Small Data Center Research Articles

Retrofitting HVAC systems for enhanced energy performance and resilience in university data center room

Data centers have been extensively implemented globally to satisfy the increasing need for IT services. Small data centers with insufficient cooling require much power to keep IT equipment running reliably. As a result, retrofitting has become an efficient method to increase energy efficiency and save operating costs. This study examines the impact of retrofitting on energy efficiency in a data center room at a university. The data center under consideration underwent a comprehensive retrofitting process. Implementing a closed in-row type data center increases cooling capacity, resulting in more efficient cooling and improved temperature control. Measurement data were collected before and after retrofitting, including power usage effectiveness (PUE), energy consumption, temperature, and velocity distribution. Computational fluid dynamics (CFD) simulations were also conducted to determine the characteristics of the data center's airflow distribution and temperature profile. Numerical simulations carried out for two operational situations have indicated that a slight modification has the potential to greatly enhance the efficiency of the cooling procedure. Numerical simulations have revealed that developing hotspots in the racks has resulted in air temperatures exceeding the allowable maximum temperature, which can be improved after retrofitting. The PUE of the data center before/after retrofitting was 2.17 and 1.52, respectively. Free cooling technology is also implemented, making the data center more efficient and reducing PUE to 1.32. The results showed a significant improvement in energy efficiency, with a PUE reduction of 0.85, which reveals that energy efficiency has been improved by 48.71 %. The retrofitting results in significant cost savings and environmental benefits.

Development and experimental study of a compact silica gel-water adsorption chiller for waste heat driven cooling in data centers

In recent years, data centers have experienced rapid construction, leading to a significant surge in energy consumption, and cooling systems have emerged as prominent energy consumers of data centers. Among the cooling technologies, liquid cooling allows higher cooling circuit temperatures and provides potential for heat recovery. As a thermally driven cooling technology, adsorption refrigeration systems offer an alternative way to compression refrigeration systems by utilizing low-grade waste heat from liquid-cooled circuits to produce chilled water for air-cooled components. In this study, a compact silica gel-water adsorption chiller is developed for a small modular data center and experimentally investigated under extremely low driving temperatures. Under the typical summer conditions with the hot/cooling/chilled water inlet temperatures of 50 °C/30 °C/23 °C, the adsorption chiller achieves the cooling power of 4.9 kW and the COP of 0.50. When the cooling water inlet temperature changes from 32 °C to 24 °C, both the cooling power and COP show significant improvements, from 3.04 kW to 10.50 kW and from 0.44 to 0.66, respectively. Moreover, the increase of the hot water inlet temperature from 48 °C to 56 °C can enhance the cooling power to some extent, while the chilled water inlet temperature has little influence on the performance. The experimental results demonstrate the feasibility and performance of the developed adsorption chiller in data center waste heat driven cooling. Besides, potential adsorbents and possible optimization strategies are presented for performance improvement based on adsorption potential method.

Joint Optimization of Memory Sharing and Communication Distance for Virtual Machine Instantiation in Cloudlet Networks

Cloudlet networks are an emerging distributed data processing paradigm, which contain multiple cloudlets deployed beside base stations to serve local user devices (UDs). Each cloudlet is a small data center with limited memory, in which multiple virtual machines (VMs) can be instantiated. Each VM runs a UD’s application components and provides dedicated services for that UD. The number of VMs that serve UDs with low latency is limited by a lack of sufficient memory of cloudlets. Memory deduplication technology is expected to solve this problem by sharing memory pages between VMs. However, maximizing page sharing means that more VMs that can share the same memory pages should be instantiated on the same cloudlet, which prevents the communication distance between UDs and their VMs from minimizing, as each VM cannot be instantiated in the cloudlet with the shortest communication distance from its UD. In this paper, we study the problem of VM instantiation with the joint optimization of memory sharing and communication distance in cloudlet networks. First, we formulate this problem as a bi-objective optimization model. Then, we propose an iterative heuristic algorithm based on the ε-constraint method, which decomposes original problems into several single-objective optimization subproblems and iteratively obtains the subproblems’ optimal solutions. Finally, the proposed algorithm is evaluated through a large number of experiments on the Google cluster workload tracking dataset and the Shanghai Telecom base station dataset. Experimental results show that the proposed algorithm outperforms other benchmark algorithms. Overall, the memory sharing between VMs increased by 3.6%, the average communication distance between VMs and UDs was reduced by 22.7%, and the running time decreased by approximately 29.7% compared to the weighted sum method.

