A Repeatable Framework for Best Fit Cloud Solution
Emmanuel Kayode Akinshola Ogunshile
Department of Computer Science, University of the West of England, Bristol, U.K.
Keywords: AH – Authentication Header, CPU – Central Processing Unit, DNS – Domain Name Server, EPM – Evans
Property Management, ESP – Encapsulation Security Payload, IaaS – Infrastructure as a Service, LAN –
Local Area Network, NAS – Network Area Storage, PaaS – Platform as a Service, SaaS – Software as a
Service, SME – Small Medium Enterprise, SMTP – Simple Mail Transfer Protocol, SSL – Secure Socket
Layer, VPN – Virtual Private Network.
Abstract: To respond to business challenges with agility, modern businesses have to evolve quickly to stay competitive.
Unfortunately, in many situations, proliferation of heterogeneous Information Technology shifts act as a
barrier to innovations instead of as a driving force. Crucially, this is due to the confusions that they sometimes
cause whilst Small Medium Enterprises (SMEs) are trying to elect the right technology solution appropriate
for a given business challenge i.e. amidst various comparable options, claims, features and benefits from
different technology vendors available in the market. To help small SMEs quickly make timely decision on
what technology solutions are appropriate for a given business challenge i.e. given the vast array of solutions
available in today’s market, this paper proposes a guideline for an implementable solution for any SME with
similar requirements to our chosen fictitious customer called EPM. The paper will cover main areas such as
introducing a generic SME business case, analysing hardware solutions and methods typically employed in
cloud networks to reduce costs. Then the paper will introduce the solutions as a repeatable framework to be
critically analysed to find a suitable solution for the customer, this will then be looked into with any other
cloud principals that could create a better fitting solution for the customer.
1 INTRODUCTION
In this paper, we will be analysing cloud and the
factors behind its popularity driving more and more
users towards its style of IT. One of the major barriers
over cloud in this era is the user misunderstanding and
confusion of what cloud is and what it can deliver.
In particular, we will relate this paper to a generic
small to medium organisation, as this is the audience
where potential growth can be easily obtained. This is
due to SMEs having expendable income that can be
invested in internal growth, and one of the major
areas that growth can be achieved is through cloud
services.
The rest of this paper is organised as follows: The
background research and organisational structure of
the chosen SME is covered in section 2. Section 3
describes the contributions of this paper. Section 4
illustrates the motivation, driving this paper. The risks
involved with cloud computing and their impact is
covered in section 4. Section 5 introduces the
technology that will be analysed and implemented
into the framework at the end of the paper. The
method of the proposed framework is explained in
section 6. Finally section 7 covers the conclusion of
the paper.
2 BUSINESS CASE
In this paper we will relate our findings back to a
specific business case for a SME. The reason behind
this is to target our findings toward a specific set of
requirements. Although the paper will cover a range
of cloud converged services and infrastructures, the
business scope will allow for some direction to be
applied to these findings. The business background is
described in the section that follows.
2.1 Business Background
The business that will be analysed is Evans Property
Management Ltd (EPM). EPM has a large portfolio
of properties which they are renting to customers.
194
Ogunshile, E.
A Repeatable Framework for Best Fit Cloud Solution.
In Proceedings of the 6th International Conference on Cloud Computing and Services Science (CLOSER 2016) - Volume 1, pages 194-205
ISBN: 978-989-758-182-3
Copyright
c
2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
The business currently employs 100 staff spread
over different departments. However the business is
looking to grow the business so more staff could be
expected in the future.
The business has been active since 2007 (9 years),
much of the equipment that is used by the business
currently is outdated and not meeting the
requirements expected by EPM. Many of the servers
currently used are failing due to high load balancing
operations that have increased gradually over the 9
years of business. Another problem with current
hardware installed in the business is that it does not
allow for sufficient data storage, file servers installed
within the business only allow for a maximum
capacity of 1TB, this has now reached its maximum
storage capacity and needs to be resolved with new
solutions.
EPM already owns an area that is currently being
as their data centre, there is no issue with size
limitations as they guarantee the area is large enough
to hold any solution.
