Modular Data Center Research Articles

MNDM: MEMS-based N-dimensional modular hybrid data center network

The rapid development of virtualization technology and cloud computing technology for the data center has brought the explosive growth of business volume and scale. The data center network is facing new challenges, such as ultra-high bandwidth requirements and energy-efficiency. The introduction of optical interconnect technology improves the performance of data center networks by virtue of less cost of devices, complexity, power consumption. However, there are still three major problems when adopting optical switching, which are constraint of centralized control system, non-modularity and poor scalability. Aiming at these problems, this letter proposes a MEMS-based N-dimensional modular hybrid data center network (MNDM) and corresponding MEMS scheduling strategy. We evaluate the performance of MNDM via OPNET simulation and results demonstrate that the MNDM architecture is stable in saturation point of latency and saturation point of throughput when the network size changes or network traffic changes.

MeGautz: A High Capacity, Fault-Tolerant and Traffic Isolated Modular Datacenter Network

The modular datacenter networks (MDCN) comprise inter- and intra-container networks. Although it simplifies the construction and maintenance of mega-datacenters, interconnecting hundreds of containers and supporting online data-intensive services is still challenging. In this paper, we present meGautz, which is the first inter-container network that isolates inter- and intra-container traffic, and it has the following advantages. First, meGautz offers uniform high capacity among servers in the different containers, and balances loads at the container, switch, and server levels. Second, it achieves traffic isolation and allocates bandwidth evenly. Therefore, even under an all-to-all traffic pattern, the inter- and intra-container networks can deal with their own flows without interfering with each other, and both can gain high throughput. meGautz hence improves the performance of both the entire MDCN and individual servers, for there is no performance loss caused by resource competition. Third, meGautz is the first to achieve as graceful performance degradation as computation and storage do. Results from theoretical analysis and experiments demonstrate that meGautz is a high-capacity, fault-tolerant, and traffic isolated inter-container network.

Energy and Exergy Analysis of Modular Data Centers

The data center industry focuses on initiatives to reduce its enormous energy consumption and to minimize its adverse environmental impact. Modular data centers provide considerable operational flexibility in that they are mobile and are manufactured using standard containers. This paper develops steady-state energy and exergy destruction models for modular data centers with the open-source EnergyPlus software package. Three different cooling approaches are examined: direct expansion (DX) cooling, evaporative cooling (direct evaporative cooling, DEC, in this study), and free air cooling (air-side economization in this study). This paper shows that for hot and arid climates like those in the southwestern U.S., augmenting DX cooling with evaporative and free air cooling can result in energy savings of up to 38% and 36%, respectively. This paper also applies exergy analysis to suggest that the Energy Reuse Effectiveness of the data center increases with decreasing ambient (outdoor) temperature and increasing server inlet–outlet temperature difference. Furthermore, simulations indicate that the use of passive cooling techniques (e.g., DEC and free air cooling) decrease data center heating, ventilation, and air-conditioning energy consumption, except in extremely hot and humid climates.

OpenFlow-Based Scalable Routing With Hybrid Addressing in Data Center Networks

In this letter, we propose an OpenFlow (OF)-based SCAlable Routing strategy (OSCAR) for modular data center networks (DCN) using a hybrid addressing mechanism. Each module in the DCN constitutes a segment in the network. Inter-segment routing is performed using virtual MAC (VMAC) ids assigned to the segments and intra-segment routing is done using IP addresses. OF provides central control over the network but it suffers from limited scalability in DCNs due to high control traffic. In OSCAR, the control traffic is minimized to achieve high scalability and flexibility in DCN routing.

