PaperlayerView exact revision on arXiv

Source-linked AI summary

A Survey on Cloud Computing Security

Hero Modares, Rosli Salleh, Amirhosein Moravejosharieh, Hassan Keshavarz, Majid Talebi Shahgoli

arXiv:1206.5468v1cs.NIcs.CR

TL;DR

Cloud computing introduces security concerns because virtual machines from different organizations may share physical resources, beyond the protection of conventional perimeter defenses. This paper adopts the NIST working definition, describes cloud service and deployment models, and outlines security concerns. It concludes that physical separation and hardware-based security cannot prevent attacks between virtual machines on the same server.

  • Problem

    Cloud security faces concerns that traditional datacenter perimeter defenses do not address when organizations’ virtual machines share physical servers [3, 4, 11, 13, 15].

    Source: §3 · p. 4

  • Method

    The paper adopts the NIST working definition and uses its cloud qualities, service models, and deployment models to frame its discussion [5, 8].

    Source: 1.1What is Cloud Computing? · p. 2

  • Results

    Physical separation and hardware-based security cannot protect against attacks between virtual machines on the same server.

    Source: §3 · p. 4

  • Takeaways & Limitations

    Cloud security considerations must extend beyond traditional perimeter defenses to risks between co-located virtual machines [3, 4, 11, 13, 15].

    Source: §3 · p. 4

    Private-cloud IT teams may need to invest in buying, creating, and controlling the cloud independently.

    Source: Limitations · p. 3

Abstract

from arXiv · show

Computation encounter the new approach of cloud computing which maybe keeps the world and possibly can prepare all the human's necessities. In other words, cloud computing is the subsequent regular step in the evolution of on-demand information technology services and products. The Cloud is a metaphor for the Internet and is a concept for the covered complicated infrastructure; it also depends on sketching in computer network diagrams. In this paper we will focus on concept of cloud computing, cloud deployment models, cloud security challenges encryption and data protection, privacy and security and data management and movement from grid to cloud.

1. Introduction

Cloud computing separates infrastructure from applications and delivers pooled resources as on-demand services, while raising questions about effects on network and information security [1, 2, 5]. The section adopts the NIST working definition and outlines cloud service models and their characteristics [5, 8].

1. Introduction

  • Cloud computing pools compute, network, information, and storage resources for on-demand allocation and consumption.The section describes cloud services as scalable and available, with potential cost reductions through optimized and efficient computing.Source: §1 · p. 2

1.1What is Cloud Computing?

  • The paper adopts the NIST working definition to provide common language and organize cloud computing around five qualities, three service models, and four deployment models [5, 8].The authors note that choosing this U.S. government source does not exclude other perspectives or geographies.Source: 1.1What is Cloud Computing? · p. 2
  • SaaS supplies standard applications or business services through cloud infrastructure or platforms rather than exposing cloud features directly.Examples include Google Docs, Salesforce CRM, and SAP Business by Design.Source: 1.1What is Cloud Computing? · p. 2
  • PaaS provides resources for developing and hosting applications, but provider-specific APIs can make applications difficult to transfer between hosts [5, 11, 12].The passage notes efforts to broaden common programming models with cloud abilities.Source: 1.1What is Cloud Computing? · p. 2
  • Compute clouds expose typically virtualized computing resources, while IaaS adds capabilities beyond simple compute access [5, 11, 12].The passage distinguishes raw compute resources from PaaS software stacks used to create applications.Source: 1.1What is Cloud Computing? · p. 3

2. Cloud Deployment Models

The section describes public, private, community, and hybrid cloud deployment models, distinguishing them by who can use or combine their infrastructure. Their trade-offs include public-cloud security concerns and the organizational costs of operating a private cloud.

2. Cloud Deployment Models

  • Public clouds are vendor-owned services accessible to the public or large industry groups, generally managed over the Internet and paid for per use.The section notes their use for affordable application development and scalable, reliable applications, while flagging security as important.Source: §2 · p. 3, Advantages · p. 3, Limitations · p. 3
  • Private clouds serve one organization and may be administered by it or a third party, on-premise or off-premise [5, 11, 14].The section associates them with improved server utilization, lower-cost hardware, and reduced costs through automation.Source: Limitations · p. 3
  • Private-cloud IT teams may need to invest in buying, creating, and controlling the cloud independently.Source: Limitations · p. 3
  • Community clouds let multiple entities share smaller-scale infrastructure; the section describes the model as still emerging, with some signs of development [6, 9].The passage also notes similarities between community clouds and GRID technology [10, 11].Source: Limitations · pp. 3–4
  • Hybrid clouds combine two or more private, community, or public clouds as distinct entities linked by technology that supports information and application portability [5, 12].Cloud bursting for load balancing is given as an example.Source: Limitations · p. 4

3. Cloud Deployment Models

Cloud security cannot rely only on conventional perimeter defenses: virtual machines belonging to different organizations may share physical resources, creating security concerns between them.

