Cloud datacenter has become key infrastructure to satisfy the growing require of computing services. However, the high-power consumption in large-scale cloud operations brings operational cost and environmental sustainability challenges. Existing resource allocation schemes lack awareness of energy inefficiency caused by dynamic workloads and co-located resources in heterogeneous and multi-tier cloud systems and, as a result, under-utilize resources and increase energy waste. This paper solve the multi-tier cloud architecture problem of wasting energy by the consideration of an Energy-Aware Multi-Tier Resource Allocation Framework. The framework is developed to dynamically distribute computing resources over a set of tires, according to the workload properties, a resource availability and costs. It incorporates a QoS-aware scheduler that meets SLAs and at the same time minimizes the energy consumption. The underlying proposed solution to the problem is based on the modeling and simulation and justifies effectiveness through the use of mathematical models identified by industry standard simulation tools such as CloudSim, Openstack. It includes tier-wise power metrics to promote resource provision, wakes up as few servers from a lower tier as possible and adapts pro-actively to potential changes in demand. .