Applying Toroidal k-ary Grids for Optimizing Edge Data Centers

IoT deployments are growing exponentially, leading to a huge increase in edge computing facilities. In order to cope with such a demand, data centers need to get customized for the specific requirements of edge computing, such as a small number of physical servers and the ability to scale and unscale according to the traffic flows running at a given time. In this context, artificial intelligence plays a key part as it may anticipate when traffic throughput will increase or otherwise by scrutinizing current traffic whilst considering other factors like historical data and network baselines. In this paper, a dynamic framework is outlined based on toroidal k-ary grids so as to organize and optimize small data centers, allowing them to increase or decrease according to the current and predicted capacity of IoT-generated traffic flows.

AVS_FD_MVITS: an agile IT service design workflow for small data centers

AVS_FD_MVITS: an agile IT service design workflow for small data centers

Experimental and numerical investigation on thermal performance of data center via fan-wall free cooling technology

With the rapid increase in the scale of data center (DC) construction, DC air conditioning energy consumption has attracted wide attention over the past years. As the DCs are located in the mild climate regions, the air-side free cooling technology that combined with the fan wall air supply demonstrates excellent energy-saving potential. However, there are quite few reports on the environmental thermal response of air conditioning in fan-wall DCs. In this article, the effect of both supply air volume (SAV) and supply air temperature (SAT) on the temperature control performance of the DC are tested through a newly built small fan-wall DC facility. The effect of supply air parameters and the DC’s total power (DCTP) are further investigated through numerical simulation. The results show that the server inlet air temperature increases linearly with the SAT and the DCTP, whereas decreases logarithmically with the SAV. The optimal SAV is 0.5 m3/s for a single rack with 2.5 kW when the SAT is 27 °C. Based on the Rack Cooling Index (RCI) evaluation method, an empirical SAV correlation is established and verified against other numerical cases for fan-wall DCs when taking the SAV, the SAT, and the DCTP into account. In the current study, according to the developed model, approximately 34% energy consumption of fans could be saved for fan-wall DCs.

Experimental study and information entropy analysis on periodic performance of a PCM thermal management system for blade servers in data centers

The high power of blade servers inside small data centers (DCs) can cause heat accumulation, which can degrade the performance of DCs. To address the problem of heat dissipation, we performed periodic experiments on a thermal management system based on a phase change material. The effects of time ratio, heating power, and cooling water temperature on the performance of thermal control and periodic stability were investigated. Information entropy reflects uncertainty by transforming a variable changes over time into a single parameter representing inhomogeneity of time. Herein, it was introduced as an evaluation index of thermal control performance. The results verified that the peak temperature of the heating surface can be maintained below 80 °C for several periods under specific conditions. Additionally, the information entropy results of the experimental data reflected the thermal control performance of the system and validated the experimental conditions. Based on the experimental results and the corresponding information entropy analysis, we proposed a design strategy for the system parameters and a method for reducing the number of measuring points. The obtained results formed the basis for the thermal management design of the server level of DCs, other electronic equipment, and electric vehicles.

Edge Data Center Organization and Optimization by Using Cage Graphs

Data center organization and optimization are increasingly receiving attention due to the ever-growing deployments of edge and fog computing facilities. The main aim is to achieve a topology that processes the traffic flows as fast as possible and that does not only depend on AI-based computing resources, but also on the network interconnection among physical hosts. In this paper, graph theory is introduced, due to its features related to network connectivity and stability, which leads to more resilient and sustainable deployments, where cage graphs may have an advantage over the rest. In this context, the Petersen graph cage is studied as a convenient candidate for small data centers due to its small number of nodes and small network diameter, thus providing an interesting solution for edge and fog data centers.

