CHAPTER VII
RESULTS AND DISCUSSIONS
This part of thesis focus on the results and discussions that was carried out and the
findings from it. To target the goal the best choice was to include individuals from
enterprises already using cloud computing or who are willing to migrate in near
future. This part also presents a discussion on the major findings of the various
components of the literature study and survey.
7.1 Results
The result was compared with existing tools, proposed tools and protocols of different
sizes of organizations work load. Our result was focused particularly of an
organization to fill the gap between their knowledge on cloud and security issues in
traditional IP networks and cloud environment in general.
The main focus of this result was to have an insight in to an organization and
respondents view on scalability, complexity, security concerns, existing costs,
predictable savings, data types and access control mechanisms. We will discuss all of
these in the next section. This work will give an understanding of how organization
needs are aligned in terms of cloud advantages (i.e. services) and how professionals
perceive some of the issues and their solutions. Following are the outcomes of our
work.
(a) Scalability
The sizes of the organizations are 38% were less than 500 employees and 27% more
than 4000. We do comparison on the base of organization size 72% scalability is
calculated in the existing cloud and 88% scalability in proposed work, is the reason
which is driving them to go for cloud computing.
(b) Problems
The problem is that company’s think that cloud computing is less reliable. They even
do not trust the secured VPN deployment for remote access of private data through
cloud service. But a ray of hope is seen when they are convinced that a disaster
recovery infrastructure along with primary infrastructure can improve availability
User Load
Data Size = 500 GB
100
Without
80 Cloud
scalability
60 Existing
Cloud
40
20 Proposed
Tools and
0 Protocols
0 100 200 300 400 500 600
User Load
Figure 38: Scalability Comparison Graph
(c) Platform
More than 50% organizations using Linux as operating system for servers. But some
of them are virtualized servers running on dedicated hardware in their own
datacenters. So they are very concerned about the security and do not want to opt for a
hosted data center or public cloud. The solution can be a private cloud on-premise. It
is because they think that service level agreements from cloud providers can affect
service availability for their organization. Also they think that private cloud in their
own data centers can protect it from becoming hacker’s target..
(d) Potential Threat
Viruses are the main security threats experienced by the network administrators.
Biggest challenges to adopt cloud computing are integration with the existing systems
and network security. The roadmap to cloud computing is to separate the public and
private data from the network. Penetration testing by customer can help building the
trust on cloud service provider security plan.
(e) Complexity
The complexity of existing system is rated 61% that means their system is slightly
hard to maintain. The results show that the organizations complexity is reduced with
proposed tools and protocols. So in this case the best solution can be cloud
computing. Deployment and maintenance are complex, but by selecting a suitable
hypervisor the system can be maintained from one place, and it becomes fairly simple
with the help of proposed cloud tools and protocols.
User Load
Data Size = 500 GB
100
80 Without Cloud
Complexity
60 Existing Cloud
40
Proposed Tools
20 and Protocols
0
0 100 200 300 400 500 600
User Load
Figure 39: Complexity Comparison Graph
(f) Security
In cloud environment data security is necessary at user and server side. In the current
environment some effort is put into the security and control related activities by
dedicating some resource percentage. The main concern is the data which is important
to organization. We have to ensure the organization that their data will be safe once
they move towards the cloud. And this can be achieved by cloud provider by taking
every step like security protocols, encrypting the data and personal firewalls. Beside
this single sign-on and single sign-off should be enforced to enhance security and user
friendliness.
(g) Access Control Management
Access control lists are used in most organizations as access control management.
Organization is more concerned about the security model being deployed by the cloud
provider. Data should be secure both at storage and transit, for that purpose we
developed two protocols DSP,CSP and Cloud Servicer User Security Tool.
User Load
Data Size = 500 GB
100
Without Cloud
80
Security
60 Existing Cloud
40
Proposed Tools
20
and Protocols
0
0 100 200 300 400 500 600
User Load
Figure 40: Security Comparison Graph
(h) Cost Saving
Approximately 11 to 20 percent of respondents organization’s IT budget was
allocated for cloud initiatives. The main reason that an organization’s shift towards
cloud computing is because of reduced cost 95% and less maintenance disaster
recovery cost in different structure of applications, platforms and network.
