Design of Data Centre for UWTSD

This report provides insight into the design of a data centre for UWTSD, including considerations for fault tolerance, scalability, and security. Recommendations for future improvements are also provided.

Data Centre Technologies

Abstract
Data centre acts as integral part within the enterprise that is designed for supporting
business application and render wide range of services. This comprises of data storage, backup &
recovery, management, e-commerce transactions, online gaming communities, artificial
intelligence and many more that leads to creation on affirmative impact on the ways in which
services are being delivered by the organisation. It is important that all the associated factors are
being considered by the organisation so that they can deal with different aspects such as load
balancing, security, management, RPO & RTO and many more. The report will furnish insight
into different aspects associated with design of the data centre and what all points organisation
must into consideration while developing their DC. In addition to this, future recommendations
will be given with respect to data centre.

Table of Contents
Abstract............................................................................................................................................2
Table of Contents.............................................................................................................................2
Introduction......................................................................................................................................1
Overview of data centre...............................................................................................................1
Design of data centre...................................................................................................................2
Points to be considered................................................................................................................8
Assets and limitations associated with data centre....................................................................11
Analysis.....................................................................................................................................12
Conclusion.....................................................................................................................................13
Recommendations..........................................................................................................................13
References......................................................................................................................................15

Introduction
Data centre refers to facility that is liable for centralisation of shared IT operations along
with equipments for rationale of storage, processing as well as dissemination of application and
data. It implies the physical facilities that are being possessed by organisations to house their
critical data and applications. The design is entirely dependent on storage resources as well as
network of computing that is liable for delivering shared information (Acton and et. al, 2017).
The broader term that is information technology which denotes communication technologies that
comprises of wireless network, middleware, social networking, cell phones, video conferencing
and other media services that allow users to access, retrieval, storage, transmission along with
manipulation of information within the digital form is defined as ICT(information and
communication technology). Basically, ICT are components and infrastructure that allow modern
computing. This report is based on UWTSD, that is formed by the merger of SMU and TSD.
They are looking forward to centralise their IT operations along with the equipments for
managing ICT infrastructure. In addition to this, the report will provide an insight into designing
of data centre along with various aspects related with this like fault tolerant, scalability, security
and many more.
Overview of data centre
Data centre is accountable for centralising the operations as well as equipments of the
organisation for ensuring that there is a secured location for storage, sharing of information along
with management of ample of data (Alishahi and et. al, 2019). This will further enable UWTSD
to have flexibility within the ways in which they create back up, store and protect their critical
assets from any kind of damage either that is natural or man-made. There are different kinds of
data centres that can be utilised by businesses for building up or storing their data within them,
they are specified beneath:
Colocation facilities: In this case UWTSD have the option to have space on rent that is
being furnished by the data centre service provider. The infrastructure is being hosted via
building owner that comprises of services and equipments for cooling systems, power,
networking, security and bandwidth. The provider of space will be accountable for renting the
space will install as well as manage components like storage, data centre servers and firewalls.
1

Enterprise: This denotes facility that is being owned as well as operated via business for
usage. This facility will be located within the own site of organisation but basically, this will be
present at off-premises at location that will render prime security, connectivity and power.
Equipping as well as building up enterprise data centre needs relevant capital investment that
enable firms to address their peculiar requirements in an appropriate manner (Avgerinou,
Bertoldi and Castellazzi, 2017).
Hyperscale data centres:They are designed to render the high computing that is required
for big data as well as cloud storage. This aspect is accountable for providing cloud service
providers with scalable and robust application along with storage capabilities to customers like
UWTSD. In this case, high fibre networks are being utilised.
Design of data centre
With reference to UWTSD, active-active architecture design can be utilised that will
enable them within having optimised uptime. Within such kind of setup, the off-site servers have
to be synced with on-site servers that are available for enabling near-instantaneous redundancy
of information. This is being carried in order to make sure that in case if any of the servers fail
then the load must be automatically rebalanced for other servers (Dandres and et. al, 2017). This
will further enable UWTSD to provide their people or consumers with adequate services as per
their requirements and their activities will also not be hampered. In addition to this, it will allow
them to minimise the network traffic load as different servers can respond to the requests that are
being made. Active-active is liable for rendering traffic load and availability sharing of
application across the DC with business continuity, load sharing, mobility, consistency and
rendering provisioning capabilities. The designing of data centre for UWTSD is provided
beneath:
2

Technical requirements
The generic technical requirements that need to be considered while designing active-
active data centre are:
 Both DC's servicing applications: These must be available at any instance of the
time and this will enable UWTSD to:
◦ Hosting applications around data centre
◦ Have optimised traffic while entering as well as exiting DC (Eltraify and et. al,
2019).
3
Illustration 1: Architectural building blocks