The departments within the business [Table 1] all
have a shared business plan and can share a solution
if this keeps in line with their main business focus of
reducing costs.
EPM current has gigabit Ethernet installed within
the company, however current hardware used by the
business cannot take full advantage of these
transmission speeds. This is something they would
like to resolve with new hardware. The main
constraint that EPM has stressed is that of money, this
is a highly important resource to EPM and SMEs in
general. The budget that has been provided by EPM
is £5,000.00.
Here the organisation structure of EPM is shown,
this gives a clear understanding of who the main
stakeholders in the business are and what their
primary concerns are.
Table 1: EPM Organisation Structure.
Department Title Name
- CEO David Evans
Finance CFO Lesley Blake
Sales VP George Wright
IT CIO Dave Clark
Table 2: Number of Employees within EPM.
Department Employees
Finance 20
Sales 65
IT 15
The stakeholders discussed above each want their
problems to be considered in the final implemented
solution. These problems are shown in Table 3, each
department head wants problems addressed.
Table 3: Current Hardware in EPM.
Server Role
Department
Owned by
Problems
Experienced
SALES-
Fileserver
File Server Sales
Insufficient Data
Storage.
Requires 1TB
FINANCE
- Fileserver
File Server Finance
Insufficient Data
Storage.
Requires 1TB
Network
Server
Proxy Server
DHCP Server
IT
Server Failure (Due
to malfunctioning
components that
cannot be replaced)
CustMan-
Server
Customer &
Property
Management
IT
Server Failure (Due
to malfunctioning
components that
cannot be replaced)
2.2 Summary
The information provided by EPM above about,
budgets, organisation structure and problems now
taken into account, EPM would also like key concerns
to be at the forefront of any implemented solutions.
Their 2 main concerns are:
1. Scalability
2. Cost Saving
The objectives of the paper will be:
Provide a solution for a budget of £5,000.00
The solution must support:
o Gigabit Ethernet
o Storage Size (Over 1TB)
o Also the solution must support functions such
as a file server, proxy server, DHCP server and
support the customer management application
owned by EPM.
Only solutions that meet the objectives given above
will be chosen for implementation into our proposed
repeatable framework given later.
3 CONTRIBUTIONS
Although there are many papers that explore the most
A Repeatable Framework for Best Fit Cloud Solution
195
befitting solution for a business. These are usually
aimed at customers experienced in a cloud
environment, overlooking customers entering a cloud
consumer market. This paper looks to:
Effectively creates a framework that looks at the
cost benefit of a variety of different solutions on
behalf of a customer.
o It will ensure that the user is offered the most
scalable solution, while minimising the costs
associated with it such as upfront costs and
running costs. This is shown in the method,
across all of section 9.
The paper also contributes to how most SMEs look at
cloud and what factors affect their decision on
moving to the cloud. This is shown in sections 4 and
9.3.
4 MOTIVATION
This paper aims to obtain an understanding of how a
framework can be implemented not only to the
business case in question, but also the wider audience
of SMEs in general. Below we have analysed the push
and pull factors of SMEs in relation to cloud and will
later implement these into a framework.
4.1 What Is the Attraction of Cloud?
Cloud has started to attract a lot of users from all
different backgrounds. Different users have different
requirements from cloud, in the business area that we
are looking at we will analyse the attraction of cloud
towards SMEs.
Figure 1 illustrates the main reasons behind UK
SMEs adopting cloud into their business.
Figure 1: Showing the main attractions of cloud relating to
SMEs (Sahandi et al., 2012).
The top three points taken away from this image
are:
‐ Cost Reduction (45.5%)
‐ Mobility & convenience in accessing applications
(44.9%)
‐ Flexibility & Scalability (38.9%)
These points made above have the most significance
to SMEs as they feel this is how a competitive
advantage will be gained. These are the main areas
we will aim to provide in the framework for a general
cloud infrastructure to be provided to customer at the
end of the paper. EPM has particular attraction to the
point made of cost reduction and how this can be
achieved through an implemented solution.
4.2 What Factors Affect Cloud Take
Up?