User Load
Data Size = 500 GB
100
80 Without Cloud
Cost Saving
60
Existing Cloud
40
20 Proposed Tools
and Protocols
0
0 100 200 300 400 500 600
User Load
Figure 41: Cost Saving Comparison Graph
(i) Data types
Cloud can increase web traffic in their networks. As far as confidentiality is
concerned it depends on the organization whether they want to move sensitive data to
cloud or not.
(j) Performance
Performance of the system can be increased by using proposed tools and protocols,
because with the help of proposed tools and protocols resources of the system are
allocated only to authorized user. This increases resource utilization and decreases
response time of the CPU by that we can increase CPU utilization and performance.
Requests / Second Vs User Load
Data Size = 100 GB
160
140 Without
120 Cloud
100 Existing
RPS
80 Cloud
60
40 Proposed
20
0
0 10 20 30 40 50 60
User load
Figure 42: CPU Utilization Comparison Graph
Response Time Vs User Load
Data Size = 4 KB
300
250
Response Time
200
150
100
50
0
0 10 20 30 40 50 60
User Load
Figure 43: Graph for 4 KB data file
Propose tools and protocol increase performance of the system if user load is increase
as compare to existing tools.
Response Time Vs. User Load
Data Size = 4 KB
250 VPN
Response Time
200 XEN Tool
150
Citrix Tool
Proposed Tool
100
50
0
0 10 20 30 40 50 60
User Load
Figure 44: Comparison 4 KB data with Existing Tools
Response Time Vs User Load
Data Size = 500 KB
400
350
Response Time
300
250
200
150
100
50
0
0 10 20 30 40 50 60
User Load
Figure 45: Graph for 500 KB data file in seconds.
Figure 46: Comparison 500 KB data with Existing Tools
7.2 Discussion
This thesis has presented in detail about the implementation of cloud environment for
a secure organization where data is prone to security threats .We have categorized our
work in a manner so that we can first find out why we should use cloud infrastructure
and later we gave the details of one of the techniques and tools used for its
implementation. We planned to provide enough proof to make them aware to use
cloud based services in their network.
We started our research by our literature review in which we read in detail and
understand the basics of the cloud computing. Our literature review provided us with
a flood of knowledge related to cloud computing which we had to filter to our
required needs. From the initial studies we came to know that cloud has different
deployment models, of which private cloud is perceived as more secure. Further we
linked our literature studies with different service models provided by vendors to
different organizations. We discussed about different common characteristics, benefits
and security concerns in cloud infrastructure. Then further a list of different vendors,
their services, their clients and the tools being used to provide these services to
different organizations were discussed. This led to much encouragement to adapt
cloud infrastructure.
To find out how these infrastructures are implemented specially related to monitoring,
securing of data and related to the present market competitors and their
implementation strategies showed us that virtualization as the most used technology.
We thoroughly discussed general benefits, security benefits, different types and most
used tools for virtualization. This led us to specified tools like hypervisors being used
to implement this technology. In our work we found out the most of the industry is
using VMware or Xen as hypervisors. We then focused on VMware (proprietary) and
Xen (open source) as a tool to monitor and secure a network. Most of the details were
related to Xen architecture and its network configurations as we should also keep in
mind the standardization issues which might cause a vendor lock in if proprietary
tools are used. We developed Cloud Service User Security tool and Cloud Security
Protocol. These tools increase cloud performance, reduce cost and provide high
security at user and server level. Then we developed Multilevel Framework and
Cloud Security Protocol to provide security at user and provider level. Data Storage
Protocol provides security for data in cloud environment. Data Storage Protocol is
developed to provide the proof at user side and security to data at server machine by
using metadata. User can remotely check the server data to ensure the data security.
Cost is one of the major role plays in the migration of organizations to cloud network.
We have taken into account the cost of public clouds and their policies related to
costs. In the work we came to know that organizations are willing to spend more
resources on cloud based environments if they are assured of better reliability and
data security. Cloud Service User Security tools, protocols and virtualization, is the
technique which can help cut costs in their network as many expenses are already
being used on maintenance. More data capacity is required for expansion of network
which becomes even more costly. Hence cloud is the cheap solution for such
hindrances.
