Energy and Thermal Aware Scheduling in Data Centers

The NLR portion of the "Cloud & Infrastructure CoP - Data Center Energy and Efficiency with AI Adoption" web meeting will cover data center locations, energy use and load growth, best practices, performance metrics, transition to direct liquid cooled data center equipment, and NLR's approach to optimizing data center.

Spatio-temporal thermal-aware job scheduling to minimize energy consumption in virtualized heterogeneous data centers

Electricity consumption for cooling purpose is known to be the most expensive operational cost factor in data centers. Inefficient cooling leads to high temperature and this in turn leads to hardware failure. This paper proposes a Thermal Aware Modified Least Slack Time Round Robin Based (MLST-RR) scheduling algorithm that can avoid high thermal stress circumstances such as large hotspots, thermal violations as well as reduce electricity consumption for cooling in data center labs. The experimental results show that, the thermal aware MLST-RR is able to significantly decrease the electricity consumption while maintaining competitive performance. Specifically, the thermal aware scheduling algorithm saves electricity consumption as much as 15000KW compared to the benchmark job scheduling algorithms such as Round Robin (RR) and First Come First Serve (FCFS); this is an electrical saving of 8.4%.

As the demand for large scale computing is rapidly increasing to serve billions of users across the world, more powerful and densely packed server configurations are being used. Often in developing countries, and in small and medium enterprises, it is hard to place such servers in sound-proof server rooms. Hence, servers are typically placed in close proximity to employees. The noise from the cooling fans in servers adversely affects employees' health, and reduces their productivity. In this paper, we provide a framework for computer architects to measure the acoustic profile in a data center along with the temperature profile, and estimate the sound power levels at points of interest. Additionally, we studied the noise levels obtained upon using algorithms targeted at homogenizing the temperature profile. We found that these algorithms result in high noise levels, sometimes above the permissible levels. So, we propose two heuristics to redistribute workloads in a data center such that noise can be reduced at certain target locations. We obtain a noise reduction of 2-13 dB when compared with uniform workload distribution, and upto 16 dB as compared to temperature aware workload placement, with a reduction of at least 5-6 dB in 75% of the cases. The performance overhead is limited to 1%.

Modern data centre energy consumption accounts for a large amount of operational cost. The idea of data centre energy optimisation deals with the job distribution between computing servers depending upon the workload. This paper projects the job role of communication material in the data centre energy consumption and proposes a scheduling approach that has both network awareness and efficient optimisation of energy consumption. This scheduling approach stabilises the energy consumption of the data centre, traffic demands and individual job performance. The proposed method enhances the server consolidation (to efficiently turn on/off the servers based on the workload) and also the distribution of traffic patterns.

In this paper we analyze the problem of optimal task scheduling for data centers. Given the available resources and tasks, we propose a fast distributed iterative algorithm, which operates over a large scale network of nodes, and allows each of the interconnected nodes to reach agreement to an optimal solution in a finite number of time steps. More specifically, the algorithm (i) is guaranteed to converge to the exact optimal scheduling plan in a finite number of time steps and, (ii) once the goal of task scheduling is achieved, it exhibits distributed stopping capabilities (i.e., it allows the nodes to distributely determine whether they can terminate the operation of the algorithm). Furthermore, the proposed algorithm operates exclusively with quantized values (i.e., the information stored, processed and exchanged between neighboring agents is subject to deterministic uniform quantization) and relies on event-driven updates (e.g., to reduce energy consumption, communication bandwidth, network congestion, and/or processor usage). We also provide examples to illustrate the operation, performance, and potential advantages of the proposed algorithm. Finally, by using extensive empirical evaluations through simulations we show that the operation of our proposed algorithm is suitable for large scale networks such as data centers.