 Recovery point objective and recovery time: This must be considered in order to
avoid automatic failover with restricted loss of services even when if any device fail
while working in normal conditions.
 Consistent policy management and quick service roll-out: This will allow UWTSD
to have strategies according to which it will become easy for them to deliver their
liabilities (Flucker, Tozer and Whitehead, 2018).
◦ Centralised provisioning will be enabled
◦ Network policies will be developed and there will be consistent security
 IO optimisation: This has to be enabled across the DC so that high performance can
be attained in an appropriate manner.
◦ UWTSD can have optimised usage of resources
Transport technologies
They acts like inter-connectors for the data centre and links along with device level
redundancies are part within the transport domain that is accountable for rendering resiliency and
HA across the site (Hammadi, 2016). This aspect will enable UWTSD to have redundancy for
multiplexers, DCI network devices, GPONs, diversity POPs in order to survive POP failure and
dark fibers along with 1+1 protection scheme. The major considerations for UWRSD with
reference to this aspect are specified beneath:
 Recovery from different failure scenarios such as module, link, node and many more.
 Applications and link latency round trip needs for traffic among DCs
 Scalability as well as bandwidth associated factors
Network Services
Through this UWTSD will be able to interconnect devices within data centres by carrying
out relevant routing functions and traffic switching. Here, the network is liable for facilitating
forwarding of load sharing along with application traffic without any kind of disruption (Jackson
and et. al, 2017). Furthermore, application mobility around the data centres will be rendered
through L2 extension, pervasive gateways, ingress along with egress path optimisation. In
addition to this, the asset that can be acquired by UWTSD through usage of active-active design
is that the service providers renders integrated VxLAN overlay solution for attainment of L2
4

extension, gateway mobility and path optimisation. The critical consideration while designing
network services are:
 Recovery from different kinds of failures like network, module and link failure
 Pervasive gateway around the infrastructure that comprises of availability of gateway
across DC and local to DC
 Stretching L2 domain this enables to expand L2 domain (either VxLAN or VLAN)
among Dcs.
 A consistent policy involves formulation of network policies that are consistent within
the on-premises along with different cloud infrastructure (Luo and et. al, 2019). Here, the
policies of UWTSD will comprise of naming and segmentation procedures that can be
integrated with distinct L4/L7 services, hypervisor integration and many more
 Egress and ingress path optimisation
 Centralised management or provisioning of network policies and managing (AAA
capabilities, inventory, traffic flow analysis, backup & restore, etc.)
L4-L7 services
The designing or building up of active-active L4-L7 facilities or services across DC is
expensive task as this comprises of placing ADC devices along with security within both DCs
5

(Mialle, Brown and Arora, 2019). With respect to this aspect, it becomes important for UWTSD
have to consider load balancers, global traffic management, firewalls and application policy
controllers as a solution in the space. They have to be deployed at distinct tiers for protecting
extranet, core server farm, perimeter, UAT segment, WAN, etc. Moreover, there are leading
service vendors that offer their customers with clustering solutions for their offerings so that the
issue of seamless fail-overs, traffic load and l4 or l5 policies can be handled in an effective
manner. Key consideration with service designing are:
 Recovery from failures that is L4-L7 device, link and module
 L4-L7 consistent policies across the cloud and on-premises infrastructure and rules for
distinct traffic types Centralised management
Storage Service
Storage along with networking solutions acts as pillar of designing the data centre and
this implies storage across serving applications of DC. In addition to this, the design must cater
capability for accepting reads along with requests for write without any kind of interruption
(Rashid and Noraziah, 2017). Thus, this will enable to have real-time mirroring as well as
seamless capabilities for preventing failures across DCs. The design storage considerations for
UWTSD are:
 Prevention from storage failures like storage array, single disk, storage controller failure
and split brain scenarios
 Synchronous replication, data needs to be written within the primary storage along with
simultaneous replica. Due to this, more bandwidth will be consumed and also requires
usage of dedicated FC links
 Redundancy and high availability involves that relevant dsisk are available for carrying
out replication of the information that is present within the server or a data centre Network device failure scenarios
Server Virtualisation
With reference to this aspect the organisations are moving forward towards containers
and micro-services. With reference to this aspect, UWTSD needs to extend container or
6