Comparing User Space and In-Kernel Packet Processing for Edge Data Centers

Telecommunication operators are massively moving their network functions in small data centers at the edge of the network, which are becoming increasingly common. However, the high performance provided by commonly used technologies for data plane processing such as DPDK, based on kernel-bypass primitives, comes at the cost of rigid resource partitioning. This is unsuitable for edge data centers, in which efficiency demands both general-purpose applications and data-plane telco workloads to be executed on the same (shared) physical machines. In this respect, eBPF/XDP looks a more appealing solution, thanks to its capability to process packets in the kernel, achieving a higher level of integration with non-data plane applications albeit with lower performance than DPDK. In this paper we leverage the recent introduction of AF_XDP, an XDP-based technology that allows to efficiently steer packets in user space, to provide a thorough comparison of user space vs in-kernel packet processing in typical scenarios of a data center at the edge of the network. Our results provide useful insights on how to select and combine these technologies in order to improve overall throughput and optimize resource usage.

Coordinated Planning of Power Grids in Information Parks considering the Time-Space Transfer Characteristics of Data Center Loads

With the development of green energy, the penetration rate of new energy in the power grid is rising rapidly. Incorporating a large amount of new energy into the power grid will affect the power supply balance of the power grid, and even in severe cases will affect the safe and stable operation of the power grid. The information park has various small and medium data centers. The adjustable load of the data center can be transferred in time and space through the optical fiber and other communication networks, which can enhance the controllability of the power grid to a certain extent and promote the consumption of new energy. This paper proposes a collaborative planning method for information parks that considers the temporal and spatial transfer characteristics of data center loads, builds a load transfer model for data centers and builds a park power grid expansion model with the goal of minimizing the total cost of construction and operation. In this paper, an example of a park is used to verify the calculation example, and it is compared with other schemes to verify the feasibility of the method in this paper.

Performance analysis of an energy-saving strategy in cloud data centers based on a MMAP[K]/M[K]/[formula omitted] non-preemptive priority queue

With the development of cloud computing, big data, artificial intelligence and other next generation information technologies, the scale of the data center industry worldwide is growing rapidly, resulting in a sharp increase in energy consumption. In order to efficiently reduce the idle energy consumption in cloud data centers, in this paper, we propose an energy-saving strategy based on two-threshold hysteresis cluster scheduling mechanism, which enables the reserved cluster to be dynamically switched on and off with load variation. In addition, the tasks to be processed are classified into real-time tasks and non-real-time tasks. To keep track of the two classes of tasks and model the correlated traffic in cloud data centers, we describe the arrival flow of tasks as a Marked Markovian Arrival Process (MMAP). Accordingly, we develop a non-preemptive priority queue as the system model to capture the working principle of the proposed strategy. By using the matrix-geometric solution and Gauss–Seidel method, the steady-state distribution of the system model is analyzed, and some key Quality of Service (QoS) metrics and Total Cost of Ownership (TCO) metrics are calculated. Results of numerical experiments show that, under the proposed energy-saving strategy, the overall power consumption in a small cloud data center can be reduced by an average of 30.20% under different scenarios, and the larger the cloud data center scale, the more obvious the energy saving effect. The results also confirm the impact of inter-class correlations of the input process on performance metrics. Furthermore, we identify the Pareto optimal solutions for trading off the overall power consumption and average waiting time of real-time tasks.

Experimental analysis of airflow uniformity and energy consumption in data centers

Temperature and airflow uniformity are experimentally investigated in this research study along with energy consumption analysis on the laboratory container size small data center. A raised floor plenum (RFP) data center system of size 6 (L) × 3.1 (W) × 2.3 (H) m with one Computer Room Air Handler (CRAH), two rows and each row contains five racks, five 8U simulator cabinets in each rack was constructed to test partial and closed cold aisle containment (CAC). Seven different cases were investigated with variables of perforated tiles (32%, 50 % & 70%), CRAH fan speed (60 Hz, 42 Hz), chiller return water temperature (CRWT) (15 °C & 17 °C) and heating load (full 50 kW, 75% load 37.5 kW). These cases were further analyzed using Rack Cooling Index (RCI) and Supply Heat Index (SHI) along with power consumption scenarios in terms of Coefficient of Performance (COP) and Power Usage Effectiveness (PUE). The uniqueness of this study is experimental investigations of system COP and PUE analysis for a wide range of important variables for partial and closed CAC. It was found that 50% porous tile provided the best airflow and temperature uniformity for the case of partial containment with full and 25% reduced heat load along with full and 30% reduced CRAH speed respectively. Slight recirculation was observed during the reduced CRAH fan speed case at the corner edge of data racks. However, due to additional supply airflow usage, these recirculation’s were within the acceptable SHI. Reduction of CRWT reduces the total Heating Ventilation and Air Conditioning (HVAC) power consumption and improves the overall system COP.

