enhanced processing time and service quality.

This thesis provides a novel architecture for the creation and management of virtual multimedia adaptation services offered by a multimedia-enabled cloud. The aim of the proposed scheme is to provide an optimal yet a transparent user access to adapted media contents while isolating them from the heterogeneity of the utilized devices, diversity of media formats, as well as the details of the adaptation services and performance variations of the underlying network. This goal is achieved through the development of service virtualization models that provide various levels of abstraction of the actual physical services and their performance parameters. Such virtual models offer adaptation functions by comprising adaptation services with accordance to their parameters. Additionally, parameters describing the functional specifics of the adaptation functions, as well as multimedia content features, are organized into a hierarchical structure that facilitates extraction of the virtual models capable of satisfying the conditions expressed by the user requests. At the same time the paramter/feature organization structure itself is flexible enough to allow users to specify media delivery requests at various levels of request details (e.g., summarize video vs. drop specific frames). As a result, in response to a user request for a multimedia content, an optimal virtual service adaptation path is calculated, describing the needed media adaptation operations as well as the appropriate mapping to the physical resources capable of executing such functions. The selection of the adaptation path is done with the use of a novel performance-history based selection mechanism that takes into account the performance variations and relations of the services in a dynamically changing environment of multimedia clouds. A number of experiments are conducted to demonstrate the potential of the proposed work in terms of the enhanced processing time and service quality.

This thesis provides a novel architecture for the creation and management of virtual multimedia adaptation services offered by a multimedia-enabled cloud. The aim of the proposed scheme is to provide an optimal yet a transparent user access to adapted media contents while isolating them from the heterogeneity of the utilized devices, diversity of media formats, as well as the details of the adaptation services and performance variations of the underlying network. This goal is achieved through the development of service virtualization models that provide various levels of abstraction of the actual physical services and their performance parameters. Such virtual models offer adaptation functions by comprising adaptation services with accordance to their parameters. Additionally, parameters describing the functional specifics of the adaptation functions, as well as multimedia content features, are organized into a hierarchical structure that facilitates extraction of the virtual models capable of satisfying the conditions expressed by the user requests. At the same time the paramter/feature organization structure itself is flexible enough to allow users to specify media delivery requests at various levels of request details (e.g., summarize video vs. drop specific frames). As a result, in response to a user request for a multimedia content, an optimal virtual service adaptation path is calculated, describing the needed media adaptation operations as well as the appropriate mapping to the physical resources capable of executing such functions. The selection of the adaptation path is done with the use of a novel performance-history based selection mechanism that takes into account the performance variations and relations of the services in a dynamically changing environment of multimedia clouds. A number of experiments are conducted to demonstrate the potential of the proposed work in terms of the enhanced processing time and service quality.