With the major attracting factors of cloud taken into
consideration above, it’s time to look at the main
concerns that SMEs such as EPM have of moving to
a cloud deployed environment.
(Sahandi et al., 2012), explained the biggest
concern facing SMEs in cloud adoption in the paper
‘SMEs perception of cloud computing’.
SMEs consider vender lock-in as a major concern
for adopting cloud computing. Cloud computing
users are concerned about losing control of their data
that could be locked-in by a cloud provider.
In a paper published by information weekly
(InformationWeek, 2015), looking at how cloud lock
in can be avoided. Gunnar Hellekson, chief
technology strategist for Red Hat's U.S. public sector
business said that SMEs need to build a detail strategy
for winding down a contract with a service provider.
This can make it easier when migrating service to
another provider if detail documents signed by cloud
providers specifically state how information and data
will be packed up in the case of provider migration.
Figure 2: Factors affecting cloud take up in SMEs (Sahandi
et al., 2012).
CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science
196
The biggest concerns facing UK SMEs on their
transferal to cloud are shown in Figure 2. The top
listed concern is:
‐ Security (VPN Security, 2008)
In the framework presented at the end of this paper,
these two influencing factors will hope to be
minimised to offer the most attraction of cloud
services to customers.
5 CLOUD COMPONENTS
In this section we will talk about the aspects of a cloud
network and how each is beneficial to its deployment.
We will also analyse three providers of these
components, one low range, one medium range and
one high range to gain an understanding of what
quality can be delivered to the customers and for what
price.
5.1 Tower, Rack and Blade Servers
In the current era most data centres incorporate the
user of rack servers, moving on from previous
generations that typically incorporated tower servers.
However blade servers are gaining popularity with
customers. These components all work in similar
ways however they have key differences between
them.
5.1.1 Tower Servers
These servers care becoming outdated and are not
typically employed by customers anymore. Reasons
behind this are that they are large and take up a lot of
space and are typically hard to manage as they cannot
be mounted on top of each other. Also tower servers
require individual monitors, keyboards and mice to
allow them to be managed. All this increases the price
of the solution dramatically, something that all SMEs
look to avoid and ECM in particular.
However all of this being said, tower servers are a
lower cost for customer to initially purchase.
Expansion with tower servers is very achievable
however this increases prices dramatically also, due
to management costs and data centre space issues.
However even though tower servers are becoming
outdated they could fit to customer requirements. For
example: A tower server could serve as a backup
system for a SME as a failover in case the main
system was to fail.
5.1.2 Rack Servers
These are the most common servers implemented by
customers. Unlike tower servers discussed above,
rack servers do not contain critical equipment such as
backup batteries, switches and storage arrays. These
are contained in rack units, in which rack servers are
installed.
Rack servers offer a lot more flexibility and
scalability than tower servers. Rack servers offer
expandability through implementable disks,
processors and RAM which can be added to the
chassis themselves. Also as multiple servers can be
added to racks (installed in bays), this means that all
components can be found together. This will then
make the network easier to manage. Many
applications or functions can be installed in a chassis
across the servers such as:
‐ Email
‐ Storage
‐ Specific business applications
However unlike tower servers, rack servers will need
to purchase chassis for the servers to be deployable,
this adds to business expenses driving up costs.
5.1.3 Blade Servers
Here we will be looking at the final server type which
is blade servers. Blade servers are like self-contained
servers, which fit into enclosures with other blades.
The enclosure provides power, cooling and
connectivity for each blade in the enclosure.
Contained in a single blade are components such as
hard drives, I/O cards, memory and storage. One of
the main selling points behind blade servers is hot
plugging, blades can be added to enclosures easily
just by pushing in the blade when demand calls for it.
This can also improve management of servers,
benefiting the business.
Blade servers are best suited to workloads such as:
Virilisation
o Server
o Desktop
Cloud infrastructure
Big Data Applications
Blade servers have the computing power to be able to
handle high intensity processes. Furthermore blade
servers take up a small area in regards to data centres,
this could cut the cost that needs to be spent on chassis
to house the servers. Blade servers offer flexibility to
customers as they can be added to chassis depending
on the business needs and can be easily scaled up if
needed.