Disaggregation and cloudification of metropolitan area networks: enabling technologies and impact on architecture, cost, and power consumption [Invited

The deployment of the new 5G system and ultra-broadband access networks, on one side, and the advent of network disaggregation and cloudification paradigms, on the other, are producing disruptive changes in metropolitan area networks (MANs). The new central offices will be more like small data centers than traditional Telco-type central offices. The presence of general-purpose servers, in partial or total replacement of specialized equipment or cards, and also hosting new functions and services, significantly changes the way the central office is organized and how the equipment internetworks. After having defined a reference architecture named cloudified-MAN, the paper briefly introduces the enabling technologies that include information technology (IT) systems, packet switches, and optical equipment. A disaggregated model is applied to all layers and systems. In the optics field, special insight on high data rate (up to 1.6 Tb/s) transceiver technology developed by the TERIPHIC project is provided as an example of low-cost and low-power technology. Through a quantitative techno-economic case study on a realistic network scenario aimed at short-term and medium-term time horizons, the paper provides indications on what could be the impact of the new architecture on costs and energy consumption. The main outcome of the study is that the IT and optical parts are those with the highest fractions of overall cost, and specifically for the optics, the cost of wavelength division multiplexing transceivers increases significantly in the medium term due to the demand for high-capacity connections ( > 400 G b / s ), while the line system cost (reconfigurable optical add-drop multiplexer) decreases in percentage. Another significant result is that in terms of energy consumption, the IT component has been found to be the dominant one (70%–75% of the total), followed by packet devices (about 20%–25%) and optical systems, the latter with significantly lower percentage values (less than 5%) in all cases analyzed.

Multi-objective prediction-based optimization of power consumption for cloud data centers

The overall development of the cloud paradigm is dominating omnipresence in the industry 4.0 business world. Over the last decade, the control measures for power utilization among the proliferative Hyper-Scale Data Centers (HSDCs) have been elucidated. However, the lack of attention to regulating power in Small and Medium-Scale Data Centers (SMSDCs) has ensued in excessive power drainage in small and medium-scale cloud data centers. The crucial factor for excessive power utilization of SMSDCs encompasses providing excessive resources, high certainty tasks. Majority of the previously reported studies zeroed-in on problems associated with hyper-scale data centers, excluding probes of the issues prevalent in small and medium-scale cloud data centers. This paper proffers a framework for a predictive optimization approach for delivering the data center services to end-users. In the first phase, the Multi-Output (MO) Random Forest Regressor (RFR) (MO-RFR) concurrently predicts the multiple-resource utilization of Virtual Machines (VMs). The predictive framework outcome was utilized by the Multi-Objective Particle Swarm Optimization (MO-PSO) framework in the second phase to resolve the issue in virtual machine placement and to accomplish better physical machine consolidation. The proposed multi-prediction-based MO-PSO to escalate the resource usage, minimizes the power utilization, and curtail the carbon footprint. The efficacy of the proposed approach was appraised via performance metrics and actual workload traces. The acquired result from the proposed method outperforms the baseline approaches.

Efficient Dynamic Deployment of Simulation Tasks in Collaborative Cloud and Edge Environments

Cloud computing has been studied and used extensively in many scenarios for its nearly unlimited resources and X as a service model. To reduce the latency for accessing the remote cloud data centers, small data centers or cloudlets are deployed near end-users, which is also called edge computing. In this paper, we mainly focus on the efficient scheduling of distributed simulation tasks in collaborative cloud and edge environments. Since simulation tasks are usually tightly coupled with each other by sending many messages and the status of tasks and hosts may also change frequently, it is essentially a dynamic bin-packing problem. Unfortunately, popular methods, such as meta-heuristics, and accurate algorithms are time-consuming and cannot deal with the dynamic changes of tasks and hosts efficiently. In this paper, we present Pool, an incremental flow-based scheduler, to minimize the overall communication cost of all tasks in a reasonable time span with the consideration of migration cost of task. After formulating such a scheduling problem as a min-cost max-flow (MCMF) problem, incremental MCMF algorithms are adopted to accelerate the procedure of calculating an optimal flow and heuristic scheduling algorithm, with the awareness of task migration cost, designed to assign tasks. Simulation experiments on Alibaba cluster trace show that Pool can schedule all of the tasks efficiently and is almost 5.8 times faster than the baseline method when few tasks and hosts change in the small problem scale.

