In recent years, the rapid increase in the number of internet users and widespread usage of internet applications have obliged large servers and networking equipment to manage large data stack and optimize the instantaneous transmission of digital information. The COVID-19 Pandemic has also caused an increase in data exchanges and digital information generation. In order to manage large-scale data, there is a need for gigantic data centers (DCs) which are tremendous energy consumers and have relatively flexible loads that are easier to control by means of shifting in time and space. Therefore, DCs can be regarded as dispatchable loads and are considered good candidates for participating in demand side management (DSM) programs for power curve smoothing and compensation of power fluctuation in electrical power systems. In this paper, the question of why DCs should participate in DSM has been investigated rather than the technical methods used in DSM. The amount of DCs' participation energy is used by peak shaving/shifting method for power curve smoothing using actual data. The possible environmental and financial effects of it for Turkey and all the world have been carried out. The study results show that DCs' participation in DSM for Turkey decreases peak load by up to 2.18%, defers up to 34% of the installed power plants launched in 2019, and improves load and loss factors by up to 2.2% and 4.3% respectively. Additionally, global DC's participation in DSM decreases the peak point by up to 0.77% and reduces CO ₂ emission by 0.03%.

The appropriate feature/predictor selection is as significant as building efficient estimation methods for the accurate estimation of power consumption, which is required for self-awareness and autonomous decision systems. Traditional methodologies define predictors by assessing whether there is a relationship between the predictors and the response variable. Contrarily, this study determines predictors based on their individual and group impacts on the estimation accuracy directly. To analyze the impact of predictors on the power-consumption estimation of an IT rack in a data center, estimations were carried out employing each prospective predictor separately using the measured data under the real-world workload. Then, the ratio of CPU usage was set as the default predictor, and the remaining variables were assigned as the second predictor one by one. By utilizing the same approach, the best combination of predictors was determined. As a result, it was discovered that some variables with a low correlation coefficient with power consumption improved the estimation accuracy, whereas some variables with high correlation coefficients worsened the estimation result. The CPU is the most power-consuming component in the server and one of the most used predictors in the literature. However, the estimation accuracy obtained using only the CPU is 10 times worse than the estimation result conducted by utilizing the predictor set determined at the end of the experiments. This study shows that instead of choosing predictors only from one point of view or one method, it is more convenient to select predictors by assessing their influence on estimation results. Examining the trend and characteristics of the estimated variable should also be considered.

Data centres can support power systems by coordinating workload scheduling, electrical storage and cooling, buttheir available flexibility depends on the operating state andrequired service duration. This paper develops an integratedmodel of IT workload shifting, uninterruptible power supply(UPS) batteries, cooling and thermal energy storage. Linked dailymixed-integer linear optimisations establish a cost-minimisingoperating baseline over 2025. A duration-aware assessment thenidentifies certified feasible power deviations for specified starttimes while enforcing workload, storage, thermal and post-eventrecovery constraints. For a hypothetical data centre with 1MW of rated IT capacity, coordinated operation reduces annualelectricity costs by GBP 29,626.91 (5.38%). Resource sensitivitiesshow the largest change in savings when the flexible workloadshare is varied, followed by UPS and thermal storage energycapacity. On the selected representative day, a 100 kW loadreduction is feasible for six hours from noon and one hourfrom 06:00. Morning reductions rely predominantly on UPSdischarge, whereas afternoon reductions rely mainly on workloaddeferral and associated cooling reductions. Load increasescombine workload execution, cooling and storage actions. Theresults quantify feasible magnitude–duration combinations underthe stated model and recovery assumptions; assessment capsand unresolved trials do not establish maximum durations.Lower electricity costs do not imply lower annual electricityconsumption or peak import. The framework provides candidateflexibility offers for subsequent assessment against networkconstraints and service-specific requirements.