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

net/publication/350545879

SECURITY IN IAAS CLOUD COMPUTING

Article · April 2021

CITATION READS

1 816

2 authors, including:

Dr. kavita
Jagannath University
141 PUBLICATIONS 98 CITATIONS

SEE PROFILE

All content following this page was uploaded by Dr. kavita on 01 April 2021.

The user has requested enhancement of the downloaded file.

 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

SECURITY IN IAAS CLOUD COMPUTING

Manoj Kumar
Research Scholar, Jayoti Vidyapeeth Women’s University, Jaipur
Dr. Kavita
Assistant Professor, Jayoti Vidyapeeth Women’s University, Jaipur

Abstract:- Securing our digital assets has become increasingly challenging

as our reliance on rapidly evolving technologies continues to grow. The
security perimeter in computing has changed from a well-defined boundary
that was relatively easy to identify and defend, to an elastic boundary that
is constantly changing and for which the threats are constantly evolving.
This paper investigates the complex security challenges that are introduced
by the trend towards Infrastructure as a Service (IaaS)- based cloud
computing. While not exhaustive, it identifies some technological and legal
issues and concerns from the perspectives of identified stakeholders, and
suggests some future directions for security research and development to
help advance the security posture of this technology.

INTRODUCTION
The term cloud computing means many things to many people and the
definitions will no doubt continue to evolve as new technologies and services
enabling this model of computing are developed. For the purpose of this
paper, the scope will be limited to IaaS cloud computing: a model in which
units of computation and/or storage are allocated to consumers, who then
access their assigned resources via some Wide Area Network. The cloud
system consumers are granted complete control of any resources assigned to
them, but have no control of the underlying virtualisation or partitioning
layer, the physical host on which it executes, or the mapping of virtual
resources to physical devices. While the security concerns for this realm are
largely applicable to any external handling and processing of an individual

556
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

or organisation’s digital assets, we will focus on security concerns for

computational cloud computing from the perspectives of cloud service users,
cloud service providers, and general security practitioners in both the
technical and legal realms.

BACKGROUND
The rapid evolution of technology coupled with our increased dependence on
the same has melded together to make securing our digital assets an
challenging problem. Originally computer systems were physically isolated
and a data-centric approach to securing systems was accomplished largely
through perimeter security. Physical security generally provided a means to
isolate and secure the systems from malicious outsiders.
As technologies continued to evolve, and connectivity and mobility
increased, it became difficult to secure an increasingly fluid perimeter. The
focus of security began to shift from physical security and securing the data
centres to protecting the endpoints themselves. This was accomplished
through many mechanisms including firewalls, confining the end point
services, changing configurations to restrict access, and similar measures.
As connectivity increased, the security focus again had to shift, to protect
the plethora of applications that depended on the network. These included
applications that request and provide data, distributed components, and
other virtual workgroups. This focus shift required that protection at the
application layer become a concern.
The new evolution towards cloud computing, both IaaS-based and (as
data and services are “outsourced” to the cloud) Software as a Service
(SaaS)-based, again demands a reconsideration of methods used to provide
security. The new critical point is that the changing perimeter that extends
further into realms that are controlled by others. The concern is how to data
in transit, in storage, and also from the service providers. The roles of the
traditional stakeholders in such a system are changing and the distinction
between insider and outsider has become increasingly blurred. In some
cases, the degree of “insider ness” associated with an stakeholder needs to

557
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

be identified in order to effectively assess risks as the traditional binary

definition that requires a clear insider/outsider boundary is longer an
appropriate model.