Feasibility Investigation on a Novel Rack-Level Cooling System for Energy-Saving Retrofit of Medium and Small Data Centers

Feasibility Investigation on a Novel Rack-Level Cooling System for Energy-Saving Retrofit of Medium and Small Data Centers

Research on traffic scheduling based on private cloud platform

With the development of cloud computing applications in recent years, the problem of IDC traffic scheduling in small and medium-sized data centers has become increasingly prominent. This paper starts from the consideration of data traffic scheduling of small and medium-sized private cloud centers, and uses the small and medium-sized private cloud built by 40 HUAWEI servers as the experimental environment. It tries to consider the classification of different service applications, the requirements and requirements of different types of traffic, and build a hybrid traffic scheduling strategy to provide a reference for traffic scheduling research of small and medium-sized private cloud.

Research on air-water coordinated operation of water side free cooling system based on improved particle swarm optimization algorithm

Water side free cooling is an important technology for cooling water system to use natural cooling source for energy saving, especially in the field of data center and industrial production. However, the operation regulation of water side free cooling system was lack of scientific basis. This process involves the two-phase circulation of air and cooling water, which requires the coordinated operation of air and water to further optimize the energy consumption. The coordinated operation of air and water includes the coordination of air and water flow in each cooling tower and the matching of the operating numbers of equipment in the system. Firstly, based on the characteristics of the cooling tower, the key equipment in the water side free cooling system, the air-water two-phase heat transfer model in the cooling tower was established. Then, according to the heat transfer model and power model of each equipment in the system, the optimization objectives and constraints are established. An improved particle swarm optimization (PSO) algorithm was proposed to solve the problem of wind water collaborative optimization of water side free cooling system. Finally, this paper takes a small data center cooling water system as an example to analyze the optimization performance of the improved particle swarm optimization algorithm. The simulation results showed that the excellent rate obtained by the improved algorithm can be increased by 36.9%, which is more energy-saving. This method can provide a reference for the energy-saving optimization of water side free cooling system.

Proactive content caching in edge computing environment: A review

AbstractEdge computing environment provides processing capability and computing at the network edge and close to users. Edge equipment includes small data centers that locally perform process and content delivery. Therefore, edge equipment management has received much attention due to the rapid growth of information and resources limitations. The content caching and proactive caching techniques are management methods of edge equipment resources. We recently witnessed the development of proactive caching mechanisms that have a crucial enabler in improving heavy traffic, energy, and bandwidth. Also, it has huge potential to increase the quick response time to users' requests that today, these services are demanded by many users and applications. This article prepares a systematic literature review content caching approach in the edge computing environment. The purpose of this study is to survey the research done on the proactive caching strategies in the edge computing environment to identify subjects that must be emphasized more in current and future research paths. This research has studied 71 articles divided into three classes: model‐based, machine‐learning‐based, and heuristic‐based. Next, we discuss content caching approaches based on critical factors such as performance metrics, case studies, utilized techniques, assessment tools, advantages, and disadvantages. Finally, open issues and challenges are presented, and the survey is concluded.

Research on Energy Saving of Air Conditioning System in Data Center

With the development of the Internet today, there are more and more small and medium data center computer rooms, the energy consumption of data centers has also become the focus of attention. Air conditioning system is a big energy consumer in data centers, approximately 40% of the total energy consumption of the data center. So reducing the energy consumption of air conditioners has become the top priority of reducing the energy consumption of data centers. This paper studies the air distribution generated by different air-conditioning methods, building a computer room model and uses the simulation software 6sigma to simulate the different air conditioning modes, thereby reducing the energy consumption of the computer room.