hypervisor clusters across DCs for attainment of seamless virtual container or machine instances
movement along with fail-over (Rogers, 2016). The options that organisation have comprises of
Microsoft, Kubernetes, KVM and Vmware Docker. Here, the major considerations are:
 Virtualisation platform for forming cross-DC virtual host cluster
 Deployment of identical services on virtual machines within the DCs so that in case if
any one is not working in appropriate manner or is unavailable that other data centre can
handle the load within the real-time scenarios
 HA functions in order to protect VM and creation of affinity rules for preferring local
hosts within their normal operational situations (Sefidcon and Yadhav, 2016)
 Centralised management of hypervisor's and computing resources Compute node devices around data centres that have been provisioned with symmetric
configuration with relevant resources in case of failover
Applications Deployment
The infrastructure is designed for applications for functioning. This is crucial for
UWTSD to ensure that there is higher availability of apps around the data centres so that fail
over can be attained along with this location proximity can be accessed (Singh and Kumar,
2019). It is crucial for the organisation that is UWTSD to have application, database tiers and
web available at distinct data centres. Key considerations are:
 Deployment of web services in vmware machines that have various servers that will
form independent clusters per DC.
 Deploying application services on VM so that cross-DC or DC forms the cluster. It is
preferred that IP based access must be furnished Deployment of database on physical machines for forming cross-DC cluster like DB2,
Oracle and SQL with WSFC (windows server failover cluster)
The active-active data centre designs for UWTSD needs that all the architecture
components of network, L4-L7, storage, virtualisation, application and computing must work
simultaneously. A seamless operation along with availability of business applications with
respect to infrastructure failure is key factor. In addition to this, the cost associated with this
design is expensive in comparison to disaster recovery but only 20% is more and more than 35%
7

capacity will be delivered that will enable to have non-stop operations. This will further enable
UWTSD to have improvised uptime, optimised asset utilisation and performance.
Points to be considered
While developing the design of data centre by taking into consideration ICT guidelines,
it is important for respective firm to take into account certain factors that will lead to creation of
affirmative influence on the overall working (Spyropoulou and et. al, 2020). In addition to this,
there will be a significant impact on the ways in which services are being delivered by the latter.
These aspects have been illustrated below:
Design: The design furnishes geo-enabled load balancing that is a software liable for
continuously measuring time to first byte for directing traffic to the servers that performs fastest.
The active-active data centre of UWTSD will enable the organisation to balance the read load
that is present across secondaries through which fast performance applications can be delivered.
Furthermore, the load balancer will provide leverage over fail-over mechanism that is built
within the modern database for enabling application transparent failover. The database will
conduct replication that will further hold inbound requests so that any kind of errors do not occur
(Sugawara and et. al, 2019), data centre is designed in such manner that replication lag can be
monitored as well as identify probable nodes that might have been fallen behind certain
threshold.
Scalability: It is important to easily add up new nodes within the data centre clusters so
that any downtime or any kind of interruption within their services. The existent nodes within the
cluster of UWTSD will lead to automatically sync up indexes as well as plugins within the each
new member that enables them to have hassle free deployment by making sure that maximum
uptime can be attained. As the data centre will be licensed by the organisation then, it will be
easy for them to predict costs as well as scale up the entire environment without extra cost of
license for new CPU or servers.
Security: This is critical aspect of data centre that will enable UWTSD to have enhanced
services without any interruption from third person or intruder. The major concern in this case is
physical security that involves that DC must be build in such manner that all the challenges that
are created by natural disaster do not lead to creation of any kind of pessimistic impact on their
operations. Moreover, the DC security team have to keep the close watch on individuals who
enter within and there must be restricted access so that intruder does not get access to sensitive
8

information. Network security is liable for securing perimeter by installation of firewall for
cleaning up traffic at peculiar point of entrance (Wang, 2019). In addition to this, replication
aware monitoring. In context of UWTSD, they need to opt for zero trust model for inspecting as
well as monitoring internal traffic in network for detecting and mitigation of threats that have
bypassed perimeter firewall.
Fault tolerant: The active-active architecture or design of data centre acts like a disaster
recovery as well as business continuity plan for the organisations. An instance can be taken to
acknowledge this aspect like if Google might be having a single data centre then each user will
not be able to access to their requests and maximum of the time the search engine might not have
been responded as well as due to overload the engine would be out of a service or unavailable.
Thus, data centre enhances the availability of any particular service that person needs at any
instance of time. In case if one server do not respond due to any fault then the traffic will be
directed to the other server. This implies that data centre is fault tolerant.
Performance: The instance has taken in consideration in above point and it clearly
specifies that it becomes easy to have access to the services by opting for data centre (Acton and
et. al, 2017). Thus, after the merger of SMU as well as TSD, it becomes important to have a
single DC so that information can be accessed easily by employees along with their consumers.
The rationale is to amplify the overall performance of UWTSD by providing spontaneous
services.
Energy usage: It is identified that huge energy is being consumed by the data centres like
as per the statistics of US, it is identified that in the year 2017 near about 90 billion kilowatt
hours of electricity and even more than this is used.
9