Qualitative Study of Cumulative Corrosion Damage of Information Technology Equipment in a Data Center Utilizing Air-Side Economizer Operating in Recommended and Expanded ASHRAE Envelope

Deployment of airside economizers (ASEs) in data centers is rapidly gaining acceptance to reduce cost of cooling energy by reducing hours of operation of computer room air conditioning (CRAC) units. Airside economization has associated risk of introducing gaseous and particulate contamination into data centers, thus degrading the reliability of information technology (IT) equipment. The challenge is to determine reliability degradation of IT equipment if operated in environmental conditions outside American Society of Heating, Refrigeration, and Air-Conditioning Engineers (ASHRAE) recommended envelope with contamination severity levels higher than G1. This paper is a first attempt at addressing this challenge by studying the cumulative corrosion damage to IT equipment operated in an experimental modular data center (MDC) located in an industrial area with measured level of air contaminants in ISA severity level G2. This study serves several purposes including: correlating IT equipment reliability to levels of airborne corrosive contaminants and studying degree of reliability degradation when IT equipment is operated outside ASHRAE recommended envelope at a location with high levels of contaminants. Reliability degradation of servers exposed to outside air via an airside economizer was determined qualitatively by examining corrosion of components in these servers and comparing the results to corrosion of components in other similar servers that were stored in a space where airside economization was not used. In the 4 years of the modular data center's servers' operation, none of the servers failed. This observation highlights an opportunity to significantly save data center cooling energy by allowing IT equipment to operate outside the currently recommended and allowable ASHRAE envelopes and outside the ISA severity level G1.

A Monte-Carlo approach to lifespan failure performance analysis of the network fabric in modular data centers

Data centers have been evolved from a passive element of compute infrastructure to become an active, core part of any ICT solution. In particular, modular data centers (MDCs), which are a promising design approach to improve resiliency of data centers, can play a key role in deploying ICT infrastructure in remote and inhospitable environments in order to take advantage of low temperatures and hydro- and wind-electric capabilities. This is because of capability of the modular data centers to survive even in lack of continuous on-site maintenance and support. The most critical part of a data center is its network fabric that could impede the whole system even if all other components are fully functional, assuming that other analyses have been already performed to ensure the reliability of the underlying infrastructure and support systems. In this work, a complete failure analysis of modular data centers using failure models of various components including servers, switches, and links is performed using a proposed Monte-Carlo approach. The proposed Monte-Carlo approach, which is based on the concept of snapshots, allows us to effectively calculate the performance of a design along its lifespan even up to the terminal stages. To show the capabilities of the proposed approach, various network topologies, such as FatTree, BCube, MDCube, and their modifications, are considered. The performance and also the lifespan of each topology design in the presence of failures of their components are studied against the topology parameters.

Spatial Division Multiplexing for High Capacity Optical Interconnects in Modular Data Centers

Modular design has recently emerged as an efficient solution to build large data center (DC) facilities. Modular DCs are based on stand-alone prefabricated modules (i.e., PODs) that can be easily installed and interconnected. PODs can generate a large amount of traffic and thus require an ultra-high-capacity interconnection network. However, current electronic and optical interconnect architectures applied to modular DCs may experience major scalability problems in terms of high energy consumption and cabling complexity. To address these problems, we investigate five optical interconnect architectures based on spatial division multiplexing (SDM), and for each architecture, we propose a resource allocation strategy. We also present an extensive comparison among the SDM architectures in terms of cost and performance (i.e., blocking probability and throughput), with the objective to find the architecture offering the best trade-off between cost and performance for given DC sizes and traffic load values. Our results demonstrate that, in small modular DCs with low traffic load, an architecture based only on SDM is the best option, while in medium DCs with medium traffic load, an architecture based on coupled SDM and flexgrid wavelength division multiplexing (WDM) with spectral flexibility is the best solution. Finally, for large DCs with high traffic load values, the best trade-off between cost and performance is achieved by an SDM architecture that is based on uncoupled SDM and flexgrid WDM.

