Enterprise level computing constantly investigates novel approaches that maximize their profits and minimize their expenses. With the rapid growth of cloud computing XaaS -'anything as a service', service providers are enabled with the rapid deployment of virtual services to service requestors. Because of the enormous growth in the variety of the services and based on the demand of the virtualized resources, cloud service providers are facing tough competition to facilitate the composite service requests made by the service requestors. QoS (Quality of Service) is considered to be a preliminary factor while composing a new cloud service out of heterogeneous and distributed atomic services. Therefore service composition is promising area that focuses on the design and development of the automated approaches to deal with diverse phases of service composition techniques that include service discovery, negotiation, service selection and optimization of the atomic services. This paper provides anatomy of existing studies addressing the problem of cloud service composition that enable to identify intended objectives of the technique along with diverse QoS aware problem solving approaches. Furthermore, the key areas of the improvement in cloud service composition are identified for future research.