Costs to be considered for a cloud migration include IT infrastructure (servers, switches, routers, firewalls, etc.), data center equipment (HVAC, racks, UPS, emergency power), real estate, software licenses, systems engineering and O&M staff. Cost models not only need to take into account resource consumption in the cloud, but also the variation in the workload (e.g. does the application get busier during certain times of the year, month, or day?).
- Decision Support Tools for Cloud Migration in the Enterprise White Paper. Cost modeling tool which uses a Unified Modelling Language (UML) diagram to build a cloud deployment profile that is built as part of the Eclipse IDE. it extends UML deployment diagrams to include: Virtual Machine, Virtual Storage, Application, Data, Database, Remote Node, Communication Path, and Deployment. Once the model is created, it can be applied to a cloud provider to determine the computational resources needed. Baseline usage is set for each resource and then an elasticity pattern defines variations in use by day, month, or year. The usage patterns for each node are multiplied by the per unit cost of that resource over the simulation period. Per unit price is determined by the cloud provider specified by the user using an WML file that stores cloud provider prices from over 600 prices from AWS, Azure, FlexiScale, Rackspace, GoGrid, and ReliaCloud.
No comments:
Post a Comment