See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/340477624

Data center maintenance: applications and future research directions

Article in Facilities · April 2020

DOI: 10.1108/F-09-2019-0104

CITATIONS READS

6 8,898

3 authors:

Mostafa Fadaee Fariborz Haghighat

Concordia University Montreal Concordia University Montreal
5 PUBLICATIONS 56 CITATIONS 481 PUBLICATIONS 20,260 CITATIONS

SEE PROFILE SEE PROFILE

Fuzhan Nasiri
Concordia University Montreal
134 PUBLICATIONS 2,567 CITATIONS

SEE PROFILE

All content following this page was uploaded by Fuzhan Nasiri on 26 April 2020.

The user has requested enhancement of the downloaded file.

 Data Center Maintenance - Applications and Future Research Directions

Fadaeefath Abadi, Mostafa1,2; Haghighat, Fariborz1 & Nasiri, Fuzhan1

1
Department of Building, Civil and Environmental Engineering Concordia University, Montreal, Canada
2
[email protected]

Abstract
One of the most critical and major infrastructures is a Data Center (DC) due to having a large
number of servers, computers and other equipment. They provide online and internet services for
various companies in the Information Technology (IT) industry. DC facilities should also
provide reliable online services while addressing the required minimum quality and performance
level considering optimized energy and power consumption. This paper provides an organized
and systematic literature review focusing on topics regarding the operation and maintenance
management of DCs. The objective is to represent and classify the main findings in this area and
to identify the main research gaps and shortcomings from the perspective of research. Although
there are several studies reported on operations and maintenance management systems for other
industrial systems, a limited number of studies with few methods and models have focused on
DCs so far and this area of application requires more attention. This paper reviews the most
recent research developments in identifying issues, problems, and challenges for DC buildings
and facilities. Main solutions and strategies found by scholars to solve or reduce the issues are
also reviewed. Finally, a conclusion of their findings is conducted in order to discover new
potential methods as future research opportunities to achieve more advancements in this area.

Keywords: Data Center (DC), Information Technology (IT), Facilities Management (FM),
Operations & Maintenance (O&M) Management.

