| Abstract [eng] |
The optimization of cloud computing provision in a Business to Business (B2B) model is highly desirable in today’s dynamic global business environment. The business workflows between business partners, suppliers and consumers need to be highly adaptable to changing business needs. As businesses move from a static to a dynamic environment, cloud computing provides a highly scalable, elastic and cost effective compute platform to run both, traditional platform 2 applications as well as the platform 3 cloud native applications. The cloud provides a common global platform for all B2B stake holders to run their development, testing as well as the production environment with pay-as-you-go cost model. The environment can scale to simulate 1000 concurrent user testing of the application one day to running a steady-state 100 machines environment the next day. The production environments can be expanded to multiple datacenters globally to provide customer or partner access close to their geographies, yet have a built-in disaster recovery capability that allows continued operation without any interruptions. Though the cloud as a platform may appear to be homogeneous, it is anything but homogeneous. The deployments in public, private, community or in a hybrid environment present number of challenges. These challenges include disparate hypervisors for running the virtual machines to different cloud management platforms to different parameters for consumption and billing. The various cloud stacks available commercially and by each cloud service provider (CSP) can create vendor lock-ins as well as information silos. This is not very desirable for an optimized B2B cloud compute platform. This thesis outlines various challenges and how those challenges can be addressed to achieve the desired state of an optimized provisioning of B2B cloud computing platform. The thesis also provides an experimental analysis of the effective approaches using commercial tools across the various cloud types. We finally conclude with a reference architecture that addresses the requirements and challenges in a global heterogeneous environment of cloud technologies and deployments. The adoption of this reference architecture will allow businesses to achieve an optimized provisioning and consumption of computing in a B2B model. |