A Repeatable Framework for Best Fit Cloud Solution
197
However blade servers are expensive to purchase
upfront, and may be out of reach of many businesses.
The blade chassis are often made to house 14, 16
blades, this needs to be utilised for customers to really
get the most out of blade.
6 METHOD
In this section of the paper we will be talking through
the generic framework that has been developed to
implement a chosen solution for EPM. In this
framework we will break the solutions down into
criteria important to EPM relating to business goals
and business constraints.
6.1 Mathematical Optimisation
The best solution for EPM will refer to achieving the
best computing power from specified servers under
specified constraints which in EPM case is cost.
Finding the best solution often refers to achieving
best performance under specified constraints (Model
Based Test Case Optimization, 2014).
In the specified case from EPM the performance
criteria that needs to be explored is:
1. Maximisation of computing power (CPU, RAM
etc.)
2. Minimisation of resources (cost)
Achieving these two areas will therefore maximise
the outcome.
These two factors contribute to cloud system as,
maximising computing power will mean customer
have access to a larger resource pool (RAM, CPU,
Hard disk space), access to services when they need
them (On demand self service). The cost saving side
of the analysis will ensure those customers reduce the
amount of resources needed to achieve this. This is
particularly important in SMEs where cost saving is
needed for company growth and survival.
Maximisation of Computing Power.
There are three main aspects consistent over all
solutions that are being analysed in this paper and
they are:
1. CPUs
2. RAM
3. Hard-drive space
We will be looking at maximising all three of these
areas in regards to the cost of each component.
CPU.
To measure the computing power we will be referring
to a publication made at CPU benchmarks (PassMark
CPU Value Chart, 2015). This analysis looks at the
computing power achieved in regards to the cost of
the component and ranks them, rating each
component of the power achieved / price of the
component.
Processor Value for Money =
£
(1)
Each processor to be analysed is included in this
paper and will be implemented into this project.
RAM.
In the analysis of RAM we will be looking at the GB
provided by the component and calculating the cost
per GB. We will be referring to a statistical study
taken, that analyses the cost per GB gained in
hardware (Average Historic Prince of RAM, 2015).
The analysis states that the cost per GB is £3.73per
GB ($5.50). Analysis of this component will look
something like:
Table 4: Showing conversion of RAM GB to total cost (£)
(Average Historic Prince of RAM, 2015).
RAM (GB) Value for money (£)
8GB 29.84
This shows that a hardware product with 8GB of
RAM is equivalent to £29.84. Taking this into
consideration when analysing all systems, this
number will look to be maximised to show customers
are achieving value for money.
Hard-drive Capacity.
In the analysis of Hard-drive space much like that of
RAM above, the cost per GB will be analysed. Hard-
drive cost per GB is relatively cheaper than that of
Ram, according to studies taken, the analysis states
that the cost per GB of hard-drive size is £0.02 per
GB ($0.03) (Average Cost of Hard Drive Storage,
2015).
Analysis of this component will look something
like this:
Table 5: Showing Hard-drive conversion GB to value (£)
(Average Cost of Hard Drive Storage, 2015).
Hard-drive capacity (GB) Value for money (£)
500GB 10.00
This analysis shows us that a hardware product
containing 500GB storage space is equivalent to
£10.00. This value will look to be maximised in the
analysis of servers later on in the paper.
CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science
198
Minimisation of Resources.
There are a few areas to take into account here apart
from the main concern which is total cost of the
solution itself. The aspects to be analysed are:
1. Total cost of the solution
2. Running costs
3. Cooling costs
The latter aspects are often overlooked by customer
when purchasing hardware. However they can make
up a substantial maintenance cost when implemented
in a business. If this was overlooked by a customer it
could mean that the cost benefit of the product was
not accurately reflected at purchase time and could
not provide the requirements that they were
expecting.
Total Cost of Solution.
The total cost of a solution will just look at the total
price of the solution. It will provide a justification
mark for performance aspects of the system.
Example, a solution costing £500 should deliver less
computing power than a solution costing £1000.
However the gap between cost and performance will
be analysed to provide the best solution to a customer.
Running Cost.