Experimental investigation on the thermal performance of water cooled multi-split heat pipe system (MSHPS) for space cooling in modular data centers

A water cooled multi-split heat pipe system (MSHPS) was proposed to cool the space in modular data centers. It combines high thermal performance heat pipe technology with cold aisle containment architecture technology to remove heat effectively and minimize the short-circuiting of the hot air within the IT server. Experimental investigations were conducted to identify the optimal refrigerant filling ratio and the effect of different inlet air temperatures and different inlet cooling water temperatures on the unit of MSHPS. As a result of experiments, the optimal refrigerant filling ratio was found in the range from 33% to 42% with cooling capacity range from 6100W to 6200W and the system thermal resistance of evaporation process ranged from 0.0028K/W to 0.0030K/W. The effect of inlet temperature of air and cooling water were analyzed in terms of the cooling capacity, refrigerant temperature and pressure distribution along the unit of MSHPS. These results are useful for the design and operational control of the MSHPS for modular data centers.

Adaptive flow scheduling for modular datacenter networks

Modular data center networks leverage recursive topology to speed up several traffic patterns and increase the network capacity for various data shuffle applications. However, the applications of online social networks and instant messaging may be deployed in the same data center for high cost-effective investment. The large amount of mice flows which produced from these applications may cause partial path congestion, and these congestion in turn can prevent the mice flow access that reduces message loss or delay for online social networks. In this stydy, we define the cost function for different parallel paths and model the path selection process, and propose an adaptive flow scheduling for modular data center networks (AFMD) to search more suitable path for both large flows and mice flows with the greedy algorithm. The experiment shows that AFMD can improve large flow throughput by up to 25.9 % and the mice flow access ratio by up to 16.5 % and AFMD exhibits effective complementarity and compatibility to BCube networks.

Impact of aisle containment on energy performance of a data center when using an integrated water-side economizer

The main purpose of this study is to analyze the impact of cold aisle containment on the thermal management performance of a data center, as well as on the amount of energy consumed for cooling. The thermal management performance of the aisle containment architecture, which has been widely employed throughout the data center industry, is demonstrated by experiment. A computational fluid dynamics (CFD) modeling strategy is used for the containment air leakage prediction validated by experiments. The thermal management performance of an uncontained and contained aisles are compared through CFD simulation. In addition, since the containment's energy performance is maximized when an economizer is incorporated, a model-based simulation of the annual cooling energy consumption of a modular data center with an integrated water-side economizer is conducted. The simulation results show that the thermal management performance is excellent in both architectures with an appropriate cooling system control, but the energy savings are much greater with the contained architecture.

Перспективы использования центров обработки данных при решении задач математической биологии и биоинформатики

The article examines the state of modern "science of life" - computational biology, discusses the problem of "big data" and the ways to resolve this extremely relevant to modern science tasks associated with storage and online analytical processing of huge volumes of information received during scientific experiments. Also it analyzes the capacity of data centers and, particularly, the modular data centers as the primary means of ensuring of scientific centers with computational resources necessary to successfully solve their research problems.

Cooling Control of Modular Data Center Utilizing Fresh-Air in Consideration of ICT Equipment Temperature Management and Energy Saving Based on Model Predictive Control

Cooling Control of Modular Data Center Utilizing Fresh-Air in Consideration of ICT Equipment Temperature Management and Energy Saving Based on Model Predictive Control

モデル予測制御法によるICT機器温度管理と省エネルギーを考慮した外気導入型データセンタの冷却制御

モデル予測制御法によるICT機器温度管理と省エネルギーを考慮した外気導入型データセンタの冷却制御

Optimum supply air temperature ranges of various air-side economizers in a modular data center

In order to reduce data center cooling energy, ASHRAE Technical Committee 9.9 expanded thermal environmental ranges for data centers and recommended using the economizer, and it has reduced data center cooling energy successfully. The main purpose of this research is to establish an energy optimization process for the air-side economizer in a modular data center with respect to various parameters (supply air conditions, server thermal characteristics, cooling system configurations, and heat exchange effectiveness), and to determine optimum supply air ranges through simulation. The cooling energy simulation is conducted for three types of air-side economizers by changing the supply air temperature (SAT) and heat exchanger effectiveness at 16 locations in South Korea based on the three thermal environmental ranges. The simulation result shows that the lowest cooling energy consumption appears when the computer room air handler (CRAH) SAT is within 18–23 °C, and at higher SAT conditions, the cooling energy consumption increases, since chiller energy reduction is offset by the increase in CRAH fan energy.

