Data warehouses are used for analytical processing and effective business report generation for knowledge workers throughout the enterprise. It is a central repository of integrated data from one or more different sources. Now a day's data warehouse considered as one of the core component of business intelligence. The performance of traditional data warehouse become degraded when the report generation needs on a fly or the report generation needed only for a specific set of data in some emergency situations. Traditional data warehouse and data marts are based on the time consuming ETL process, hence on-the-fly decision makings are not possible. Virtual Data Warehouse (VDW) is the concept where the data are not required to be cleaned and transformed through ETL. Data in the VDW are schema-less, so the time requirement to form ETL is reduced drastically. Innovation of Service Oriented Computing Paradigm helps to provide the exact functionality to the users on the pay or need basis. The functionalities of VDW could be provided in the form of Software as a Service (SaaS) in cloud computing environment using service modeling. Here we have proposed two different architectures to deploy VDW in Service Oriented Architecture (SOA) with the provision of offering cuboids as a service.