1. Introduction
A Data Center (DC) is a major infrastructure and facility which contains a high quantity of
servers and computers to provide internet services for many companies in the world (such as
Google, Amazon, etc.). The term “Data Center” may have different meanings from various
points of view and there are also some other similar names such as data hall, data farm, data
warehouse, computer room, server room, etc. which have been used by researchers and scholars.
According to U.S. Environment Protection Agency (EPA), DC is a type of principle electronic
equipment that is used for processing data (server devices), storing the data (storage equipment)
and communications including networking tools and devices. DC requires special power
conversion and backup equipment to preserve high quality and reliable power system. The
system maintains important environmental factors such as proper temperature and humidity for
the Information and Communication Technology (ICT) equipment (Geng, 2015). The principal
components and equipment located in DCs are computers, servers, cooling systems, storage
devices, and supply devices. Three main subsystems located in a general (traditional) DC are
continuous power supply, air conditioning and network connectivity (Saha et al., 2016).
Corresponding to different references, there have been various changes in the technology of DCs
such as other technologies in the world. As there were computer rooms in the middle of the 20th
century, DC began to operate in those computer rooms as a shelter or house. Later, these
computer housings became large and complex infrastructures. The ARC net is recognized as the
first DC, placed in a special area. About 10,000 ARC net DCs were used globally by 1980. From
the beginning of the 1990 decade, as the usage of personal computers was increasing in the
world, users had access to a network. Therefore, the real difference appeared and the mainframe
which existed before, was converted into places hosting many servers. These places were later
named “Data Centers” (Carey, Raisinghani and White, 2017).
DCs are industrialized centers located worldwide and the number of them is growing due to
high demands for several services. As specified by references and reports, the total number of
DCs worldwide in 2017, was about 8.4 million. Based on the literature, there are about 8.6
million enterprise data centers operating globally. Also, the number of Cloud Data Centers
(which will be elaborated in the next pages) has been increased recently. Moreover, DC facilities
have been combined and became hyperscale DCs. Therefore, because of the stated reasons, the
total quantity of DCs have been decreased from 8.55 million in 2015 to about 8.4 million in 2017
but the demand for data storage and processing is rising. It is estimated that this number will
reduce to 7.2 million until 2021 (Statista, 2017; Williams, 2018). According to statistics, the
amount of data center storage consumed by big data worldwide will require about 124 exabytes
of DC storage space globally until the end of 2018 and will grow up to 403 exabytes by 2021
(Cisco Systems, 2018b). In addition, in 2018, the storage capacity of DCs reached 1450 exabytes
in the world with 250 exabytes being consumed by collaboration activities and data. The storage
capacity will increase up to near 2500 exabytes by 2021 (Cisco Systems, 2018a).
Regarding the provided statistics related to DCs, there are several challenges and issues in
managing DC facilities and their components. Especially, the amount of power being consumed
by servers is noticeable. Concerning energy consumption, it should be noted that DCs utilize
about 40 times energy than other standard buildings such as offices. Thus, energy management
tools that have been used for regular office building could not be easily applied for DCs since the
energy simulation of DCs is complicated. There are considerable differences in terms of energy
loads and the structure of buildings between DCs and commercial buildings (Cho et al., 2015).
Research and studies have been performed on DC operations by scientists to effectively apply
the DC operations management framework for reducing the staff workload. In addition, the other
goal is increasing the work efficiency to bring maximum overall effectiveness for the DC
operations management system. (Bai and Geng, 2015). In terms of DC power and energy usage,
the indicated numbers express high and increasing energy demands. The servers and DCs have
used 61 billion kWh, costing $4.5 billion in the year 2006 and it was expected to be doubled by
the year 2011. Based on recent studies, one server with 300W power has about $330 energy cost
in a year (Guenter, Jain and Williams, 2011). In addition, the total electricity consumption of
DCs in the United States has reached 60 billion kWh per year in 2014 and will rise to 70 billion
kWh per year until 2020. Fortunately, based on the presented numbers, responsiveness about
infrastructure operations in DC has increased and thus the efficiency has been improved,
specifically in cloud DCs. Moreover, recently, the cloud DCs are designed to maximize the
efficiency of the DC infrastructure. Generally, Cloud DCs host around 10,000 servers for
offering various services to be used by the applications (Saha et al., 2016; Shehabi et al., 2016).
DCs play an essential role in our fast-developing society as main Information Technology
(IT) infrastructures. Due to the complexity of IT equipment placed in DCs, the reliability,
availability, and maintenance of its components require more attention. Based on recent studies,
DCs should adjust themselves to the fast-developing industry. In addition, the Operations and
Maintenance Management System should carefully be considered and formed based on the
required standards to obtain a standard management system to meet DC’s acceptable
performance (Zhan and Zhang, 2014).
2. Literature Review
In this paper, at first, review papers have been searched in the field of operations and
maintenance management systems, generally. The purpose was to find previous similar literature
reviews, published in the last five years, which have discussed these systems generally or for
specific applications or systems in the industry. Based on findings, up to now, there are a few
papers that have focused on various models and methods of operations and maintenance
management. The most relevant review papers have been selected and discussed in the next
paragraphs of this section. They present a wide range of studies about the maintenance and
operations, their evolution and application in different systems. They have also discussed,
classified and reviewed the applied methodologies.
Generally, the maintenance management process in any system, including DCs, could be
classified in two main steps. First, the approach is determined and then, the strategy should be
implemented in the system (Márquez, 2007). The latter phase consists of (1) Maintenance
planning for all assets in the system, (2) Maintenance scheduling, (3) Data collection to manage
maintenance actions, (4) Maintenance assessment by having appropriate measurements, (5)
Checking the improvement of maintenance tasks by conducting the required modifications, and
(6) Analyzing any necessary modifications for re-designing any components in the system
(Márquez, 2007).
Various models of maintenance management systems have been categorized by researchers to
demonstrate the empirical evidence for using these models in real worldwide projects. They
obtained a solution to improve the Empirical Evidence Rates (EER) for maintenance. The
scholars have searched for various maintenance management models in several articles and
finally found 37 different models (Fraser, Hvolby and Tseng, 2015).
According to a literature review focusing on maintenance and facility management, while the
warehouse’s operational performance and logistics are influenced by facility management, there
is a lack of research in this area to highlight the importance of facility management for logistics
operations (Mangano and de Marco, 2014). In addition, other researchers have identified the
main issues of maintenance operations by comparing them with the problems found in the
manufacturing operations. A systematic approach is performed to classify the literature and to
analyze them. Based on calculations, the maintenance operations, and manufacturing operations
have similar problems such as not having top management support, not measuring the Overall
Equipment Effectiveness (OEE) and not presenting strategic planning and implementation. In
regard to their discussions, both performance and productivity of machines should be considered
for proper maintenance (Phogat, Gupta, Kumar and Gupta, 2017). Research on building’s
dynamic energy performance was reviewed to highlight the main developments of the major
phases such as monitoring, diagnosing, and retrofitting. The goal was to concentrate on the
operations and maintenance phase, as a part of the diagnosing phase, and to review main
advances for reducing the carbon emissions in this phase of the buildings. The operational rating
system, which is a measurement tool for applying in the building’s operation and maintenance
was discussed (Hong et al., 2015). A comprehensive literature review is published in two parts
concerning performance indicators for facilities operations and maintenance. In the first part (Lai
and Man, 2018a), the most important performance indicators (about 70) are mapped and
classified based on the Phase-Hierarchy (P-H) model. These indicators were defined and
classified from the Facility Management (FM) perspective. It could be a useful source of various
indicators which is essential in most types of buildings. In the second part, the Key Performance
Indicators (KPIs) are discussed with groups of Operations and Maintenance (O&M) experts in
the commercial buildings sector to obtain their comments and ideas for shortlisting the total
performance indicators found in the first part. This shortlist of KPIs have the potential to be used
by professionals of FM in commercial buildings. Also, the same approach for obtaining a
shortlist of the important indicators could be applied to other building sectors (Lai and Man,
2018b). In the case of office buildings, extensive research based on interviews and surveys have
been performed to develop an FM process model. This model includes essential parameters for
office buildings to improve their profit. Operations and maintenance have been implemented in
this process model as an important factor (Shin et al., 2018). Operations and maintenance
expenses have been studied by reviewing different factors of building maintenance. In the study,
different problems and issues in building maintenance scenarios in Malaysia are discussed to find
solutions to reduce the costs (Che-Ghani, Myeda and Ali, 2016).
A review of maintenance management systems was conducted (Yousefli, Nasiri and Moselhi,
2017) in order to identify and classify current literature and studies in the field of maintenance
management for healthcare facilities. Several applied methods, such as Integrated performance-
based maintenance management and area-based weighting method were discussed by authors. In
addition, another paper has reviewed different maintenance strategies (Pro-active, Preventive,
Predictive and Corrective) and investigated their relationship with customer satisfaction in the
case of healthcare facilities and hospitals to review different types of maintenance strategy
applied to maintain facilities while satisfying the end-user to obtain a correlation result between
types of maintenance strategies and customer satisfaction (Rani et al., 2015).
Recent works on the Data Center Network (DCN), which is an important network connecting
various devices in the DC infrastructure, was reviewed (Xia et al., 2017). The researchers have
classified main studies in this area and used surveys to describe their findings and to represent
future research opportunities. In addition, a review of the FM and maintenance strategies has
been recently conducted focusing on facility maintenance and management concepts of DC
operations. The study provides future research work for having more reliable operations in DCs
(Bieser and Menzel, 2019).
In addition to review papers regarding operations and maintenance, some recent research
work discussing the relation of FM and maintenance have been found and included in this
section. The impact of failures such as natural disasters or manmade disasters on the facility
management organization has been studied to analyze its vulnerabilities. The study was
performed by simulating and modeling the organization network using surveys, defining the
failure modes and evaluating the system’s stability. The operations and maintenance of the
infrastructure were considered with various elements and eventually, the essential parameters for
the organization’s maintainability have been obtained (Tang, Cao and Shao, 2017). In order to
study the role of maintenance in facility management, research has been done in the case of
building facilities by applying a Fuzzy-AHP-based Decision Support System (FADSS). By using
this method, the optimal and proper maintenance strategy considering the improvement of
efficiency and minimizing maintenance cost has been achieved (Pun et al., 2017).
In the areas of Building Information Modeling (BIM), FM and maintenance, some research
works have been done. BIM and Augmented Reality (AR) are integrated together to provide an
environment that is Ambient Intelligent (AmI). This approach could assist facility management
in performing maintenance tasks. The research considered healthcare facility management as the
area of application (Irizarry et al., 2014). The importance of BIM and big data has been also
discussed to present a model that could be used in maintenance and FM. Based on the research,
maintenance information, as an important part of FM, is required to create the Conditions Data
Model (CDM). Researchers have concluded that applying the CDM framework has several
advantages for building owners, service providers and users by enhancing the quality of service
(Halmetoja, 2019).
The field of green maintainability in buildings has become an interesting topic for researches.
Studies have been conducted on the green building considering FM, operations and maintenance.
In addition, high-rise Vertical greenery systems (VGS) have been studied by conducting
interviews and surveys to discover their main challenges and issues regarding maintenance costs
and considerations (Chew, Conejos and Asmone, 2017; Conejos, Chew and Azril, 2019).
A comprehensive search has been performed to find published research works on topics
related to operations, maintenance, and management of DCs. Based on findings, literature review
regarding operations and maintenance management issues for DC and IT facilities is very
limited. Therefore, according to the lack of such a comprehensive review, this paper is focused
on identifying and organizing the available literature, mostly during the last 10 years, about
current strategies and solutions proposed by researchers for managing the operations and
maintenance issues for the DC infrastructure. Also, optimization algorithms were implemented
and applied by researchers to adjust operational components in DCs. These methods are
reviewed systematically and included in this paper.
In this review, the authors have discovered the most recorded keywords in publications related
to operations and maintenance management in the Engineering Village – Elsevier database. To
limit the results, the search has been narrowed down to specific terms including Maintenance,
Decision making, Reliability and Scheduling. More than 50 important keywords representing the
topics and methods of these publications are found and presented in Figure 1. Achieving these
keywords is useful for demonstrating the main concerns and trends regarding the operations and
maintenance management field. In addition, frequent keywords used in the literature by
researchers about DC’s operations and maintenance have been obtained separately and shown in
Table 1.
 Figure 1: Most used keywords in publications related to operations and maintenance management (Elsevier, 2019)
Table 1: Most frequent keywords in the topic of Data Center Design, Planning, Operations and
Maintenance Management
Data Center Operations and
Data Center Design and Planning
Maintenance (O&M) Management
Data Center Maintenance
Data Center Infrastructure Management (DCIM) Facility Management
Cooling Systems and Technologies Performance
Keywords

Cloud Computing (Cloud) Data Center Simulation

Project Management Operations and Maintenance (O&M) Management
Network Topology Energy
Building(s) Cost
 Modular Data Center (MDC) Maintenance Scheduling
Data Center Networking (DCN) Quality
Internet Data Centers (IDC) Reliability-Centered Maintenance (RCM)
Event Management System Availability
Sustainable Construction Management Power Management
HVAC System Key Performance Indicator (KPI)
Internet of Things (IoT) Information Technology Infrastructure Library (ITIL)
Failure
Efficiency
Maintenance effect
Preventive Maintenance
Reliability Based Maintenance
Task Scheduling

The next sections of this review paper are organized based on the following categories: The
Infrastructure Management (IM) topic in DCs is introduced and discussed in Section 2. A review
of recent studies and research work in the field of Operations Management (OM) in the DC
facility is expressed in section 3. Section 4 describes literature and publications regarding the
Maintenance Management of DC infrastructure by illustrating relevant methods and models. In
the fifth section, a discussion based on previous sections (2 to 4) is provided in order to
summarize the reviewed papers and determine the main research gaps and limitations, obtained
from the literature. Finally, in the conclusions section, a summary of this paper including the
main research gaps and limitations is presented and according to the findings from the previous
section, appropriate ideas for future research directions are demonstrated. To represent the topics
covered and reviewed in this paper more clearly, the paper’s structure diagram has been provided
in Figure 2.