The aspects that must be considered by UWTSD while making use of data centre while
delivering their liabilities is that they need to decline the cooling dependencies, conduct
experiments with temperature (by doing so cost can be declined), synchronise server load and
capacity (Alishahi and et. al, 2019). Along with this, they must also identify as well as kill
zombie servers, optimise the space and have strong supplier partnership.
Management: Data centre management is accountable for making sure reliability of
connections along with mission critical information that is present in the storage provided by the
data centre. This is also accountable for effectively placing workloads of information within cost
efficient compute resources that are available with UWTSD.
RPO & RTO: The time that is being taken up by the application to come back into
normal state after failure is referred to as RTO (recovery time objective). RPO (recovery point
objectives) implies amount of data that is lost before the relevant harm that occurs to the
business. It is important that UWTSD makes use of significant backup technologies for avoiding
data loss. In order to furnish support for active-active DC, it is crucial for avoiding traffic hair-
10
Illustration 3: Power Utilisation

pinning along with furnish write/read capabilities for storage. This will lead to decline RTO as
storage will be leveraged via application.
Assets and limitations associated with data centre
The different aspects that can be attained via UWTSD while making use of active-active
data centre are specified beneath:
Assets:
 This mode will enable them within aggregation of two or more servers through which
network traffic load can be maintained while working like a team and distribute within
the network servers.
 Through utilisation of active-active data centre UWTSD will be able to render
information from cache in which network server do not have to respond. This will enable
them within having processes that will decline network traffic load (Avgerinou, Bertoldi
and Castellazzi, 2017).
 The data centre will enable UWTSD to have adaptability within the varying business
situations through which improvised flexibility can be attained in context of related
metrics.
 Enhanced system performance will be delivered by distribution of load that is present
across the cluster nodes. Moreover, the instant scalability is dependent on the alterations
within the capacity needs of their customers. The relevant standards will be created that will create a probable impact on the uptime
reliability through which lots of money can be saved. Moreover, the client centric
solutions like real time lookup, cloud computing and many more will be available for the
UWTSD through which services can be delivered in an appropriate manner.
Limitations:
 By setting up load balancers within this mode means that the firm is rendering their
liabilities at full capacities. In case, if load balancers fail then the network servers might
slow down and session of users might be timed out.
 UWTSD will not have complete control locally as the hardware is usually present within
the remote location. In addition to this, the quality along with usage of data will also vary
depending upon the internet connectivity at the premises of their customers (Dandres and
et. al, 2017).
11

 The information might be at critical risk from malware infections, virus or corporate
theft. Furthermore, latency issues might prevail on the basis of bandwidth.
Analysis
The above section clearly illustrates the designing of the data centre along with this the
certain points are specified that needs to be considered. When organisation is delivering their
liabilities within the distributed computing environment in which application workloads interact
across public as well as private network, it becomes important to make use of encryption in order
to protect data from being vulnerable. In addition to this, there are certain legal requirements for
safeguarding from potential threats and making sure that DC complies with emerging standards
like ISO 27001 and regulations such as HIPAA and PCI. This will enable UWTSD to have
robust security strategies for safeguarding their services and operations from probable threats. In
addition to this, adaptive security measures must be utilised for allowing workload to move
freely without any kind on intervention from third person (Eltraify and et. al, 2019). For this, it is
mandatory that regular auditing is carried out and be attentive to the alerts related with any kind
of potential risks. Through the usage of proper planning it becomes easy to design data centre
along with this it has to be ensured that temperature aspect are maintained so that overall cost
can be declined like even if 0.5 temperature is reduced then this will have probable impact on the
expenses that will be created on the annual basis. The option of scalability can only be attained in
case space for expansion is left and it is important as information needs are changing drastically.
Conclusion
From above, it can be concluded that, data centre refers to centralised locations in which
networking as well as computing equipments are concentrated for the rationale of gathering,
storage, processing, distribution and rendering assess to ample of data. This involves backup
power supplies, environmental controls, communication connections and many more security
devices. It completely depends upon the organisation that what kind of DC is required by them
either in-house or from vendor. Each provides different benefits as well as possesses certain
restrictions. It is important that firm considers both the aspects while opting for anyone of them
so that there rationale can be attained and relevant standards of services are furnished.
12