Running costs will take into account the electricity
required to run a server over as given timescale and
the cost of this power consumption.
A formula has been taken of how to convert the
electricity consumption of a hardware solution
(Weijdema, 2011):
OperatingHours ∗ Watt UsagePerHour
1000
∗
(2)
The wattage of a solution can be taken using the
formulae:
W = Amps * Volts (3)
Analysis of the power usage of solutions will look to
find the solution with the lowest power consumption.
Although energy prices will vary from location to
location, for the simplicity of the framework to be
applied a constant value of 0.09 (£) per kWh will be
used across all solutions analysed in this paper. This
price has been taken from a quote on UK power, from
the supplier extra energy, (Gas and Electricity Tariffs,
2015).
Cooling Costs.
Cooling costs is also another aspect of solution
running that needs to be taken into account. A
formula can also be implemented to calculate cooling
costs (Weijdema, 2011).
Operating Hours ∗ BTU perhour
∗ 0.293
1000
∗
(4)
Cooling, another aspect that will look to be minimised
across all the solutions.
Summary.
In summary, the analysis of hardware components
that was performed earlier in the paper will now be
implemented into the proposed framework explained
above. Weighing up the initial cost of the propose
system against the provided computing power will
indicate a cost benefit to the customer. This will
provide collated analysis information that will
propose the best solution for EPM.
6.2 Implementing the Business
Scenario
In this section we will be implementing the
framework explained above into the different
solutions that were analysed earlier in the paper. We
will first shown each solution compared in its group,
e.g. Blade, Rack & Tower. We will then go on to
compare each solution group against each other.
Tower Server Comparison.
We will first be looking at the tower server solution
for EPM.
Table 6: Lenovo ThinkServer, Tower Server.
Analysis of this allow us to see the other costs of
a server that are sometimes not taken into
consideration along with the computing output of this
solution.
The numbers highlighted in red shown the
dynamic figures of the solution, they will vary from
solution to solution and indicate the computing value
provided to the customer. These values are displayed
in the component values section.
The Lenovo ThinkServer TS140 allows for up to
32GB of RAM to be installed along with 16TB of
Hard-disk space. These values do not come installed
A Repeatable Framework for Best Fit Cloud Solution
199
in the solution however they offer room for expansion
which will be required in EBS in coming time. These
components will have to be purchased separately,
however this is the same with every solution that will
be displayed in this paper. The CPU value displayed
in this figure, shows that the bench mark performance
of the CPU was relatively good compared to the
performance of similar Intel Xeon processors with
similar GHz. However the low cost of this CPU
allows a much greater value for money to be
achieved.
The running cost shown in this figure illustrates
the amount of electricity required to run this solution.
The running time has been specified by EPM as 24/7
(8760 hours). This is often an overlooked cost by
customers and real costs of a solution are not
explored. This paired with the cooling cost of the
solution, which is calculated in a similar manner
allows us to see the overall costs of a system over a 1
year period.
Table 7: Dell Power Edge, Tower Server.
Table 8: HP ProLiant, Tower Server.
All of the information from the three tower solu-
tions taken into consideration, a graph can be
produced to show the best value for money solution.
This is calculated by subtracting the component costs
from the total cost of the solution.
This illustrates that the dell power edge system is
the best value for money as it delivers the most
amount of potential computing power for the smallest
cost to the customer.
Figure 3: Best Value Solution, Tower Server.
The values shown are negative as each system
does not actually deliver more potential computing
power than is paid for, however the graph indicates
that the closer a solution value is to 0 means that it
provides the same potential computing power as the
overall costs of the solution, providing a better return
on investment.
Rack Servers.
Now we will look at the next solutions recommended
in the analysis. The rack servers will be compared to
each other in the same method that was applied
above, comparing the value of money achieved from
each solution.
The first rack server that will be looked at is the
Asus RS100-X7, this is the entry level rack server for
EPM.
Table 9: Asus, Rack Server.
The medium range server from IBM was the next
solution to be analysed.
£-
£1.000,00
£2.000,00
Lenovo Dell HP
Cost to the User
CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science
200
Table 10: IBM, Rack Server.