Figure 2: Structure of this literature review paper

Literature Review

Operations & Maintenance Management

Area of Application
Methods/Models Operations Maintenance
(Case study)

Reliability-Centered Project
Data Centers Scheduling
Maintenance (RCM) Management
3. Data Center Infrastructure Management (DCIM)
It is worthwhile to note that the DC infrastructure expense makes up to one-third of the total
IT budget. This amount of investment is also used for DC operations. Unfortunately, companies
have not completely focused on applying new methods, models and technologies to manage their
capital expenses related to the DC infrastructure operations. This section focuses on the DC
Infrastructure Management (DCIM) definition which has been introduced recently in order to
assist in managing and monitoring several systems in a DC (Geng, 2015). Other studies have
considered the DCIM defining it as a combination of the IT systems management with FM
systems (Hubbell et al., 2015). Before introducing DCIM, IT professionals were engaged only
with IT equipment (servers, storage, networks, etc.) and facility managers took care of the DC
building. But now due to increasing investments in High Performance Computing (HPC) and
cloud services infrastructure, it would be more effective if DC and its IT equipment be
considered as one single system to bring more benefits. Some useful tools and methods such as
the Big Data model, 3D gaming technology and 3D monitoring and management (MM3D),
which has been developed at MIT Lincoln Laboratory, were used to provide in-depth analysis of
historical trends and conditions to optimize operations for reaching higher efficiency and
reliability. Managing to compute resources and identifying troubles in the system and providing
real-time insight into critical facilities systems, IT infrastructure, lifecycle management, cost
reduction strategies, historical analytic capabilities with less effort and cost are other objectives
of these methods. The US government along with major companies (Google, Microsoft, and
Facebook) are studying the expenses for the DCIM implantation. Based on results, some future
works could be conducted towards developing a smart tool for green DCs (such as delaying the
jobs until the evening hours or off-peak hours due to lower electric rates) (Hubbell et al., 2015).
Researchers have noted that DCIM is used to monitor, check, control and manage the energy
usage of IT devices in a DC. It has principle points which are input, process, and output. Devices
such as sensors are considered as input components and are transmitted in the process stage to
collect and analyze various data and finally are offered to the users as the output component. The
information of DCIM assists DC asset management as a useful source. Important parameters
such as space, power and cooling capacity could be predicted well. The availability of DC is
improved by receiving information from the current power level and cooling capacity.
Operational costs will be decreased by improving total efficiency (Abbas, Orozco and Villalba,
2015).
4. Data Center Operations Management
An operation is defined as a task that should be conducted in a service many times during an
application’s performance (Diouri et al., 2015). Since DC should always provide the required
services and operations, any interruption in its services will cause major problems and will have
negative effects on the whole system. Therefore, fast recovery of operations in an essential issue
in the case of DCs. Providing cloud services has become an important factor for DCs due to
giving information to numerous users worldwide which are using several types of intelligent and
portable gadgets and devices. Reports prepared by organizations indicate that the total DC traffic
will reach up to 8.6 zettabytes in 2018 (Janacek and Nebel, 2017). In section three of this paper,
the authors have presented recent research and studies on various aspects of DC operations such
as DC cooling technologies, quality and project management in a DC, issues in DC networks
operations and the operational expenses.
 Selecting the appropriate cooling system will also reduce the operations and maintenance
costs and increase the overall performance of a DC. The investigation is in progress on newer
technologies such as free-air cooling and liquid cooling to be used instead of the traditional air
cooling system (Li et al., 2016). Researchers are evaluating economic cooling systems for DCs
to have them energy-optimized. By performing numerical and simulation studies and considering
a dedicated cooling system and design conditions, selecting the best system could be possible
(Cho et al., 2015). Since cooling devices consume the most energy in DCs, energy-saving
methods are being tested in case studies to obtain significant energy saving (Hassan et al., 2015).
4.1. Quality Control and Project Management in DCs
Quality Management System and Quality Control procedures are important facts in managing
efficient DCs. A recent study has identified service quality assessment as a required tool to be
applied during the operation of a DC. Several analyses were performed on optimizing the cost of
service quality by considering some factors such as job numbers to decrease waiting times and
loss rate which have a significate effect on improving quality and finally minimizing its cost
(Bayati, 2016). In addition, a quality service model was implemented by scholars to improve the
system reliability in a DC. To reach customer expectations, service quality is a key issue that
should be have been taken into consideration seriously. The Analytic Hierarchy Process (AHP)
method is used to develop a quality service model. Several values of service quality (such as
reliability, credibility, security, access, competence, etc.) were selected and organized to make
the best decision for the costumer’s satisfaction. Their proposed model integrates the system
from the design and planning stage to the controlling and operations level. Based on the
methodology in their study, conversations have been performed with DC companies, IT
specialists, operation managers and other technicians in this area to find the main design
standards, DC design shortcomings, and its framework. As mentioned, the AHP method is
utilized to obtain the service success for problems that occurs in a DC complex. However, there
are some limitations, such as addressing only controllable and productive factors (internal
factors), in their study (Wiboonrat and Kaewsiri, 2014). About DCPM, newer research has
implemented the integrated ISO 26702 and IEEE 1490 standards in the Data Center Project
(DCP) while investigating the stakeholder’s participation. The target was developing a
framework for DC Systems Engineering (SE) via process integration regarding both ISO and
IEEE standards. The highlighted objectives of the study were minimizing user’s complaints on
requirements that not match with DC specifications and design, finding the effect of these
modifications on the system breakdown structure (SBS), comparing the effectiveness and trade-
off analysis in the system and finally, obtaining the role of standards in DC project success. The
researcher has applied international project management standards with project management
tools and techniques for DC. Also, modeling, prototyping, and development strategies were
utilized (Wiboonrat, 2016). The project success model has been defined and elaborated in the
literature, which considers the time domain of several activities. These tasks include operation
management, marketing management, and service management. Leading a project to success is
an uncertain strategy and dynamic solutions. It has both short project success corresponding to
reliable operations and long project success which is related to time. Thus, operations
management and marketing management are the most important parameters of project success.
In a DC, the most critical factors based on recent studies are the design and planning stage and
then, implementing, monitoring and controlling the system. Moreover, the system maintenance
and operations process is taken into consideration. The devices and components such as
networks, servers, storage, operating systems, applications are important. Also, the installation
and operation of applications need to be considered to find critical factors that have an effect on
project success or service quality (Wiboonrat and Kaewsiri, 2014).
4.2. Data Center Network Operations
The DC Network or DC Networking (DCN) has been considered by researchers (Xia et al.,
2017). It is an interconnection in the DC infrastructure which is designed specifically. In order to
optimize DC operations, the DCN has an important effect. In a DC, a trade-off among operation
costs and the availability of required services should be considered. This research reviews some
critical issues of DCN operations such as the huge scale of DC because of having numerous
servers. This causes economic, design, operations, and other considerations. For instance, there
are more than 100 Microsoft DCs worldwide each having a tremendous number of servers.
Another challenge is a high quantity of online services available in a DC which creates high
traffic and workload. High power and energy usage, as mentioned before in the introduction
section, is one of the main concerns in many studies and it is specially investigated in this survey.
Finally, avoiding or reducing failures during a DC operation is evaluated. While DC should be
available all the time, having a strong system in an essential matter. Thus, monitoring various
component failures and providing recovery strategies is a must in the DCN (Xia et al., 2017). An
important cause of failure in a DC, discussed by researchers, is the heat waste produced by the
high-density blade servers located in the DCs. This heat wastage increases the failure of
hardware and thus, the DC maintenance costs will rise significantly (Wang et al., 2009). Internet
Data Centers (IDCs) have been studied in terms of operational issues (Wang et al., 2012). As
they are an important type of Cyber-Physical Systems (CPCs), they should provide dependable
operations for many online services (email, blogging, social networks, online business, etc.).
While these services are increasing every day, their operational costs (running the system,
cooling equipment, etc.) are also rising since IDCs use a considerable portion of the electricity
market. Study on how to minimize the power usage expenses was performed by presenting a
supply function by researchers. Their objective was to optimize electricity costs related to IDC
with the help of non-linear programming. The Corrected Marginal Cost (CMC) algorithm has
been introduced and utilized in order to solve the optimization problem in an efficient manner.
Based on studies, the operational costs could be reduced up to 45 percent by taking advantage of
various energy price in different locations (Wang et al., 2012).
4.3. DC Operational Costs
Efforts have been taken to reduce the operational risk, improve the responsiveness and
facilitate behavior predictions in the DC infrastructure by using low-cost and low-power wireless
Sensors to monitor variables such as power, temperature, humidity, airflow, differential air
pressure, water, closures, motion, and vibration in the DC facility (Levy and Hallstrom, 2017).
The objective of their study was to gather important information to assist users to make more
