Security Measures for Virtual and Cloud Environment

Dr. Aly, O.
Computer Science

Introduction

The purpose of this discussion is to discuss and analyze security measures for virtual and cloud environments. It also discusses and analyzes the current security models and the possibility for additional enhancements to increase the protection for these virtual and cloud environments. 

Virtualization

Virtualization is a core technology in Cloud Computing technology.  The purpose of Virtualization in Cloud Computing is to virtualize the resources to Cloud Computing Service Models such as Software-as-a-Service (SaaS), Infrastructure-as-a-Service (IaaS), and Platform-as-a-Service (PaaS) (Gupta, Srivastava, & Chauhan, 2016).   Virtualization allows creating many instances of Virtual Machines (VMs) in a single physical operating system.  The utilization of these VMS provides flexibility, agility, and scalability to the Cloud Computing resources.  The VM is provided to the client to access resources at a remote location using the virtualization computing technique.  Key features of Virtualization include the resource utilization using isolation among hardware, operating systems, and software.  Another key feature of Virtualization is the multi-tenancy for simultaneous access of the VMs residing in a single physical machine. After the VM is created, it can be copied and migrated.  These features of the Virtualization are double-edged as they provide flexibility, scalability, and agility, while they cause security challenges and concerns.  The security concerns are one of the biggest obstacles to the widespread adoption of the Cloud Computing (Ali, Khan, & Vasilakos, 2015). 

The hardware Virtualization using the physical machine is implemented using hypervisor.  The hypervisor has two types:  Type 1 and Type 2. Type 1 of the hypervisor is called “Bare Metal Hypervisor” as illustrated in Figure 1.  Type 2 of the hypervisor is called “Hosted Hypervisor” as illustrated in Figure 2.   The “Bare Metal Hypervisor” provides a layer between the physical system and the VMs, while the “Hosted Hypervisor” is deployed on the Operating System.

Figure 1.  Hypervisor Type 1: Bare Metal Hypervisor. Adapted from (Gupta et al., 2016).

Figure 2: Hypervisor Type 2: Hosted Hypervisor. Adapted from (Gupta et al., 2016).

Virtualization has many security flaws to intruders.  The traditional security measures that control physical systems are found inadequate or ineffective when dealing with the virtualized data center, hybrid and private Cloud environment (Gupta et al., 2016).  Moreover, the default configuration of the hypervisor does not always include security measures that can protect the virtual and cloud environment.

One of the roles of the hypervisor is to control the management between the VMs and the physical resources.  In Type 1 Hypervisor “Bare Metal Hypervisor,” the single point of failure increases the security breaches for the whole virtualized physical environment on the physical system.  In Type 2 Hypervisor “Hosted Hypervisor,” the configuration exposes more threats than the “Bare Metal Hypervisor.”  The VMs, which are hosted in the physical system, communicate with each other which can cause the loopholes to the intruders. 

Virtualization is exposed to various types of threats and vulnerabilities.  These vulnerabilities in Virtualization Security include VM Escape, VM Hoping, VM Theft, VM Sprawl, Insecure VM Migration, Sniffing and Spoofing.  Figure 3 illustrates the vulnerabilities of the Virtualization. 

Figure 3.  Vulnerabilities of Virtualization. Adapted from (Gupta et al., 2016).

As indicated in (Gupta et al., 2016), Hypervisor should be inbuilt with the firewall security and disable access console (USB, NIC) to prevent unauthorized access.   The access to the Role Based Access Control (RBAC) is effective to control Hyper jacking of VMs.  The role and responsibilities should be defined to the users of the VMs to check the access authorization. 

Security Principles, Security Mode. Security Models and Security Implementation

As indicated in (Abernathy & McMillan, 2016), the primary goal of all security measures is to provide protection and ensure that the measure is successful.  Three major principles of security include confidentiality, integrity, and availability (CIA).  These Security Principles are known as CIA triad.  The confidentiality is provided if the data cannot be read either through access control and encryption for data as it exists on the hard drive or through encryption as the data is in transit.   Confidentiality is the opposite of “disclosure” (Abernathy & McMillan, 2016).  The Integrity is provided if the data is not changed in any way by unauthorized users.  The integrity principle is provided through the hashing algorithm or a checksum.  The availability principles provide the time the resources or data is available. The availability is measured as a percentage of “up” time with 99.9% of uptime representing more availability than 99% uptime.   The availability principle ensures the availability and access of the data whenever it is needed.  The availability principle is described as a prime goal of security.  Most of the attacks result in a violation of one of these security principles of confidentiality, integrity, or availability.  Thus, the defense-in-depth technique is recommended as an additional layer of security.  For instance, even if the firewall is configured for protection, access control list should still be applied to resources to help prevent access to sensitive data in case the firewall gets breached.  Thus, the defense-in-depth technique is highly recommended.

Security has four major Security Modes which are typically used by the Mandatory Access Control (MAC).  These four security modes include Dedicated Security Mode, System High-Security Mode, Compartmented Security Mode, and Multi-Level Security Mode.  The MAC operates in different security modes at different times based on variables such as sensitivity of data, the clearance level of the user, and the actions users are authorized to take.  In all the four security modes, a non-disclosure agreement (NDA) must be signed, and the access to certain information is based on each mode.

Security Models provide a mapping technique for the security policymakers to the rules which a computer system must follow.  Various types of the Security Models provide various approaches to implement such a mapping technique (Abernathy & McMillan, 2016). 

• State Machine Model,
• Multi-Level Lattice Models, 
• Matrix-Based Models,
• Non-Interface Models, and
• Information Flow Models.

Moreover, there are formal Security Models which are incorporating security concepts and principles to guide the security design of systems. These formal Security Models include the following seven Models (Abernathy & McMillan, 2016).  The detail for each model is beyond the scope of this discussion.