Recommendations
It has been identified that data centre can be designed but there are different standards
that must be taken into consideration while doing so. It is recommended that while carrying out
the designing process ISO27001 standard must be utilised. This is international standard that is
accountable for setting up specifications for ISMS. This will also enable them within addressing
the security aspect that might occur while delivering their services. In addition to this while
doing so or designing the data centre, it is suggested that UWTSD must leave the room for
further growth as the entire process is extremely expensive and thus it is important that when
investment is made then space is left for scalability. In addition to this, power requirements are
massive and planning must be carried out effectively through which future needs can be
acknowledged. It is important that experiments must be carried out with respect to the
temperature so that overall cost can be minimised. It is recommended that relevant temperatures
are maintained while building up data centre and in case if facilities get hotter than this will
probably lead to catastrophic failures that will lead to lose millions of cost.
This is recommended that it becomes crucial for UWTSD to take into consideration time
within planning of cooling along with airflow for data centre. With respect to this, it is crucial
that forecasting capabilities minimise over provisioning as well as enable data centre operators
for anticipating the influence of future installations on the critical infrastructure. Furthermore, it
is suggested that there needs to be relevant airflow in each server rack as it is crucial. It is
important that filler panels must be utilised that will lead to planning strategies in an effectual
manner. The smart airflow plan will lead UWTSD to minimise cooling expenses by near about
40% in the most of the cases. In addition to this, it is important to automate operational
functionalities that are being carried out the employees. This will further lead to conduct
activities in an affirmative manner as well as this will also enable to make sure that infrastructure
must be deployed in consistent manner within identical data centre that will minimise cost and
time spend within maintaining activities.
13

References
Books & Journals
Acton, M. and et. al, 2017. 2018 Best Practice Guidelines for the EU Code of Conduct on Data
Centre Energy Efficiency. Publications Office of the European Union, Luxembourg,
Tech. Report. EUR 29103 EN, 2018.
Alishahi, S. and et. al, 2019. Design and implementation of a data centre for smart grids based on
Software Defined Network Technology (SDN).
Avgerinou, M., Bertoldi, P. and Castellazzi, L., 2017. Trends in data centre energy consumption
under the european code of conduct for data centre energy efficiency. Energies, 10(10),
p.1470.
Dandres, T. and et. al, 2017. Consideration of marginal electricity in real-time minimization of
distributed data centre emissions. Journal of Cleaner Production, 143, pp.116-124.
Eltraify, A.E. and et. al, 2019, July. Experimental Evaluation of Server Centric Passive Optical
Network Based Data Centre Architecture. In 2019 21st International Conference on
Transparent Optical Networks (ICTON) (pp. 1-5). IEEE.
Flucker, S., Tozer, R. and Whitehead, B., 2018. Data centre sustainability–Beyond energy
efficiency. Building Services Engineering Research and Technology, 39(2), pp.173-182.
Hammadi, A.A., 2016. Future PON data centre networks(Doctoral dissertation, University of
Leeds).
Jackson, C. and et. al, 2017, March. Demonstration of the benefits of SDN technology for all-
optical data centre virtualisation. In 2017 Optical Fiber Communications Conference
and Exhibition (OFC) (pp. 1-2). IEEE.
Luo, Y. and et. al, 2019. A framework for waste heat energy recovery within data centre. Energy
Procedia, 158, pp.3788-3794.
Mialle, P., Brown, D. and Arora, N., 2019. Advances in operational processing at the
international data centre. In Infrasound Monitoring for Atmospheric Studies (pp. 209-
248). Springer, Cham.
Rashid, M. and Noraziah, A., 2017. Review on Green Technology Implementation Challenges in
University Data Centre. Advanced Science Letters, 23(11), pp.11134-11137.
Rogers, P., BRIPCO bvba, 2016. Data centre. U.S. Patent 9,347,233.
Sefidcon, A. and Yadhav, V., Telefonaktiebolaget LM Ericsson AB, 2016. Method, system,
computer program and computer program product for monitoring data packet flows
between virtual machines, vms, within a data centre. U.S. Patent Application
15/026,306.
Singh, A.K. and Kumar, J., 2019. Secure and energy aware load balancing framework for cloud
data centre networks. Electronics Letters, 55(9), pp.540-541.
Spyropoulou, M. and et. al, 2020. Towards 1.6 T datacentre interconnect technologies: The
TWILIGHT perspective. Journal of Physics: Photonics.
Sugawara, T. and et. al, 2019. Oscillator without a combinatorial loop and its threat to FPGA in
data centre. Electronics Letters, 55(11), pp.640-642.
Wang, B., 2019, July. Construction Research of University Data Centre of Cloud Computing
Based on Virtualization Technology. In IOP Conference Series: Materials Science and
Engineering (Vol. 569, No. 5, p. 052022). IOP Publishing.
14