informed decisions, reduce the downtime and cost of operations, improve energy efficiency,
perform energy audits, plan space usage, and develop predictive behavior models. Finally, a real-
time monitoring and management approach for the DC infrastructure was proposed. As some
future research opportunities in this field, predictive modeling can be considered to simulate and
forecast the DC’s performance and behavior in the future. Moreover, the simulation tool is useful
for validating the model and analyzing different scenarios (Levy and Hallstrom, 2017). Other
methods or models have been also used to address reducing the costs of operations in DC. An
approach is proposed (Saha et al., 2016) to optimize the use of resources in DC. According to
their study, the operational expenses in a DC, depending on the type of operations, could be
generally between 10 to 25 million dollars per year (Saha et al., 2016). The hardware and
software in a DC, both spend about 40% of DC’s operational costs. Therefore, a revenue
optimization model is developed and implemented for maximizing the revenue in a DC while
minimizing the investment cost. To achieve this objective, the Cobb-Douglas production
function is used and calculated by the researchers while considering budget constraints.
Maximizing the production, minimizing the cost and maximizing the profit for a typical DC
enterprise were modeled. The inputs were cost of servers, infrastructure investment, power costs,
and networking costs. Finally, the solutions for having the maximum revenue in a Cloud DC are
obtained (Saha et al., 2016). Different energy scenarios and power management strategies are
analyzed based on the number of servers in the DC (Janacek & Nebel, 2017a; Postema &
Haverkort, 2018). To evaluate and test the quality of power management strategies in DCs,
important parameters including stability, efficiency, adaptability, and robustness have been
measured for each power management strategy. Regarding the desire quality metric, the
optimized strategy may be different. Thus, the selected power strategy depends on which quality
parameter is the most important one for a typical DC (Postema & Haverkort, 2018).
4.4. Information Technology Infrastructure Library (ITIL) method
To explore and improve the DC operations in a cloud computing environment, the
Information Technology Infrastructure Library (ITIL) method was introduced by scholars (Bai
and Geng, 2015). ITIL is defined as a set of IT service management standard libraries and based
on a process. It is applied to manage the IT infrastructure process. Having efficient, safe and
stable operations in DCs, forming an operation management practice under the cloud computing
environment were the objectives of their research work. In addition to the ITIL, the advantages
of automation and maintenance tools in DC operations have been elaborated. These tools are
helpful in making a unified management system consisting of physical and virtual resources. By
applying automation tools, the operation and maintenance (O&M) management could become
effective, appropriate, smart, united and integrated to be used in a Cloud DC (Bai and Geng,
2015).
4.5. DC Reliability and Failure Analysis
Regarding the reliability analysis, a study on Modular Data Centers (MDCs), as a relatively
new type of DCs is performed recently. MDCs are defined as the DCs which have improved
design and high resiliency, especially in remote regions. They could extend the Information and
Communication Technology (ICT) to places that have limitations in terms of maintenance
support. MDCs are made up of independent and self-directed modules. They are constructed in
standard scales so that they could be transported more easily in containers (Farrahi Moghaddam
et al., 2017). Different components of an MDC, such as serves, switch and links are considered
in terms of their failure during operation. For this analysis, the Monte-Carlo approach has been
used to investigate each components’ performance during its lifespan. The scholars have selected
three main network topologies of the MDC and evaluated their performance and resiliency
(Farrahi Moghaddam et al., 2017).
Continuity of the service is a key aspect of any system. Particularly, in a DC, proper facility
management is an essential issue for improving energy efficiency and reducing operational costs.
A benchmark reliability analysis identifying failure modes of key components in a DC is needed
which can estimate the distribution of them over their service life. In order to design DC facilities
with excellent reliability, minimizing DC’s downtime will save organizations large amounts of
money in the future. It should also be noted that DC carries the information system related to the
host network and stores them in devices. Thus, for having the best DC performance, managing
and maintaining its system a key parameter Online services use the data stored in DCs, which
should be accessible and available instantly. The quantity of stored data in DCs is increasing (as
mentioned in the introduction) because of using Infrastructure as a Service (IaaS) applications
(computers and virtual machines used by remote users which are using services such as Amazon
Web Services) and social media (Janacek and Nebel, 2017). Recently, practices have been taken
place in DCs to assure successful data recovery. Moreover, critical applications should be able to
restore for enabling the continuity of operations when a disruption occurs in DC services (Gómez
et al., 2017; Bai & Geng, 2015a).
5. Data Center Maintenance Management
In the area of maintenance management, several studies have been performed and various
methods and models have been used, implemented and developed by scholars. In this section,
efforts have been taken to cover most of the recent studies related to this topic by investigating
the maintenance management issues of DCs and IT infrastructures. The methods and models
applied in various industry sectors will be also discussed. This approach could assist in finding
new methods to develop an optimized maintenance management schedule and framework for
DCs.
The maintenance stage in every system is very important and requires major attention.
Especially in the case of DCs, with having several critical components running all time, a proper
maintenance management system proving various maintenance jobs is needed in order to
monitor DC operations, find component failures and provide repairing strategies (Xia et al.,
2017). The importance of DC maintenance has been highlighted by some scholars. Studies on
various solutions have been performed by them to reduce energy consumption and improve the
reliability, quality, and performance of most important systems, subsystems, and components in
DCs. Design methods and techniques are implemented to overcome main issues by improving
cooling systems and technologies and managing the servers and other key components while
minimizing operations and maintenance costs. Due to the conducted research and estimations,
near 80% of total costs in a DC are related to the maintenance expenses (Saha et al., 2016).
A brief study on the economic evaluation of effective Computer Aided Facility Management
(CAFM) software implementation in manufacturing companies was performed. The focus of the
study was on the manufacturing part of a mechanical company in the Czech Republic. The goal
was analyzing machinery maintenance as a critical element of FM in the industry (Poór,
Kuchtová and Šimon, 2014). The human performance during maintenance operations is an
important issue which has its own impacts. This factor has been presented and focused on by
researchers. They have presented a data collection and analysis procedures for maintenance
operations in Ships and marine systems. Information was gathered from Engine Departments
(ED) and Deck Departments (DD) through a structured questionnaire. However, the study is
limited to process marine systems and no specific maintenance plan, schedule or system is
proposed and developed (Islam et al., 2018)
 The Reliability-Centered Maintenance (RCM) optimization method, as a popular maintenance
management model, has been used by researchers, especially for power distribution units
(Yssaad, Khiat and Chaker, 2014; Piasson et al., 2016). RCM is used to perform preventive
maintenance in a system. Introducing and implementing an economic maintenance program for
electric power distribution systems is performed by studying an electrical distribution station
located in Algeria, as a case study. The objective of the research was to reduce the maintenance
cost by focusing on the system’s most important functions and ensuring safety by performing
preventive maintenance actions and maintain the functionality at a low cost (Yssaad, Khiat and
Chaker, 2014). In another case study, an Electric Power Distribution System (EPDS) located in
Brazil was selected and a complex mathematic model was developed and used focusing on
scheduled maintenance actions and the unreliability index caused by the maintenance programs.
The model minimized preventive maintenance costs while improving the reliability index for
more than 700 components and 3 feeders operating in this EPDS. (Piasson et al., 2016). Another
study has also implemented the RCM strategy in 2016 by using the maintenance history data of a
process plant manufacturing titanium dioxide. Reaching high availability of production
equipment and reducing maintenance costs were the main objectives (Vishnu and Regikumar,
2016).
The Availability-Based Maintenance (ABM) strategy has been recently applied in a domestic
hot water subsystem (DHW) in the building’s HVAC system. The reason for using this
framework was to Reduce maintenance costs while improving the system’s reliability. Also, with
the assistance of the Keeping System Available (KSA) method, optimization of maintenance
scheduling has been performed regarding the system’s reliability. However, the impact of other
types of subsystems (such as parallel and standby) on the maintenance scheduling and the
system’s availability was not discovered. Moreover, the maintenance cost was not minimized in
their study (Pourhosseini and Nasiri, 2018). It should be noted that the KSA method has been
previously presented and applied to optimize the maintenance programs for critical elements in
thermal power stations. The researchers found that by using this method, a more appropriate
maintenance schedule could be obtained and thus, the failures are reduced and the maintenance
expenses are decreased (Zhang, Nakamura and Hatazaki, 2002). Moreover, scholars have studied
further for improving the method based on reliability for increasing total system availability
(Zhang and Nakamura, 2005).
6. Research Gap Analysis, Limitations and Future Research
Based on the reviewed publications and research works, which were discussed in the previous
sections, it is observed that a considerable number of works have been done in the fields of
system operations, maintenance, reliability and availability along with proposing new
frameworks, methods and models to improve the performance of several systems in the industry.
This paper’s objective was to find and discover studies which have focused on DCs. Only a few
research works were found in this area of application. In this section, all publications have been
organized in the following table (Table 2) based on their topic of research, methodology and
their area of application.
 2014)
2012)
2011)
2009)
2005)
(Zhang,