Communication in data centers often involves many parallel flows that all share the same performance goal. A useful abstraction, coflow, is proposed to express the communication requirements of prevalent data parallel paradigms. The multiple coflow routing and scheduling problem faces challenges when deriving a good theoretical performance ratio because coexisting coflows will compete for the same network resources such as link bandwidths. In this paper, we focus on the coflow problem in the most popular data center infrastructure: the Leaf-Spine topology. We first formulate the problem and study the path selection issue on this two-tier structure. In order to minimize the average coflow completion time (CCT), we propose a Multi- hop Coflow Routing and Scheduling strategy (MCRS) for inter-coflows and intra-coflows and prove that our method has a reasonably good competitive ratio. Extensive experiments show that MCRS outperforms other state-of- art schemes.

ABSTRACTTask scheduling algorithms for data centers must be capable of making instantaneous decisions based on the current state of the system. However, due to information limitations, these scheduling algorithms often fail to achieve optimal scheduling plans. To address the information bottleneck faced in DAG (Directed Acyclic Graph) task scheduling within data centers, this paper proposes a deep reinforcement learning scheduling model based on a DAG attention mechanism. This model utilizes the attention mechanism to capture the potential relationships between dependent tasks, thereby improving scheduling efficiency and system performance under limited information conditions. The experimental results indicate that our DAG attention mechanism can significantly reduce makespan.

Eventually-consistent federated scheduling for data center workloads

Two data centers in this study were within a co-location facility located on the sixth floor of a multi-story building in downtown Los Angeles, California. The facility had 37,758 gross square feet floor area with 2-foot raised-floors in the data services area. The two data centers were designated as the west data center (DC No.18) and the east data center (DC No.19). The study found that 56% of the overall electric power was consumed by sixth floor critical loads in both data centers, 33% of the power was consumed by HVAC systems, 3% of the power was consumed by UPS units, 3% of the power was for generator losses, and the remaining 5% was used by lighting and miscellaneous loads in the building. The power density of installed computer loads (rack load) in the two data centers was 20 W/ft{sup 2} and 56 W/ft{sup 2}, respectively. The power density was relatively lower in DC No.18 compared to other data centers previously studied. In addition, HVAC to IT power demand ratio was 0.6 in DC No.18 in this study, and was 0.4 in DC No.19. Two out of three chillers were running at a low partial load, making the operation very energy inefficient. The operation and control of the chillers and air-handling units should be optimized while providing sufficient cooling to the data centers. Although arranging hot aisle/cold aisle design to separate airflow streams would be difficult in such a co-location data center, optimizing air distribution should be pursued. General recommendations for improving overall data center energy efficiency include improving the design, operation, and control of mechanical systems serving the data centers with various critical loads in place. This includes chiller operation, chilled water system, AHUs, airflow management and control in data centers. Additional specific recommendations or considerations to improve energy efficiency are provided in this report.

Existing flow scheduling schemes in Data Center Network (DCN) are designed mainly to minimize the flow complete time (FCT) of short flows and do not consider optimizing the FCT of latency-sensitive long flows (e.g. VR video streaming, interactive artificial intelligence questiona and the information-agnostic scheme (i.e. PIAS), which is based on the premise that flow size is not known a priori, requires a central server, causing a poor scalability in large network scales. Given the limitations of existing solutions, in this paper, we propose a distributed information-agnostic flow scheduling scheme (DIAS), which minimizes the FCT of both short flows and latency-sensitive long flows. In DIAS, packets are forwarded complying with their priorities, which are determined based on packets wait-time that is defined as staying time in end hosts' send buffers, and the longer a packet stays in a send buffer, the lower its priority. Meanwhile, instead of utilizing a central server to collect traffic load information, each switch feeds traffic load information which is used to adjust the thresholds of determining packets priority back to end hosts via ACK packets. The experimental results in ns-3 simulator show that DIAS reduces FCT by up to 54.7% and 50.1 % over DCTCP and $L^{2}\mathbf{DCT}$ , respectively. Besides, DIAS ensures a smaller FCT of latency-sensitive long flows and performs better than PIAS.