TECHNICAL CHALLENGES
While IaaS-based cloud computing brings many advantages, this model also
raises significant technical security considerations and questions. Among
these issues are operational trust modes, resource sharing, new attack
strategies, and digital forensics. These are important areas of concern and
are further complicated by the issues associated with giving up control in
any environment. As you increasingly rely on others to provide you with
functionality, you Correspondingly give them control. Further this control is
given to someone who most likely does not share your vested interest in
your mission. As you relinquish control, you lose access to information and
as a result, give up the ability to answer some of the important questions
regarding technical and jurisdictional issues. In some cases, the information
required to make informed security decisions is no longer available, and that
data which is available, may no longer be as trustworthy as it would be in a
system over which you have complete control. As a result it is vital that the
security implications of cloud computing be carefully considered and
factored into a decision about the appropriateness of a cloud-based solution
to a given set of IT requirements.
Operational Trust Modes
A primary technical challenge that results from the decision to use IaaS-
based cloud computing is the level of trust accorded the resource provider.
These providers are part of your system and their roles as stakeholders in
your business are complicated by the blurring of the insider/outsider line.
Thus, there are two aspects here. The first is to determine what operational
levels of trust are needed to capture the trust in the resource providers. The
second is to use those levels in the risk assessment.
In our cloud computing scenario above, the cloud resource provider
maintains sole access to the underlying physical components of the system,

558
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

and provides the cloud consumer with full control over some portion of it. In
practice this means that the provider has access to all of the consumer’s
operations and data in the cloud environment. Much of this may be
business sensitive, and the security plan must take this into consideration.
Among the approaches that can be used are:
 If the cloud is used only to store data, well- established
cryptographic algorithms allow a cloud storage consumer to encrypt
the data prior to insertion in the cloud, and decrypt it after moving
it back to her own systems. It should be noted, however, that such
algorithms are not effective if the data is encrypted for use within
the cloud (as the computation device in the cloud would require
access to the decryption key, which would then also make the key,
and consequently the data, available to the cloud resource
provider.
 In some cases it may be possible for a computational engine to
perform operations on encrypted data, producing results without
exposing the input data or the result in plaintext form. Such schemes
are known as homomorphic encryption. While this is an area of active
research, current results tend to be applicable to fairly narrow
operations. However, even in this mode of operation, a malicious cloud
provider may be able to infer useful information from the operations
being performed, and the characteristics of the results.
We must also consider the extension of the network in the cloud
environment, because that portion of the network provides the connection
between the consumers and the cloud resources. In a more traditional
environment, it may be reasonable for a corporation to control that entire
pathway, or at least have some well-understood pathway between their
offices and the corporate data centre. In the cloud model, the WAN
connections are likely to be far more complicated and dynamic. While
encryption of network traffic on such links can provide some level of
protection, we must again consider that even encrypted content can
provide interesting information that might be exploitable.

559
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

Can we run in a mode in which we do run not need to trust the cloud
providers and hosts? Do we need to trust the networks and the network
providers? Not trusting the latter may be simpler than not trusting the
former, but is non-trivial if attacks against the core network infrastructure
and routing algorithms are considered.
Resource Sharing
In the current corporate computing model, resources such as storage and
hosts tend to be used exclusively by a single corporate entity. However, in
the cloud model it is entirely reasonable that a resource allocated to one
corporation may be instantiated on some physical infrastructure that also
hosts resources allocated to other corporate users. For example, a virtual
machine may be instantiated on a physical server hosting several virtual
machines, each allocated to a different corporation. In this case, it is quite
possible that two competitors may be allocated resources on the same
physical infrastructure. Then security policies and procedures must
consider the possibility that data may leak between competing corporations,
or that the actions of one corporation could impact the ability of a
competitor to conduct business. Some work has already demonstrated this
potential.
Given this concern, we should consider whether a cloud provider can
provide some minimal level of assurance that such conflict will not arise.
Obviously the easiest way to do this is to allocate any physical component
in the cloud to a single consumer, but this would significantly impact the
flexibility of the cloud. A more practical approach may be to implement an
analog of the Brewer and Nash model, in which corporate cloud consumers
are grouped into conflict of interest classes. The cloud provider would then
be free to allocate resources in the cloud with the limitation that no two
corporations in the same conflict of interest class can share the same
physical resources.