Nakamura,
(Zhang and
Author(s)/

(Wang et al.,
(Wang et al.,

and Williams,
(Guenter, Jain

(Irizarry et al.,
Nakamura and
Developer(s)

Hatazaki, 2002)
Project and Quality Management



Operational Cost

Operations Efficiency

Cooling Systems and Equipment



Energy and Power Usage Minimization

System Performance
Topics of Research





Maintenance Management, Planning and Scheduling


System Monitoring and Management
Operations & Maintenance Management



Maintenance Cost




Reliability and Failure Analysis

Reliability-Centered Maintenance (RCM)



Keeping System Available (KSA)

Analytic Hierarchy Process (AHP)

Survey - Interview

Service Quality Assessment

Methodology

Monte-Carlo Approach





System Optimization



Simulation Methods

Information Technology Infrastructure Library (ITIL)

Table 2: Taxonomy of recent publications based on their topic, methodology and case study

Big Data / Data Collection





Data Centers



Power Plants (Units) / Related Components

Other Industrial Systems



Commercial Buildings
Area of Application

No Specific Case Study

(Poór, Kuchtová
and Šimon,       
2014)
(Wiboonrat and
Kaewsiri, 2014)
      
(Yssaad, Khiat
and Chaker,    
2014)
(Zhan and
Zhang, 2014)
   
(Bai and Geng,
2015)
   

(Cho et al.,
2015)
    

(Hassan et al.,
2015)
     

(Hubbell et al.,
2015)
    

(Khan and
Zomaya, 2015)
   

(Rani et al.,
2015)
    

(Bayati, 2016)     

(Cho and Kim,

2016)
    

(Kadir et al.,
2016)
   

(Li et al., 2016)     

(Piasson et al.,
2016)
   
 (Saha et al.,
2016)
     

(Scheffer et al.,
2016)
  
(Vishnu and
Regikumar,      
2016)
(Wiboonrat,
2016)
   
(Farrahi
Moghaddam et   
al., 2017)
(Levy and
Hallstrom, 2017)
    

(Pun et al., 2017)       

(Tang, Cao and

Shao, 2017)
    

(Beauregard and
Ayer, 2018)
      

(Islam et al.,
2018)
  
(McArthur and
Bortoluzzi,    
2018)
(Postema and
Haverkort, 2018)
   

(Pourhosseini
and Nasiri, 2018)
     

(Shin et al.,
2018)
    
(Braglia,
Castellano and    
Gallo, 2019)
(Conejos, Chew
and Azril, 2019)
   

(Erozan, 2019)   

(Gómez et al.,
2019)
  

(Halmetoja,
2019)
   

(Hameed et al.,
2019)
   
(Kian, Bektas
and Ouelhadj,   
2019)
(Mollahassani-
Pour,
Rashidinejad and   
Pourakbari-
Kasmaei, 2019)
(Sayyah et al.,
2019)
    

(Sheng and
Prescott, 2019)
  

(Wang et al.,
2019)
  

(Zhong et al.,
2019)
   
