Modelling and Control of Power Dissipation in Internet Data Centers

Abstract

Large-scale data centers involve a set of server racks for storage and computations for which they require huge amount of power along with some cooling arrangement. It is observed in the literature that the size of internet data center has increased 10 times in the last 10 years and energy cost is going up in a similar way. So there is a need of proper power management in internet data centres so that the power consumption can be reduced. In this thesis, our focus is modelling and simulation of internet data centers and comparing two different control techniques at different workloads on the servers. The first technique is a CRACs ON-OFF method where the power of computer room air conditioning (CRACs) is automatically controlled on the basis of servers output temperature. In particular, if the temperature of a server rack in internet data centre is greater than some fixed temperature, the CRACs are turned on, otherwise the CRACs will stay off. The second method is the multi-steps ON/OFF control in which the CRACs are partially turned on and off on the basis of outer air temperature of servers. We vary the intermediate steps to 1 and 3 in the multi-step ON/OFF control. It is observed that the two different control techniques can ensure a desired output temperature of server racks, but CRACs ON-OFF control method involves more and sharp power peaks which can cause problems in the operation of IDCs. The multi-step CRAC’s ON/OFF control involve small power peaks and therefore can be considered a better option as compared to CRACs ON-OFF method. The behaviour of the control system is almost similar under different workload conditions. The proposed multi-step CRAC’s ON/OFF control is an important contribution of this thesis which can minimize the total power consumption of the data center. An important future work will be to consider state estimation in the modelling and control strategy under different workloads.