NEW ATTACK STRATEGIES

The cloud model makes new attack strategies possible. These strategies

560
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

must be considered when assessing and attempting to secure cloud

computing resources. One example of such a strategy is the ability of an
attacker to co-locate its resources with that of their target. Once the attacker
has gained such a foothold, subsequent efforts to attack the targets may be
possible, both by using current attacks and by attacking the virtualisation
layer and/or physical hardware directly.
Digital Forensics
Several issues make digital forensics in the cloud more challenging than
with ordinary systems and networks. These issues include:
 The ephemeral nature of cloud resources. Hosts in the cloud may be
instantiated for the duration of some processing event (e.g., monthly
accounting), after which they are decommissioned and their resources
(CPU, RAM, storage, etc) are returned to the cloud for use by other
hosts. Unless an event worthy of forensic investigation is discovered
during the lifetime of the virtual host, or possibly very soon after that
virtual host is decommissioned, it is likely that the “system” to be
examined is lost. Once the resources are used by some other host in
the cloud, recovering any useful data from the original virtual host is
probably impossible, as would be attributing any data actually
recovered to the original virtual host with any degree of confidence.
 Seizing a “system” for examination. In traditional digital forensics, it is
relatively easy to seize an entire system, including all processing and
storage components, for examination offline. We can also feel relatively
confident that the data written by the system was written to these
devices, and that data found on these storage devices was written by
the host being examined. However, in cloud systems, what comprises
the system to be examined is far less straightforward. At a logical
level, we could attempt to seize the virtual disks associated with the
cloud host in question, but this may consist of virtual disk file that
was written to many areas of multiple physical storage devices, by
many different physical hosts, as the cloud resource was migrated.
While finding the current version of the file is probably quite easy,

561
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

finding historical artifacts is not, because they may be found on other

areas of the physical disk, no longer be associated with this virtual
disk file, or be on other physical disks no longer associated with this
virtual host.
Some other operations currently used by digital forensics investigators
(such as live analysis) can be performed more easily in virtual environments
such as those commonly found in the cloud.
A full examination of digital forensics in a cloud computing
environment is beyond the scope of this paper. But it is clear that digital
forensics in the cloud presents challenges, and provides opportunities, not
found in digital forensics as it is more commonly practiced today.
Legal Issues
In addition to the identified technical challenges, legal issues associated
with IaaS-based cloud computing need to be incorporated into a risk
analysis plan. That way, potential consumers can make informed decisions
about the appropriateness of utilising this potentially valuable resource.
These legal issues include consideration of jurisdictional issues, an
understanding and the evolution of cloud stakeholder rights, and technical
approaches to addressing the associated legal and jurisdictional issues.
Jurisdictional Issues
Resources in the cloud are often not fixed to any geographical location such
as a specific data centre. They may migrate between physical locations
during their lifetime. The decision as where a resource is instantiated or
migrated may be based on a variety of factors, including load balancing by
the cloud provider, network and datacenter performance and availability,
and even the characteristics of the current clients. The result is that the
resource may exist in multiple legal jurisdictions, each of which may have
different, and even conflicting, rules about important security issues such
as intrusions and data protection. In such a dynamic environment, it may
not be possible for a cloud resource consumer to remain in compliance with
the legal requirements of the jurisdictions in which their assigned resources
may operate.