 In the next step, based on the literature discussed in the previous sections, significant
challenges and issues which have been addressed by researchers, are presented. However, some
important topics about O&M management in DCs have not yet studied by scholars. These topics
are summarized and discussed as future research developments in this section. Hence, remaining
research works in this field and the most important research gaps and limitations are obtained.
According to the literature, authors have presented valuable methods and models for DC
operations monitoring and proposed a monitoring system to improve the reliability and reduce
the downtime and cost of operations (Hubbell et al., 2015; Levy and Hallstrom, 2017). In
addition, in the design and planning stage of DCs, important studied have been conducted about
the Data Center Project (DCP) and developing a framework for DC Systems Engineering (SE) to
improve customer satisfaction (Wiboonrat and Kaewsiri, 2014; Wiboonrat, 2016). Their studies
could be extended by adding a complete integrated reliability and availability analysis for DC
and calculation of availability and maintenance cost for each component in the DC infrastructure.
In terms of DC operations and maintenance, scholars have conducted research on operational
issues of Internet Data Centers (IDCs) and proposed a supply function to model the market
power of IDC and minimize the electricity cost (Wang et al., 2012). They also applied the
Information Technology Infrastructure Library (ITIL) V3 as an operations management strategy
to explore and improve the DC operations in a cloud computing environment. O&M services for
cloud computing DC are also utilized to provide an overall, comprehensive, standard and
automated monitoring (Zhan and Zhang, 2014; Bai and Geng, 2015). Further topics such as
failure analysis, proposing an optimized maintenance plan and model and validating the results
in real DC case studies are highly recommended.
In the area of system reliability, a study has taken valuable effort to analyze the resiliency for
Modular Data Centers (MDCs) by identifying the critical components and calculating their
failures. The performance of the component’s design has been evaluated (Farrahi Moghaddam et
al., 2017). This research procedure can be applied in general and traditional DCs by adding
system availability analysis and obtaining an efficient maintenance schedule.
In other industrial systems and case studies such as power plants (Yssaad, Khiat and Chaker,
2014; Yssaad and Abene, 2015; Vishnu and Regikumar, 2016) and HVAC systems
(Pourhosseini and Nasiri, 2018), several studies have been done to achieve system’s availability,
obtain a maintenance schedule, reduce maintenance costs and improve the reliability. These
methods and models could be also applied to DCs in future research to obtain an appropriate
maintenance scheduling model.
According to the gaps mentioned above, future research directions could be classified into two
main categories from theoretical and practical perspectives:
6.1. Theoretical perspectives
From the theoretical point of view, it should be noted that there are different levels of
availability requirements for DCs provided by the Telecommunications Industry Association
(TIA) and Uptime Institute standards (ADC Telecommunications Inc., 2006; Gabriel, 2014;
Institute, 2014). Based on these requirements, The DC infrastructure is categorized into four
different tiers and levels. Tier 1 DC has a basic level of availability (99.671%), while Tier 4 DC
has the maximum availability level (99.995%). These levels of availability and downtime,
express the importance of DC infrastructures and the need for an availability-based maintenance
scheduling model. Most of the current maintenance management and scheduling models have
considered cost minimization while the risk parameter has not been fully implemented. In the
case of DCs, a maintenance management optimization model has to be implemented integrating
cost and availability factors regarding DC’s availability requirements. We are also in need of
models to incorporate organizational (such as change management capabilities) and external
risks (such as extreme weather conditions) into failure prediction and reliability analysis of data
center infrastructures, which has been widely studied in information technology domain
(Mafakheri et al., 2012) but is missing in case of data centers facility management research.
6.2. Practical perspectives
In future research, the maintenance management models for DCs need specific management
practices and benchmarks, which are not much available in the current literature. Because of the
nature of assets available in DCs, there are several types and categories of assets in a DC which
should run 24/7. Thus, to perform best maintenance management practices for DCs, companies
and their facility managers need to share their experiences in this matter. Specific benchmarks,
codes, and handbooks are recommended to be developed and customized for DCs. to obtain and
implement a specialized availability-based optimized maintenance scheduling model for the DC
industry. Also, the condition monitoring process is highly important when dealing with various
types of mechanical, electrical, networking, etc. assets in DCs.
Although this paper has reported several methods, techniques, models, and frameworks
regarding operations and maintenance management systems in different cases, specifically in
DCs, there are still remaining gaps and limitations in this area that should be concerned. In both
the academic and industrial sectors, the attention to the DC infrastructure and facility is
increasing in terms of obtaining reliable and effective solutions for optimizing DC’s operations.
In addition, these solutions should assist maintenance managers to reduce maintenance costs. So
far studies have been conducted mostly on the maintenance monitoring for DCs and new
technologies in this area. Also, several publications have investigated the solution for minimizing
power and energy consumption of DCs. Although each of these research works has reported
valuable information and solutions, new studies are required to be performed for addressing an
optimized maintenance schedule and management system. Similar works are done in other
industrial systems but in the case of DCs, research works have remained limited and some
important issued are not considered.
Regarding the issues discussed, a comprehensive framework is required to analyze the failure
rates and modes for the components in a DC and then calculate the availability of the
components while considering the cost of maintenance tasks for each component. With this
approach, the subsystems and components could be organized based on their maintenance
schedule, the parts which need preventive maintenance are defined and the issue of system’s over
maintenance is avoided.
7. Conclusions
In this paper, efforts have been made to represent a critical literature review on the topic of
operations and management. The objective was to discover new publications in this area and
especially for DCs. In the previous section, a summary of the literature is reported to clarify the
research topics, methodologies and case studies to obtain the remaining unstudied works, which
are the research shortcomings. As represented, the literature in the field of DC maintenance is not
rich and thus, more studies have to be performed in the future.
An extensive reliability, failure and availability analysis, integrated with maintenance
scheduling, must be implemented for DCs. Moreover, the maintenance plan could be optimized
by applying various optimization techniques to decrease the costs of maintenance tasks while
addressing the system’s availability requirements. The DC facility and infrastructure with all
systems, sub-systems, and components should be carefully considered in future research works.
Identified gaps and limitations in the reviewed studies express the essential need for more
efforts in conducting research in the field of maintenance planning, scheduling, and management
for DCs to reduce the maintenance costs while improving the performance and reliability of the
DC system. As DCs have been dealing with several practical issues such as high maintenance
costs, this paper provides a structured way of classifying the literature with respect to these
issues. In addition, several of the reviewed papers were based on real case studies. Therefore, the
outcomes of this paper could have many benefits for practitioners.
 8. References
1. Abbas, C. J. B., Orozco, A. L. S. and Villalba, L. J. G. (2015) ‘Monitoring of Data Centers
