Source-linked AI summary
An Analysis of the Cloud Computing Security Problem
Mohamed Almorsy, John Grundy, Ingo Müller
TL;DR
Cloud computing’s adoption is constrained by security concerns that become more complex through multi-tenancy, elasticity, architecture, and layer dependencies. The paper analyzes these dimensions across cloud perspectives and derives a detailed security-problem specification and required solution features. Its conclusion identifies isolation, security management, and cross-layer control coordination as central requirements.
Problem
Cloud security remains a major adoption concern because multi-tenancy, elasticity, architecture, and dependent layers introduce additional security dimensions.
Method
The paper analyzes cloud security from architecture, characteristics, stakeholders, and service-delivery-model perspectives.
Results
The analysis derives a detailed specification of the cloud security problem and identifies key features that proposed security solutions should cover.
Takeaways & Limitations
Cloud security requires tenant isolation, coordinated management of heterogeneous controls, and protection spanning the cloud’s dependent layers.
Abstract
from arXiv · showhide
Cloud computing is a new computational paradigm that offers an innovative business model for organizations to adopt IT without upfront investment. Despite the potential gains achieved from the cloud computing, the model security is still questionable which impacts the cloud model adoption. The security problem becomes more complicated under the cloud model as new dimensions have entered into the problem scope related to the model architecture, multi-tenancy, elasticity, and layers dependency stack. In this paper we introduce a detailed analysis of the cloud security problem. We investigated the problem from the cloud architecture perspective, the cloud offered characteristics perspective, the cloud stakeholders' perspective, and the cloud service delivery models perspective. Based on this analysis we derive a detailed specification of the cloud security problem and key features that should be covered by any proposed security solution.
I. INTRODUCTION
Cloud computing offers scalable, on-demand IT services without major upfront investment, but security concerns remain a major barrier to adoption. The paper examines these concerns across architecture, characteristics, stakeholders, and delivery models to specify requirements for cloud security solutions.
- Cloud computing provides scalable, internet-based computing resources as services that can be rapidly provisioned and released with minimal management effort.
- The model supports IT adoption without upfront infrastructure and software investments, benefiting small and medium businesses and attracting government interest.
- USD 58.6B in 2009, USD 68B in 2010, and USD 148B by 2014 illustrate rapid cloud-market growth and increased attacker interest.
- Security is the major consumer concern because outsourcing reduces control, co-locates tenants, weakens SLA guarantees, and exposes assets to public infrastructure.
- The paper analyzes cloud security across multiple perspectives and derives dimensions that proposed security solutions should cover.
II. LITERATURE REVIEW
The literature review identifies prior analyses of cloud use cases, risks, threats, and service models. It also introduces the three deployment models and three delivery models whose varied implementations complicate standardized security management.
- Prior work examines cloud use cases and requirements from customer, developer, and security-engineer perspectives.
- ENISA analyzes cloud risks together with affected assets, likelihood, impacts, and vulnerabilities.
- Cloud deployment models are private, public, and hybrid, with public clouds described as the most vulnerable because malicious users can host services there.
- IaaS provides computation, storage, and network resources; PaaS provides development and deployment platforms; SaaS provides hosted applications.
- Different implementations and coexistence of delivery models complicate development of standard security models and security management.
IV. CLOUD COMPUTING CHARACTERSTICS AND SECURITY IMPLICATIONS
Multi-tenancy and elasticity improve resource utilization but introduce confidentiality and isolation risks. The paper emphasizes secure tenant separation and migration of security requirements as resources change.
- Multi-tenancy shares computational resources, storage, services, and applications among tenants to improve resource utilization and reduce cost.
- Approaches 3 and 4 are identified as most risky because tenants coexist within the same process, memory, and hardware.
- Resource sharing can violate tenant confidentiality, requiring isolation among tenants’ data and other resources.
- Elasticity reallocates resources as demand changes, creating a risk that a new tenant may deduce contents previously held by another tenant.
- Security requirements should migrate with services and be enforced in newly assigned environments during resource placement.
V. CLOUD COMPUTING’S DEEP DEPENDENNCIES STACK
Cloud security spans a deep stack of dependent infrastructure, virtualization, platform, API, service, and application layers. Breaches or inconsistent controls at lower layers can affect the whole platform, motivating unified control management.
- Cloud functionality and security depend across layers including physical infrastructure, virtualization, virtual resources, platforms, APIs, services, and applications.
- A breach of any cloud object can affect the whole platform because each layer’s security depends on lower layers.