3. Cloud Deployment Models

  • Physical separation and hardware-based security do not protect against attacks between virtual machines on the same server.Source: §3 · p. 4
  • Cloud providers co-place virtual machines from several organizations on shared physical resources to gain virtualization efficiencies.Source: §3 · p. 4
  • The section frames cloud security as requiring attention beyond the strong network perimeter used in traditional datacenters [3, 4, 11, 13, 15].Source: §3 · p. 4

A. Administrative Access To Servers And Applications

Cloud self-service shifts administrative access from direct or on-premise connections to the Internet, increasing disclosure and risk; the passage stresses limiting and monitoring that access.

A. Administrative Access To Servers And Applications

B. Dynamic Virtual Machines: VM State And Sprawl

Virtual machines can be rapidly reverted, stopped, restarted, cloned, and moved, making consistent security harder to maintain. The paper highlights the difficulty of auditing VM security state and confirming it regardless of location.

B. Dynamic Virtual Machines: VM State And Sprawl

C. Vulnerability Exploits And VM-To-VM Attacks

Cloud servers and localized machines use the same operating systems and applications, whose vulnerabilities can be exploited remotely. Co-located VMs enlarge the attack surface and make VM-level intrusion detection and restraint important.

C. Vulnerability Exploits And VM-To-VM Attacks

D. Data Integrity: Co-Location, Compromise And Theft

The section links shared cloud infrastructure with increased security concerns and calls for auditable verification that resources remain uncompromised. It also recommends recording operating-system and application files and activities.

D. Data Integrity: Co-Location, Compromise And Theft

E. Encryption And Data Protection

Encryption requirements for sensitive data and compliance are heightened in multi-tenant clouds, where credential availability and security pose challenges. Data management concerns include consistency, uncontrolled scattering, interoperability, and scaling access across many nodes.

E. Encryption And Data Protection

  • PCI DSS and HIPAA requirements call for encryption of crucial data such as cardholder data and PII, while multi-tenancy heightens challenges for encryption-credential availability and security.Source: E. Encryption And Data Protection · p. 5

4. Data Management

  • Growing web data and application throughput make storage and bandwidth management important, while maintaining consistency across replicated or concurrently accessed tenant data grows more complex.Transactional assurance for multi-tier software stacks remains a major analysis gap, including at scale without data separation or moderate consistency [11].Source: §4 · p. 5
  • Uncontrolled data partitioning and scattering can create legislative, proficiency, and reintegration problems, prompting a need for greater infrastructure control to support environment analysis and QoS.Source: §4 · p. 5
  • Heterogeneous, unstructured data requires new approaches to separation and information operation; transferring data or services between clouds can be difficult and calls for standards and long-term interoperability.The passage identifies XDR as relevant to systems coupled insecurely and notes shared data across tenants.Source: §4 · p. 5
  • At large node counts, traditional data management and classical transaction control face scaling and disk-delay problems, while cache coherency and storage approaches remain concerns.The passage notes that only a few studies examine cloud file and data access requirements [7].Source: §4 · p. 6

5. Privacy & Security

Cloud security concerns arise from shared, potentially location-unknown resources and from security models that may not fit diverse cloud services. Data must be protected while remaining controllable and compatible with legal requirements.

5. Privacy & Security

  • Shared resources and potentially undefined data locations raise protection concerns, including legislative issues around data location and the need for system control.The passage also notes that information about an industry's use of cloud infrastructure may suffice for industrial espionage [3, 11].Source: §5 · p. 6
  • Cloud diversity can make classical authentication insufficient to distinguish aggregators or vendors from real users, particularly in IaaS.Source: §5 · p. 6
  • Collected or resold cloud systems may combine security procedures incompatibly and leave users without insight into the model, motivating new security management models and operations [3, 4].Source: §5 · p. 6

6. Movement from Grid to Cloud

The passage presents European Grid vendors as potential cloud providers because their business motivations and infrastructure requirements resemble those of cloud providers. It frames the transition around business development, its necessity, and how to manage implementation difficulties.

6. Movement from Grid to Cloud

  • Similar business motivations and infrastructure requirements make it relatively easy for European Grid vendors to move toward cloud provision, including protective devices and middleware.The passage describes the European market, especially in Grid, as ready to acquire cloud services despite slower industrial cloud progress and less available infrastructure than the US.Source: §6 · p. 6
  • The passage identifies existing Grid dealers and adopters in Europe as a substantial group relevant to the transition.Source: §6 · p. 6
  • Grid vendors are asked to clarify how the move can develop their business, why it is necessary, and how to accomplish it and overcome difficulties [11].Source: §6 · p. 6

7. Conclusion

The conclusion presents cloud computing as increasingly popular while identifying security as a major unresolved research concern. It also stresses legal and security risks that businesses must address when choosing and using cloud services.

7. Conclusion

  • Security remains a major unresolved concern in a public literature the passage describes as insufficient, alongside other unanswered cloud-computing issues.Source: §7 · p. 7
  • The passage describes pay-as-you-go use as reducing capital expenditure and new-system and protection costs, and calls for useful policies at local and international levels as cloud use grows.Source: §7 · p. 7
  • Cloud risks include compliance, bandwidth accessibility, global execution, intellectual-property violation, client-data use and location, transparency, and legal disputes.Source: §7 · p. 7
  • Businesses require certainty and assurance about service providers and flexibility to use services amid changing legal challenges.Source: §7 · p. 7
Loading 1206.5468v1…