using Wireless Sensor Networks’, in Handbook on Data Centers. New York: Springer, pp.
1171–1183.
2. ADC Telecommunications Inc. (2006) ‘Tia-942 Data Centre Standards Overview’, p. 8.
Available at: www.adc.con.
3. Bai, W. and Geng, W. (2015) ‘Research on Operation Management under the Environment of
Cloud Computing Data Center’, International Journal of Database Theory and Application,
8(2), pp. 185–192. doi: http://dx.doi.org/10.14257/ijdta.2015.8.2.17.
4. Bayati, M. (2016) ‘Managing Energy Consumption and Quality of Service in Data Centers’,
in 5th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS).
Rome, pp. 1–9. Available at:
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7951363&isnumber=7951246.
5. Beauregard, M. A. and Ayer, S. K. (2018) ‘Maintaining performance: Understanding the
relationship between facility management and academic performance at K-12 schools in the
State of Arizona’, Facilities, 36(11–12), pp. 618–634. doi: 10.1108/F-11-2017-0111.
6. Bieser, J. and Menzel, K. (2019) ‘Assessing Facility Maintenance Models for Data Centres:
Status and Deficits of Current Facility Management and Maintenance Concepts’, Applied
Mechanics and Materials, 887, pp. 255–263. doi:
https://doi.org/10.4028/www.scientific.net/AMM.887.255.
7. Braglia, M., Castellano, D. and Gallo, M. M. (2019) ‘A novel operational approach to
equipment maintenance: TPM and RCM jointly at work’, Journal of Quality in Maintenance
Engineering. Emerald (J. Qual. Maint. Eng. (UK)), 25(4), pp. 612–634. doi: 10.1108/JQME-
05-2016-0018.
8. Carey, C., Raisinghani, M. S. and White, B. (2017) ‘Foundations of Data Center: Key
Concepts and Taxonomies’, in Engineering and Management of Data Centers. Springer, pp.
1–13.
9. Che-Ghani, N. Z., Myeda, N. E. and Ali, A. S. (2016) ‘Operation and Maintenance Cost for
Stratified Buildings : a Critical Review’, in MATEC Web of Conferences IBCC 2016, pp. 1–
9. doi: 10.1051/matecconf/20166600041.
10. Chew, M. Y. L., Conejos, S. and Asmone, A. S. (2017) ‘Developing a research framework for
the green maintainability of buildings’, Facilities, 35(1–2), pp. 39–63. doi: 10.1108/F-08-
2015-0059.
11. Cho, J. et al. (2015) ‘Development of an energy evaluation and design tool for dedicated
cooling systems of data centers: Sensing data center cooling energy efficiency’, Energy and
Buildings. Elsevier B.V., 96, pp. 357–372. doi: 10.1016/j.enbuild.2015.03.040.
12. Cho, J. and Kim, Y. (2016) ‘Improving energy efficiency of dedicated cooling system and its
contribution towards meeting an energy-optimized data center’, Applied Energy. doi:
10.1016/j.apenergy.2015.12.099.
13. Cisco Systems (2018a) Data center storage capacity worldwide from 2016 to 2021, by
segment (in exabytes), Statista. Available at:
https://www.statista.com/statistics/638593/worldwide-data-center-storage-capacity-cloud-vs-
traditional/.
14. Cisco Systems (2018b) Volume of big data in data center storage worldwide from 2015 to
2021 (in exabytes), Statista. Available at:
https://www.statista.com/statistics/638621/worldwide-data-center-storage-used-by-big-data/.
15. Conejos, S., Chew, M. Y. L. and Azril, F. H. Bin (2019) ‘Green maintainability assessment of
high-rise vertical greenery systems’, Facilities, 37(13–14), pp. 1008–1047. doi: 10.1108/F-09-
2018-0107.
16. Diouri, M. E. M. et al. (2015) ‘Providing green services in hpc data centers: A methodology
based on energy estimation’, in Handbook on Data Centers. New York: Springer, pp. 287–
323.
17. Elsevier (2019) Engineering Village.
18. Erozan, hsan (2019) ‘A fuzzy decision support system for managing maintenance activities of
critical components in manufacturing systems’, Journal of Manufacturing Systems. Elsevier
B.V., 52, pp. 110–120. doi: 10.1016/j.jmsy.2019.06.002.
19. Farrahi Moghaddam, R. et al. (2017) ‘A Monte-Carlo approach to lifespan failure
performance analysis of the network fabric in modular data centers’, Journal of Network and
Computer Applications, 87, pp. 131–146. doi: 10.1016/j.jnca.2017.03.015.
20. Fraser, K., Hvolby, H.-H. and Tseng, T.-L. (Bill) (2015) ‘Maintenance management models: a
study of the published literature to identify empirical evidence’, International Journal of
Quality & Reliability Management, 32(6), pp. 635–664. doi:
http://dx.doi.org/10.1108/02683940010305270.
21. Gabriel, C. (2014) ‘Data Center Disaster Recovery and High Availability’, in Data Center
Handbook. John Wiley & Sons, Inc Hoboken, NJ, pp. 639–657.
22. Geng, H. (2015) Data Center Handbook, Data Center Handbook. Hoboken, New Jersey: John
Wiley & Sons, Inc. doi: 10.1002/9781118937563.
23. Gómez, J. F. et al. (2019) ‘Risk-based criticality for network utilities asset management’,
IEEE Transactions on Network and Service Management, 16(2), pp. 755–768. doi:
10.1109/TNSM.2019.2903985.
24. Guenter, B., Jain, N. and Williams, C. (2011) ‘Managing cost, performance, and reliability
tradeoffs for energy-aware server provisioning’, Proceedings - IEEE INFOCOM, pp. 1332–
1340. doi: 10.1109/INFCOM.2011.5934917.
25. Halmetoja, E. (2019) ‘The conditions data model supporting building information models in
facility management’, Facilities, 37(7–8), pp. 484–501. doi: 10.1108/F-11-2017-0112.
26. Hameed, A. et al. (2019) ‘A decision support tool for bi-objective risk-based maintenance
scheduling of an LNG gas sweetening unit’, Journal of Quality in Maintenance Engineering.
Memorial University, St John, CanadaSultan Qaboos University, Muscat, Oman: Emerald (J.
Qual. Maint. Eng. (UK)), 25(1), pp. 65–89. doi: 10.1108/JQME-04-2017-0027.
27. Hassan, S. F. et al. (2015) ‘Free Cooling Investigation of SEECS Data Center’, Energy
Procedia. Elsevier B.V., 75, pp. 1406–1412. doi: 10.1016/j.egypro.2015.07.233.
28. Hong, T. et al. (2015) ‘A review on sustainable construction management strategies for
monitoring, diagnosing, and retrofitting the building’s dynamic energy performance: Focused
on the operation and maintenance phase’, Applied Energy. Elsevier Ltd, 155, pp. 671–707.
doi: 10.1016/j.apenergy.2015.06.043.
29. Hubbell, M. et al. (2015) ‘Big Data strategies for Data Center Infrastructure management
using a 3D gaming platform’, 2015 IEEE High Performance Extreme Computing Conference,
HPEC 2015. doi: 10.1109/HPEC.2015.7322471.
30. Institute, U. (2014) Data Center Site Infrastructure Tier Standard : Operational Sustainability,
Uptime Institute, Llc.
31. Irizarry, J. et al. (2014) ‘Ambient intelligence environments for accessing building
information: A healthcare facility management scenario’, Facilities, 32(3), pp. 120–138. doi:
10.1108/F-05-2012-0034.
32. Islam, R. et al. (2018) ‘Human error assessment during maintenance operations of marine
systems – What are the effective environmental factors?’, Safety Science. Elsevier, 107(April
2017), pp. 85–98. doi: 10.1016/j.ssci.2018.04.011.
33. Janacek, S. and Nebel, W. (2017) ‘Engineering and Management of Data Centers’, in Gómez,
J. M. et al. (eds). Cham, Switzerland: Springer International Publishing AG 2017, pp. 163–
185. doi: 10.1007/978-3-319-65082-1.
34. Kadir, E. A. et al. (2016) ‘Wireless monitoring for big data center server room and
equipments’, Proceedings - 2015 International Conference on Science in Information
Technology: Big Data Spectrum for Future Information Economy, ICSITech 2015, pp. 187–
191. doi: 10.1109/ICSITech.2015.7407801.
35. Khan, S. U. and Zomaya, A. Y. (2015) ‘Data Center Modeling and Simulation Using
OMNeT++’, in Handbook on Data Centers. New York: Springer Science + Business Media,
pp. 1–1334. doi: 10.1007/978-1-4939-2092-1.
36. Kian, R., Bektas, T. and Ouelhadj, D. (2019) ‘Optimal spare parts management for vessel
maintenance scheduling’, Annals of Operations Research. Dept. of Ind. Eng., Univ. of Tabriz,
Tabriz, Iran: Springer (Ann. Oper. Res. (Germany)), 272(1–2), pp. 323–353. doi:
10.1007/s10479-018-2907-y.
37. Lai, J. H. K. and Man, C. S. (2018a) ‘Performance indicators for facilities operation and
maintenance (Part 1) Systematic classification and mapping’, Facilities, 36(9), pp. 476–494.
doi: http://dx.doi.org/10.1108/MRR-09-2015-0216.
38. Lai, J. H. K. and Man, C. S. (2018b) ‘Performance indicators for facilities operation and
maintenance (Part 2) Shortlisting through a focus group study Joseph’, Facilities, 36(9/10), pp.
495–509. doi: http://dx.doi.org/10.1108/MRR-09-2015-0216.
39. Levy, M. and Hallstrom, J. O. (2017) ‘A New Approach to Data Center Infrastructure
Monitoring and Management (DCIMM)’, in Computing and Communication Workshop and
Conference (CCWC), 2017 IEEE 7th Annual, pp. 1–6.
40. Li, L. et al. (2016) ‘Data center power minimization with placement optimization of liquid-
cooled servers and free air cooling’, Sustainable Computing: Informatics and Systems, 11, pp.