562
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

For example, consider a resource which is instantiated in a

jurisdiction that does not require personally identifiable information to be
encrypted. The data is subsequently migrated to a jurisdiction for which
such protection is mandatory. Should the cloud resource consumer
configure her system to comply with the most stringent legal requirements
in any jurisdiction in which the cloud has nodes—indeed, would the
consumer even know what jurisdiction this was? Would it even be possible—
for example, the cloud may span two jurisdictions with conflicting legal
requirements, in which case it would not be possible to configure a system
to meet the most restrictive case.
This issue is certainly not new. There are many examples of
companies with data centres in multiple jurisdictions. Companies with
multiple data centres have at least some known set of locations in which the
resources reside, and the resource consumer and resource provider are often
functionally equivalent entities. Even in cases where one corporate entity is
purchasing resources in a data centre belong to another entity, there is still
generally a contract describing the location of service, and any co-location
services that will be provided.
But the cloud model is somewhat different. In that model, the cloud
resource consumer and cloud resource provider are seldom the same entity,
and the contract between cloud consumer and provider tends to describe the
resource that will be provided rather than the locations at which these
services will be provided. The cloud provider is then free to provision
resources to meet these requirements at their discretion, and subsequently
migrate them to address their continually changing system status.

CLOUD STAKEHOLDER RIGHTS

Given that a resource logically located “in the cloud” can be instantiated in
and migrated among multiple physical locations, we consider the
implications this has for the stakeholders in this process.
 Cloud Provider: Given that the act of migrating a (virtual) host may
change the legality of the activity taking place on that host, what

563
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

restrictions should be placed on the provider, and to what extent are

they liable for illegal activity that results in such a move? For example,
consider a host that provides a social networking site. If this host is
migrated from jurisdiction A, which has no cyber-bullying legislation,
to jurisdiction B, which prohibits cyber bullying, is activity that was
legal yesterday now illegal based only on this migration within the
cloud? From another point of view, can the cloud provider migrate
hosts into a jurisdiction if they know that the activity already
occurring on the host is illegal in that jurisdiction? Is such a
migration legal to aid law enforcement, and if so under what
conditions?
 Cloud Resource End Users: Can a user of some resource in a cloud-
based system be expected to know when their activities are illegal?
Revisiting our cyber bullying example, if jurisdiction A has no cyber
bullying provisions, and jurisdiction B does, a user could post three
identical messages from the same location on the same service, and
that same action could be legal on the first and last days, and illegal
on the second day, because the server was migrated from jurisdiction
A to B and back again.
Technical Solutions
One solution to this issue would be for the cloud resource consumer to “tag”
their resources in a manner that would indicate which components could be
migrated, to where, and under what conditions. The cloud provider could
make informed decisions about migration based not only on the available
resources in the cloud, but also on the requirements of the consumers as
indicated by the tags. This would simplify the legal issues, particularly for
the consumers. But it would also quite likely limit the cloud provider’s ability
to efficiently manage their resources, resulting in a model that looks far
more like the current “rack space in a datacenter” approach than the free-
flowing and flexible cloud many currently envision.

564
 ISSN: 0374-8588

Volume 21 Issue 3, October 2019

FUTURE CONSIDERATIONS
While not insurmountable, the challenges associated with IaaS-based cloud
computing need to investigated in order to protect our digital assets. An
increased understanding of cloud computing and the roles of various
stakeholders in this realm are important, as is more research into the
technical and legal issues that resource-based cloud computing introduce to
the threat horizon. This requires cloud-oriented research into identification
of technological issues including trust modes, resource sharing, attack
strategies, and digital forensics implications. Also, legal issues such as
jurisdictional issues, cloud stakeholder roles and rights, and technological
approaches to solving these problems should be paramount in resource-
based cloud computing research and development.

REFERENCES
1. Change, C, and S. Tsu. Arithmetic operations on encrypted data. International
Journal of Computer Mathematics. International Journal of Computer
Mathematics, Volume 56, Issue 1 &21995,pages1-10
2. Micciancio, D. 2010. A first glimpse of cryptography's Holy Grail. Commun. ACM 53,
3 (Mar. 2010), 96-96. DOI= http://doi.acm.org/10.1145/1666420.1666445
3. Brewer, D.F.C., and M. Nash. "The Chinese Wall Security Policy," pp. 206, 1989
IEEE Symposium on Security and Privacy, 1989.

565

View publication stats