Table 11: Cisco, Rack Server.
Now that all of the solution values have been
taken into consideration. The figure below illustrates
the value for money of each solution. Looking at the
overall running costs of the solution subtracting the
potential computing power available.
Figure 4: Best Value Solution, Rack Server.
From the analysis of cost shown in the figure
above, the IBM x3650 rack server showed to be the
best investment for EPM. The x3650 returned
£1713.00 worth of potential computing power to the
solution from the £2, 339.43 upfront cost and running
costs. The other two servers (Asus & Cisco) didn’t
return as much potential computing power as the IBM
rack server. The Cisco solution that was analysed has
a very large upfront cost and the large power
consumption of the solution means that the potential
computing power provided cannot overcome the
large costs, extracting valuable resources from EPM
which would be hard for them to invest in a system
like this.
Blade Servers.
The next solution group that will be looked at is the
blade servers. These were analysed earlier in the
paper and appeared to have very large upfront costs
and high power supply’s required. In the analysis
below it will reveal if blade servers can justify the
high prices of various suppliers products.
The firs solution that will be analysed is the Cisco
UCS B230 M2.
Table 12: Cisco, Blade Server.
The values expressed in the above figure also
takes into account the cost of the blade rack to support
the individual modules.
Table 13: Cisco, Blade Server.
The Dell Power Edge adopts the same method
used in the last blade analysis, the cost of the blade
enclosure has been included in the unit price of the
solution.
£-
£1.000,00
£2.000,00
Asus
RS100-X7
IBM
x3650 MD
Cisco
X220 M3
Value for Money
A Repeatable Framework for Best Fit Cloud Solution
201
Table 14: HP ProLiant, Blade Server.
Now that the values of all of the blade solutions
have been taken into consideration, they can be
analysed together to illustrate which solution
provides the best value for money to EPM.
Figure 5: Best Value Solution, Blade Server.
The graph illustrates that the HP ProLiant was the
best value for money, as it delivered high RAM and
Hard drive capacity unmatched by the other two
solutions. However the CPU power provided is not
the best for the money paid for the solution. All of this
taken into consideration the HP ProLiant Blade server
is the best blade solution for EPM.
What this analysis does not take into
consideration is the amount of blades required by
EPM. As EPM is a small business with fairly simple
computing requirements, the blade solutions are a bit
too high tech for the intended business and therefore
would provide the needs of the business with one
server blade. The values shown above indicate that it
is relatively worse value for money than the other
solutions. However the total price of the solution
incorporates the one off payment of an enclosure to
support up to 8 blades, so sharing this value over 8
blades would decrease the overall upfront enclosure
value. Also as the power is supplier through the
enclosure to blades the running costs would be shared
over the 8 servers, also lowering the total price of the
solution and providing a better value for money
result.
However for true value for money to be
recognised on this solution the customer would have
to purchase 4 or 5 server blades, this is out of scope
for the customer and does not represent a real solution
to the customer’s needs.
Other Solutions.
File Servers. Another server that was considered in
the analysis at the beginning of this paper was the file
server. As EPM require two file servers to be
implemented into the business the below analysis will
look at identifying if a suitable solution lies within file
servers.
Table 15: NetGear ReadyNAS 102, File server component
values and running costs.
Table 16: Asustor, File Server.
The next file server analysed here has slightly
higher specifications than the previous server,
however some areas have not been delivered at a high
value/cost rate.
The two standalone file server products have been
analysed using data about their computational power
and total cost of ownership. The graph below
illustrates the products in comparison.
£-
£5.000,00
£10.000,00
£15.000,00
Cisco UCS B230 M2Dell Power EdgeHP ProLiant
Value for Money
CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science
202
Figure 6: Best Value Solution, File Server.
The two servers taken into consideration do both
deliver great value for money, more so in the case of
the NetGear ReadyNas. The total cost of the product
including running and cooling costs is worth less
value compared to the potential computing power
offered by the solution. The Asustor AS-604T also
offer good value for money.
Analysing these products has shown that they
would be great value for money in a home scenario
where a lot of data storage is required, however they
would not be suitable for SME and commercial use.