3–15. doi: 10.1016/j.suscom.2016.02.001.
41. Mafakheri, F., Breton, M., & Chauhan, S. (2012) 'Project-to-organization matching: An
Integrated risk assessment approach', International Journal of Information Technology Project
Management (IJITPM), 3(3), 45-59. doi: 10.4018/jitpm.2012070104.
42. Mangano, G. and de Marco, A. (2014) ‘The role of maintenance and facility management in
logistics: A literature review’, Facilities, 32(5), pp. 241–255. doi: 10.1108/F-08-2012-0065.
43. Márquez, A. C. (2007) The Maintenance Management Framework: Models and Methods for
Complex Systems Maintenance, The Maintenance Management Framework. doi:
10.1007/978-1-84628-821-0.
44. McArthur, J. J. and Bortoluzzi, B. (2018) ‘Lean-Agile FM-BIM: a demonstrated approach’,
Facilities, 36(13–14), pp. 676–695. doi: 10.1108/F-04-2017-0045.
45. Mollahassani-Pour, M., Rashidinejad, M. and Pourakbari-Kasmaei, M. (2019)
‘Environmentally constrained reliability-based generation maintenance scheduling
considering demand-side management’, IET Generation, Transmission and Distribution,
13(7), pp. 1153–1163. doi: 10.1049/iet-gtd.2018.5713.
46. Phogat, S., Gupta, Kumar, A. and Gupta, A. K. (2017) ‘Identification of problems in
maintenance operations and comparison with manufacturing operations A review’, Journal of
Quality in Maintenance Engineering, 23(2), pp. 226–238. doi: 10.1002/2016WR019831.
47. Piasson, D. et al. (2016) ‘A new approach for reliability-centered maintenance programs in
electric power distribution systems based on a multiobjective genetic algorithm’, Electric
Power Systems Research. Elsevier B.V., 137, pp. 41–50. doi: 10.1016/j.epsr.2016.03.040.
48. Poór, P., Kuchtová, N. and Šimon, M. (2014) ‘Machinery maintenance as part of facility
management’, Procedia Engineering. Elsevier B.V., 69, pp. 1276–1280. doi:
10.1016/j.proeng.2014.03.119.
49. Postema, B. F. and Haverkort, B. R. (2018) ‘Evaluation of Advanced Data Centre Power
Management Strategies’, Electronic Notes in Theoretical Computer Science, 337, pp. 173–
191. doi: 10.1016/j.entcs.2018.03.040.
50. Pourhosseini, O. and Nasiri, F. (2018) ‘Availability-Based Reliability-Centered Maintenance
Scheduling: Case Study of Domestic (Building-Integrated) Hot Water Systems’, ASCE-
 ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering,
4(1), pp. 1–13. doi: 10.1061/AJRUA6.0000935.
51. Pun, K. P. et al. (2017) ‘A fuzzy-AHP-based decision support system for maintenance
strategy selection in facility management’, PICMET 2017 - Portland International Conference
on Management of Engineering and Technology: Technology Management for the
Interconnected World, Proceedings, 2017-Janua, pp. 1–7. doi:
10.23919/PICMET.2017.8125300.
52. Rani, N. A. A. et al. (2015) ‘Perception of Maintenance Management Strategy on Healthcare
Facilities’, Procedia - Social and Behavioral Sciences. Elsevier B.V., 170, pp. 272–281. doi:
10.1016/j.sbspro.2015.01.037.
53. Saha, S. et al. (2016) ‘A novel revenue optimization model to address the operation and
maintenance cost of a data center’, Journal of Cloud Computing. Journal of Cloud Computing,
5(1), pp. 1–23. doi: 10.1186/s13677-015-0050-8.
54. Sayyah, A. et al. (2019) ‘A multi-objective modelling approach to assess the impact of
maintenance scheduling on equipment procurement methods’, International Journal of
Industrial and Systems Engineering. Faculty of Industrial and Systems Engineering, Tarbiat
Modares University, Tehran, Iran. Management and Technology Development Center, Tarbiat
Modares University, Tehran, Iran: Inderscience Publishers (IEL), 31(2), pp. 184–208. doi:
10.1504/IJISE.2019.097737.
55. Scheffer, M. et al. (2016) ‘SIMULATION OF MAINTENANCE STRATEGIES IN
MECHANIZED TUNNELING’, in Proceedings of the 2016 Winter Simulation Conference,
pp. 3345–3356.
56. Shehabi, A. et al. (2016) United States Data Center Energy Usage Report, Lawrence Berkeley
National Laboratory. Berkeley, California.
57. Sheng, J. and Prescott, D. (2019) ‘A coloured Petri net framework for modelling aircraft fleet
maintenance’, Reliability Engineering and System Safety. 93582 Troops, Chinese Air Force,
Shanxi; 038300, ChinaResilience Engineering Research Group, University of Nottingham,
Nottingham; NG7 2RD, United Kingdom: Elsevier Ltd, 189, pp. 67–88. doi:
10.1016/j.ress.2019.04.004.
58. Shin, H. et al. (2018) ‘Facility management process of an office building’, Journal of
Infrastructure Systems, 24(3). doi: 10.1061/(ASCE)IS.1943-555X.0000436.
59. Statista (2017) Number of data centers worldwide in 2015, 2017, and 2021 (in millions),
Statista. Available at: https://www.statista.com/statistics/500458/worldwide-datacenter-and-it-
sites/ (Accessed: 23 December 2018).
60. Tang, H., Cao, J. and Shao, Z. (2017) ‘Network Simulation and Vulnerability Analysis on
Organization of Facility Management’, in 2017 IEEE International Conference on Systems,
Man, and Cybernetics (SMC) Banff Center, Banff, Canada. Banff, Canada. doi:
10.5771/9783845255507_437.
61. Vishnu, C. R. and Regikumar, V. (2016) ‘Reliability Based Maintenance Strategy Selection in
Process Plants: A Case Study’, Procedia Technology. The Author(s), 25(Raerest), pp. 1080–
1087. doi: 10.1016/j.protcy.2016.08.211.
62. Wang, L. et al. (2009) ‘Towards Thermal Aware Workload Scheduling in a Data Center’, in
IEEE 10th International Symposium on Pervasive Systems, Algorithms, and Networks
(ISPAN), pp. 116–122.
63. Wang, L. et al. (2019) ‘An Extended FMECA Method and Its Fuzzy Assessment Model for
Equipment Maintenance Management Optimization’, Journal of Failure Analysis and
Prevention. School of Mechanical and Electrical Engineering, China Jiliang University,
Hangzhou; 310018, ChinaShanghai Rail Transit Maintenance Support Centre, Shanghai;
200070, China: Springer New York LLC, 19(2), pp. 350–360. doi: 10.1007/s11668-019-
00611-3.
 64. Wang, P. et al. (2012) ‘D-Pro: Dynamic data center operations with demand-responsive
electricity prices in smart grid’, IEEE Transactions on Smart Grid, 3(4), pp. 1743–1754. doi:
10.1109/TSG.2012.2211386.
65. Wiboonrat, M. (2016) ‘Applying ISO 26702 and IEEE 1490 for data center project
management’, 2016 IEEE International Conference on Management of Innovation and
Technology, ICMIT 2016, pp. 298–303. doi: 10.1109/ICMIT.2016.7605051.
66. Wiboonrat, M. and Kaewsiri, U. (2014) ‘A Chronological Transformation of Data Center
Project Management’, in 2014 World Automation Congress (WAC), pp. 1–6. doi:
10.1109/WAC.2014.6935754.
67. Williams, J. (2018) State of the Data Center Industry. Olympia, WA, USA. Available at:
http://www.commerce.wa.gov/wp-content/uploads/2018/01/Commerce-Data-Center-Study-
and-appendices-2017.pdf.
68. Wu, C. and Guo, J. (2015) ‘Software monitoring in data centers’, in Handbook on Data
Centers. New York: Springer, pp. 1209–1253. doi: 10.1007/978-1-4939-2092-1_42.
69. Xia, W. et al. (2017) ‘A Survey on Data Center Networking (DCN)- Infrastructure and
Operations.pdf’, IEEE Communications Surveys and Tutorials, 19(1), pp. 640–656. doi:
10.1109/COMST.2016.2626784.
70. Yousefli, Z., Nasiri, F. and Moselhi, O. (2017) ‘Healthcare facilities maintenance
management: a literature review’, Journal of Facilities Management, 15(4), pp. 352–375. doi:
10.1108/JFM-10-2016-0040.
71. Yssaad, B. and Abene, A. (2015) ‘Rational Reliability Centered Maintenance Optimization
for power distribution systems’, International Journal of Electrical Power and Energy
Systems. Elsevier Ltd, 73, pp. 350–360. doi: 10.1016/j.ijepes.2015.05.015.
72. Yssaad, B., Khiat, M. and Chaker, A. (2014) ‘Reliability centered maintenance optimization
for power distribution systems’, International Journal of Electrical Power and Energy
Systems. Elsevier Ltd, 55, pp. 108–115. doi: 10.1016/j.ijepes.2013.08.025.
73. Zhan, H. and Zhang, W. (2014) ‘The Operation and Maintenance Management System of the
Cloud Computing Data Center Based on ITIL’, Advances in Computer Science and Its
Applications, 421, pp. 1103–1108. doi: 10.1007/978-3-642-41674-3_153.
74. Zhang, T. and Nakamura, M. (2005) ‘Reliability‐based Optimal M aintenance Scheduling by
Considering Maintenance Effect to Reduce Cost’, Quality and Reliability Engineering
International. Wiley Online Library, 21(2), pp. 203–220.
75. Zhang, T., Nakamura, M. and Hatazaki, H. (2002) ‘Optimizing maintenance scheduling of
equipment by element maintenance interval adjustment considering system availability’, in
Power Engineering Society Winter Meeting, 2002. IEEE. IEEE, pp. 205–210.
76. Zhong, S. et al. (2019) ‘A reliability-and-cost-based fuzzy approach to optimize preventive
maintenance scheduling for offshore wind farms’, Mechanical Systems and Signal Processing.
Elsevier Ltd, 124, pp. 643–663. doi: 10.1016/j.ymssp.2019.02.012.

View publication stats