• Bell-LaPadula Model.
• Biba Model.
• Clark-Wilson Integrity Model.
• Lipner Model.
• Brewer-Nash Model.
• Graham-Denning Model.
• Harrison-Ruzzo-Ullman Model.

With respect to the Security Implementation, there are standards which must be followed when implementing security measures for protection.  These standards include ISO/IEC27001 and 27002 and PCI-DSS.   The ISO/IEC27001 is the most popular standards, which is used by the organization to obtain certification for information security.  These standard guides ensure that the information security management system (ISMS) of the organization is properly built, administered, maintained and progressed.  The ISO/IEC 27002 standard provides a code of practice for information security management. This standard includes security measures such as access control, cryptography, compliance.  The PCI-DSS v3.1 is specific for payment card industry. 

Security Models in Cloud Computing

As Service Model is one of the main models in Cloud Computing.  These services are offered through a Service Provider known as a Cloud Service Provider to the cloud users.  Security and privacy are the main challenges and concern when using Cloud Computing environment.  Although there is a demand to leverage the resources of the Cloud Computing to provide services to clients, there is also need and the requirement for the Cloud servers and resources not to learn any sensitive information about the data being managed, stored, or queried (Chaturvedi & Zarger, 2015).   Effort should be exerted to improve the control of users to their data in the public environment.  Cloud Computing Security Models include Multi-Tenancy Model, Cloud Cube Security Model, the Mapping Model of Cloud, Security and Compliance, and the Cloud Risk Accumulation Model of CSA (Chaturvedi & Zarger, 2015).

The Multi-Tenancy Model is described to be the major functional characteristic of Cloud Computing allowing multiple applications to provide cloud services to the clients.  The user’s tenants are separated by virtual partitions, and each partition holds clients tenant’s data, customized settings and configuration settings.  Virtualization in a physical machine allows users to share computing resources such as memory, processor I/O and storage to different users’ applications and amends the utilization of Cloud resources.  SaaS is a good example of Multi-Tenant Model which provides scalability to serve a large number of clients based on Web service.  This model of Multi-Tenancy is described by the security experts to be vulnerable and expose confidentiality which is regarded to be one of the Security Principles to risk between the tenants.  Side channel attack is a significant risk in the Multi-Tenancy Model.  This kind of attack is based on information obtained from bandwidth monitoring.   Another risk of the Multi-Tenancy Model is the assignment of resources to the clients with unknown identity and intentions.  Another security risk associated with Multi-Tenancy involves data storage of multiple tenants in the same database tablespaces or backup tapes. 

The Cloud Cube Security Model is characterized by four main elements; Internal/External, Proprietary/Open, Parameterized/De-parameterized, and Insourced/Outsourced.  The Mapping Model of Cloud, Security, and Compliance Model is another Model to provide a better method to analyze the gaps between cloud architecture and compliance framework and the corresponding security control strategies provided by the Cloud Service Provider, or third parties.  The Cloud Risk Accumulation Model of CSA is the last Security Models of Cloud Computing.  The three Cloud Models of IaaS, PaaS, and SaaS have various security requirements due to the layer dependencies.

Security Implementation: Virtual Private Cloud (VPC)

The VPC Deployment Model is a model that provides more security than the Public Deployment Model.  In this Model, the user can apply Access Control at the instance level as well as at the network level.  Policies are configured and assigned to groups based on the access role.   The VPC as a Deployment Model of the Cloud Computing did solve problems such as the loss of authentication, loss of confidentiality, loss of availability, loss, and corruption of data (Abdul, Jena, Prasad, & Balraju, 2014).  The VPC is logically isolated from other virtual networks in the cloud.  As indicated in (Abdul et al., 2014), VPC is regarded as the most prominent approach to Trusted Computing technology.  However, organizations must implement the security measures based on the requirements of the business.  For instance, organizations and users have control to select the IP address range, create a subnet, route tables, network gateway and security as illustrated in Figure 4.

Figure 4.  Virtual Private Cloud Security Implementation.

In summary, security measures must be implemented to protect the cloud environment.  Virtualization imposes threats to the Cloud environment.  The hypervisor is a major component of Virtualization.  It is recommended that the Hypervisor should be inbuilt with the firewall security and disable access console (USB, NIC) to prevent unauthorized access.   The access to the Role Based Access Control (RBAC) should be effective to control Hyper jacking of VMs.  The role and responsibilities should be defined to the users of the VMs to check the access authorization.  Virtual Private Cloud as a trusted deployment model of the Cloud Computing provides a more secure cloud environment than the Public Cloud. The Security Implementation must follow certain standards.  The organization must comply with these standards to protect organizations and users.

References

Abdul, A. M., Jena, S., Prasad, S. D., & Balraju, M. (2014). Trusted Environment In Virtual Cloud. International Journal of Advanced Research in Computer Science, 5(4).

Abernathy, R., & McMillan, T. (2016). CISSP Cert Guide: Pearson IT Certification.

Ali, M., Khan, S. U., & Vasilakos, A. V. (2015). Security in cloud computing: Opportunities and challenges. Information Sciences, 305, 357-383. doi:10.1016/j.ins.2015.01.025

Chaturvedi, D. A., & Zarger, S. A. (2015). A review of security models in cloud computing and an Innovative approach. International Journal of Computer Trends and Technology (IJCTT), 30(2), 87-92.

Gupta, M., Srivastava, D. K., & Chauhan, D. S. (2016). Security Challenges of Virtualization in Cloud Computing. Paper presented at the Proceedings of the Second International Conference on Information and Communication Technology for Competitive Strategies, Udaipur, India.