Both the products do not offer any additional services
besides file storage.
EPM could implement a standalone file server to
the business and this product would be suitable for the
data requirements needed. However additional
products would have to be purchased to support a
proxy and DHCP server. This would involve
implementing other servers also, adding to the total
cost of the infrastructure for EPM. Also the
standalone file servers would not allow for room with
scalability if the data storage requirements of EPM
became larger. For these reasons file servers will not
be considered for final implementation in EPM.
6.3 Results and Reasoning
As we have now analysed three different areas of
servers, and analysed the key features that relate to
EPM. We will now produce the findings of the
analysis and my recommendation of the solution to be
implemented.
The below figure has been produced to illustrate
all of the server solutions analysed.
The best value solution area for EPM is rack
servers. These provide the highest amount of
potential computing power (RAM Capacity, Hard-
drive capacity & CPU value for money) while
minimising the total cost of the solution.
The importance of the potential computing power
is to ensure that EPM can grow the server (swapping
resources such as RAM, hard-disk space and CPUs
Figure 7: Overall best value solution, looking at blade, rack
and tower servers.
when needed) as the business grows. The initial costs
of the solution is greater than the tower server
solution however this is made up for with the
computational power provided.
IBM x3650 MD has shown to be the most cost
efficient solution to the customer’s problem so we
will suggest that this is implemented as it takes into
account EPM requirements as well as the cost aspect.
Blade servers have not provided a viable solution
for EPM as they offer minimal computing power
(when only one is implemented) for a very large cost,
straining EPMs cost resources. This solution may be
viable in a very large business that requires a lot of
computing power and has the available resources to
purchase the blades and enclosures as well as run
them. However in EPMs case the solution is not an
option that would be available to them.
When choosing the solution for EPM, other areas
must be considered that may not be specific to them
but will be specific to SMEs in general. If these
factors are not considered it could affect the take up
and implementation of the chosen solution and EPM
could simply reject it.
Other factors that need to be considered in the
chosen solution are points that we have made
previously in the paper and been supported by
statistical data (Sahandi et al., 2012; YouTube Virtual
Computing, 2015; Virtualisation Best Practices for
SMEs, 2015; IBM Data Center Operational
Efficiency, 2015), section 4.
How can the solution reduce cost and meet
scalability aspects?
How does the solution offer security?
The solution to be implemented has met scalability
£(50,00)
£-
£50,00
£100,00
NetGear Ready
Nas 102
Asustor AS-
604T NAS
Value For Money
£- £5.000,00 £10.000,00 £15.000,00
Lenovo…
Dell Power…
HP ProLiant…
Asus RS100-X7
IBM x3650 MD
Cisco X220 M3
Cisco UCS…
Dell Power…
HP ProLiant…
Best Value
Solution For EPM
A Repeatable Framework for Best Fit Cloud Solution
203
aspects by providing swappable drive bays for disk
space and RAM. This means that EPM can add
hardware components when needed to support their
network.
The cost reducing aspect could also be
implemented from the solution. We have already
talked about how the best value product has been
chosen for EPM as this is one of their main priorities,
but how can costs be reduced further?
The IBM rack server chosen as EPMs
implementable solution will support the installation
of virtual managers so many users can be supported
from one server. This would reduce the costs as only
one server would need to be purchased even if
multiple Operating systems wanted to be run from it.
Section IX also refers to how EPM could adapt their
server room to operate more efficiently. This could
also ensure that EPM are getting the most potential
out of their servers for the money they paid.
The security aspect has also been considered in
the choice of solution. The IMB rack server comes
with administrative user’s controls as standard, such
as password protection and trusted platform module
(TPM). Furthermore the solution supports security
protocols such as SSL and PPTP. These are the two
fundamental protocols used in VPN connectivity.
This means EPM could implement a VPN network to
the business to secure their connections and reduce
the risk of security breaches. The organisation
structure of EPM would allow for a VPN connection
to be implemented, as the sales and finance
departments may want to keep their files separate
from each other. VPN would allow for the security of
these departments to be maintained while still
operating from the same server.