In distributed computing frameworks like MapReduce, Spark, and Dyrad, a coflow is a set of flows transferring data between two stages of a job. The job cannot start its next stage unless all flows in the coflow finish. To improve the execution performance of such a job, it is crucial to reduce the completion time of a coflow, as it can contribute more than 50 percent of the job completion time. While several coflow schedulers have been proposed, we observe that routing, as a factor greatly impacting the Coflow Completion Time (CCT), has not been well considered. In this article, we focus on the coflow scheduling problem and jointly consider routing and bandwidth allocation. We begin by providing an analytical solution to the problem of optimal bandwidth allocation with pre-determined routes. In the following, we formulate the problem of scheduling a single coflow as a Non-linear Mixed Integer Programming problem and present its relaxed convex optimization problem. We further propose two algorithms, CoRBA and its simplified version: CoRBA-fast that solve the single coflow scheduling problem with a joint consideration of routing and bandwidth allocation. Lastly, to address multiple coflows in online scheduling, we propose an online scheduler named OnCoRBA. By comparing with the start-of-the-art algorithms and schedulers via simulations, we demonstrate that CoRBA and CoRBA-fast reduce the CCT by 30-400 percent and the OnCoRBA scheduler reduces the average online CCT by 20-230 percent. In addition, CoRBA-fast can be hundreds times faster than CoRBA with around 8 percent performance degradation compared to CoRBA, which makes the use of CoRBA-fast very appropriate in practice.

The electricity cost of cooling systems can account for 30% of the total electricity bill of operating a data center. While many prior studies have tried to reduce the cooling energy in data centers, they cannot effectively utilize the time-varying power prices in the power market to cut the electricity bill of data center cooling. Thermal storage techniques have provided opportunities to store cooling energy in ice or water-based tanks or overcool the data center when the power price is relatively low. Consequently, when the power price is high, data centers can choose to use less electricity from power grid for cooling, resulting in a significantly reduced electricity bill.In this paper, we design and evaluate TStore, a cooling strategy that leverages thermal storage to cut the electricity bill for cooling, without causing servers in a data center to overheat. TStore checks the low prices in the hourahead power market and overcools the thermal masses in the datacenter, which can then absorb heat when the power price increases later. On a longer time scale, TStore is integrated with auxiliary thermal storage tanks, which are recently adopted by some data-centers to store energy in the form of ice when the power price is low at night, such that the stored ice can be used to cool the datacenter in daytime. We model the impacts of TStore on server temperatures based on Computational Fluid Dynamics (CFD) to consider the realistic thermal dynamics in a data center with 1,120 servers. We then evaluate TStore using workload traces from real-world data centers and power price traces from a real power market. Our results show that TStore achieves the desired cooling performance with a 16.8% less electricity bill than the current practice.

In all types of data center, keeping the right temperature with less cost and energy is one of important objective as energy saving is crucial in increased data driven industry. Energy saving is global focus for all industry. In Information technology, more than 60% of energy is utilized in data centers as it needs to be up and running. As per Avocent data center issue study, across globe more than 54% of data centers are in redesigning process to improve their efficiency and reduce operational cost and energy consumption. Data center managers and operators major challenge was how to maintain the temperature of servers with less power and energy. When the densities of data center energy nearing 5 kilowatts (kW) per cabinet, organizations are trying to find a way to manage the heat through latest technologies. Power usage per square can be reduced by incorporating liquid-cooling devices instead of increasing airflow volume. This is especially important in a data center with a typical under-floor cooling system. This research paper uses Rear-Door Heat eXchangers (RDHx) and cool logic solutions to reduce energy consumption. It gives result of implementation of Cold Logik and RDHx solution to Data center and proves that how it saves energy and power. Data center has optimized space, cooling, power and operational cost by implementing RDHx technology. This will enable to add more servers without increasing the space and reduce cooling and power cost. It also saves Data center space from heat dissipation from servers.

Greening cloud data centers in an economical way by energy trading with power grid