Energy saving potential of various air-side economizers in a modular data center

With the recent development of the IT technology, the demand for data centers has significantly increased, and modular data centers have attracted considerable attention because of their excellent stability, scalability, and economic feasibility. This research quantitatively analyzed the applicability of various air-side economizers and their energy-saving potential in modular data centers. A detailed cooling load estimation process was established for modular data centers, and annual cooling energy simulations were carried out using various air-side economizers. The various air-side economizers yielded cooling coil load savings of 76–99% in comparison to conventional cooling systems in data centers, and the total cooling energy savings of the economizers ranged from 47.5% to 67.2%. Indirect air-side economizers with high-effectiveness heat exchangers were found to yield significant energy saving (63.6%) and have simple system configurations.

Simplified server model to simulate data center cooling energy consumption

This paper proposes a server model that can be applied to simulating the annual cooling energy consumption of data centers and an analysis of its impact. Recently, aisle containment architecture and the variable air volume system have been widely adopted in the data center industry because of their superior indoor thermal management and cooling energy savings. However, the current simulation method, which fixes the server heat generation and the supply and return air temperature difference of the computer room air handler (CRAH), does not correctly reflect current cooling systems. In this study, a server model was developed that accounts for server thermal characteristics – including server power, fan airflow, and exhaust temperature – according to the given CPU utilization and temperature. After the proposed server model was validated, the annual cooling energy consumption was simulated for modular data centers with an air-side economizer and high ambient temperature. Unlike the conventional method, the proposed model showed that the cooling energy consumption increased when the CRAH supply air temperature was higher than 19°C because of the increase in fan energy consumption.

모듈러 데이터 센터에 적용된 다양한 외기냉방 시스템의 급기 조건 및 지역별 에너지 절감효과 분석

The main purpose of this paper is to analyze energy saving potentials of various economizers in modular data center by thermal conditions and locations in response to explosive IT demand and data centers' energy consumption increase. To investigate holistic energy saving potentials, various server thermal characteristics, economizers, and locations are analyzed on a modular data center. As a results, economizers can save cooling coil load 75-100% according to their running conditions, and indirect air-side economizer showed better performance than direct air-side economizer. Even though there is more chiller energy saving with server inlet temperature increase, total energy saving decreased more sharply due to fan energy increase. Furthermore, there should be more energy saving potentials in high latitude climates regardless of type of economizers.

A Model Transformation Approache of Complex Systems - Case Study in Design Evolutions of Data Centers

Unpredictable transaction requirements of IT business lead to miss design the right size of data center. Over design data center contributed to surplus capital investment and lifetime operations. Legacy data centers designed before the 2ndmillenniums over design capacity more than 60% of actual load. The research objectives are created a model transformation approach from legacy data center to mobile and modular data center M2DC and proposed multivariate optimization for the right sizing of data center as business needs by using case study. The research method is investigation and assessment through 21 sample data centers and in-depth interviews with IT managers (32) and data center consultants (8). The fact findings have shown the standardized modular of M2DC force requirements to fit in the building boxes and expansion as needs.

Distributed thermal aware load balancing for cooling of modular data centres

Thermal management in data centres requires a complicated trade-off between cooling costs and thermally induced equipment failure rates. Using ideas from cooperative control and distributed rate limiting, in this study the authors describe a distributed architecture that can be used for thermal aware load balancing for a common type of modular data centre; namely an algorithm that, for a given demand D *, distributes load to individual machines such that the temperatures in the individual modules are equalised. The benefit of shifting load based on thermal considerations is that significant gains in cooling cost can be achieved. We evaluate the performance of the algorithm using computational fluid dynamics and Matlab simulations. The results show that significant cost savings can be made by applying such algorithms, and that these can be achieved without the need for detailed modelling and tuning of controllers.

A new high performance and fault-tolerant datacenter network for modular datacenters proposed by researchers from NUDT

A new high performance and fault-tolerant datacenter network for modular datacenters proposed by researchers from NUDT