7 CONCLUSIONS
In this section we will talk through how a network
solution has been applied to EPM and how it meets
their unique requirements. We will then go on to
explain how other measures could be implemented
which have been talked about in areas of this paper.
If we relate back to the business background
introduced at the beginning of this paper.
1. EPM required a network solution for under
£5,000.00
2. The solution must support functions such as File
server, Proxy server, DHCP server and support
the customer management application owned by
EPM.
3. Maximise potential scalability, which may be
needed by EPM in the coming years.
The solution that we have chosen for implementation
(IBM X3650 MD) meets all of the criteria stated by
EPM.
The framework that was created to analyse each
potential solution for EPM, found that products more
focused towards a home use audience generally
provided more value for money and a lower TCO.
However the solutions focused towards a SME
audience delivered a lower value for money in terms
RAM, CPU and Hard-Drive, and a higher TCO. This
is to be expected in a higher grade solution as the
extra costs are justifiable by the extra computing
power provided and the extra functionality that the
solutions can provide.
In this paper we feel that all of EPMs unique
requirements have been implemented with the chosen
solution. However we also feel a more general
approach has been taken relating to all SMEs, as the
views and concerns of various SMEs have been taken
into consideration to provide a solution that creates an
all-round fit for the customer.
REFERENCES
Sahandi, R., Alkhalil, A. and Opara-Martins, J. O.-M.
(2011) L.M. Camarinha-Matos, L. Xu, and H.
Afsarmanesh (Eds.): PRO-VE 2012, IFIP AICT 380,
pp. 186–195, 2012. IFIP International Federation for
Information Processing 2012 SMEs’ Perception of
Cloud Computing: Potential and Security.
InformationWeek, 'Avoiding Cloud Lock-In -
InformationWeek', 2015. [Online]. Available:
http://www.informationweek.com/cloud/avoiding-
cloud-lock-in/d/d-id/1111982. [Accessed: 23- Apr-
2015].
T.G HKSAR, 'VPN SECURITY', 2008.
YouTube, 'A simple explanation of virtualization (virtual
computing', 2015. [Online]. Available:
https://www.youtube.com/watch?v=hPkEqOoQSu4.
[Accessed: 23- Apr- 2015]
Cisco, 'Virtualization Best Practices for SMEs', 2015.
[Online]. Available: http://www.cisco.com/web/IN/
about/network/virtualization_smes.html. [Accessed:
23- Mar- 2015].
IBM Data center operational efficiency best practices,
2015. [Online]. Available: http://www-
935.ibm.com/services/multimedia/IBM_Global_Data_
Centre_Study.pdf. [Accessed: 01-May-2015].
‘Model Based Test Case Optimization of UML Activity
Diagrams using Evolutionary Algorithms', Academia
EDU, vol. 2, no. 211, 2014.
Cpubenchmark.net, 'PassMark CPU Value Chart -
Performance / Price of available CPUs', 2015. [Online].
Available: https://www.cpubenchmark.net/cpu_value_
available.html. [Accessed: 23- Apr- 2015].
CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science
204
Statisticbrain.com, 'Average Historic Price of RAM |
Statistic Brain', 2015. [Online]. Available:
http://www.statisticbrain.com/average-historic-price-
of-ram/. [Accessed: 23- Apr- 2015].
Statisticbrain.com, 'Average Cost of Hard Drive Storage |
Statistic Brain', 2015. [Online]. Available:
http://www.statisticbrain.com/average-cost-of-hard-
drive-storage/. [Accessed: 23- Apr- 2015].
Ukpower.co.uk, 'Gas & electricity tariffs per kWh | energy
prices per unit - UK Power', 2015. [Online]. Available:
http://www.ukpower.co.uk/home_energy/tariffs-per-
unit-kwh. [Accessed: 22- Mar- 2015].
Weijdema, E.. 'How to calculate electrical costs for cooling
and power consumption', VMGuru, 2011. [Online].
Available: https://www.vmguru.com/2011/06/how-to-
calculate-electrical-costs-for-cooling-and-power-consu
mption/. [Accessed: 23- Apr- 2015].
A Repeatable Framework for Best Fit Cloud Solution
205