- Heterogeneous requirements and controls may conflict across layers, producing an inconsistent security model.
- The paper therefore calls for a unified security-control management module to coordinate controls across layers according to security needs.
VI. CLOUD COMPUTING STAKEHOLDERS AND SECURITY IMPLICATIONS
Cloud stakeholders have distinct security responsibilities, requirements, and capabilities, creating conflicts and complicating runtime enforcement and transparency across hosted services.
- Cloud providers, service providers, and service consumers each maintain security-management processes with distinct expectations and capabilities.
- Different tenants may define conflicting security requirements for the same service, requiring configurations to be enforced per instance and at runtime.
- Providers cannot deliver fully effective controls without knowing hosted-service architectures, while changing requirements complicate administration and control updates.
VII. CLOUD COMPUTING SERVICE DELIVERY MODELS AND SECURITY IMPLICATIONS
Security responsibilities and threats differ across IaaS and PaaS delivery models, spanning virtual machines, shared infrastructure, SOA services, and management APIs.
- Cloud providers and consumers share responsibility for security issues across service delivery models, with responsibilities varying by issue.
- A. IaaS Issues: IaaS consumers secure VM operating systems and workloads, while providers secure VM image repositories, virtual boundaries, and shared infrastructure.
- A. IaaS Issues: Offline VM images can be compromised or stolen, and templates may retain previous owners’ information for later consumers.
- B. PaaS Security Issues: PaaS inherits SOA threats including denial-of-service, man-in-the-middle, XML, replay, dictionary, injection, and input-validation attacks.
- B. PaaS Security Issues: PaaS APIs require authentication, authorization, security standards, and memory isolation, with API security assigned to service providers.
C. SaaS Security Issues
SaaS security is shared between cloud and software providers and inherits data, network, and web-application security concerns from underlying models.
- SaaS security enforcement and maintenance are shared by cloud providers and software vendors.
- SaaS inherits data-security requirements covering locality, integrity, segregation, access, confidentiality, and backups, along with network security.
- Cloud-hosted web applications should be scanned for vulnerabilities using current vulnerability and attack-path databases.
- Web application firewalls should mitigate discovered vulnerabilities by examining HTTP traffic.
- The Cloud Management Layer coordinates services and security management, making breaches capable of giving attackers administrator-like control over the platform.
E. Cloud Access Methods Security Issues
Cloud resources are accessed through distinct protocols and interfaces, so security controls must protect communications and manage identities across interacting domains.
- SaaS, PaaS, and IaaS expose resources through browsers, service protocols and APIs, or remote connections, VPN, and FTP.
- Security controls should target vulnerabilities in these access protocols to protect data transferred between cloud platforms and consumers.
- Identity systems recognize users, services, servers, clouds, and other entities while avoiding disclosure of personal information.
- Cloud identity management includes provisioning, federation, single sign-on, authentication, and authorization.
- SPML, SAML, OAuth, and XACML are identified as standards for securely federating identities across domains and cloud platforms.
B. Key Management
Cloud security depends on managing encryption keys across data, processes, communications, and PaaS APIs. The paper also situates key management within broader security requirements, controls, and lifecycle activities.
- Key management: Encryption supports confidentiality for cloud data, processes, and communications, but both symmetric and asymmetric approaches require secure key management.Key management covers securely generating, storing, accessing, and exchanging secret keys.
- Key management: PaaS applications require keys for APIs and service calls, expanding key-management needs beyond stored data.
- Security management: Cloud security management must coordinate security requirements, policy-based control configurations, and feedback from the environment and controls.The paper proposes this management function as a plug-in for CML.
- Secure development: Secure cloud software development includes security-requirement elicitation, threat modeling, and augmenting system models and generated code with security requirements.PaaS can provide reusable security-enabling components for developing secured cloud applications.
E. Security-Performance tradeoff optimization
Cloud security decisions must balance stronger protection against performance costs, while supporting adaptive configurations and federated enforcement across integrated clouds.
- Security-performance tradeoff: Higher security levels consume more resources and can affect performance, so cloud management must address the security-performance trade-off.The paper proposes using utility functions to balance the two objectives.
- Security-performance tradeoff: Adaptive security configures controls according to current and expected threat levels while considering other trade-offs.
- Federated security: Applications spanning multiple clouds require security requirements to remain enforced on each cloud and across the connections between them.
- Federated security: Federated cloud platforms must enforce shared security requirements when clouds integrate to provide combined resources or services.
IX. CONCLUSION
The conclusion characterizes cloud security as a layered, multi-tenant management problem requiring holistic modeling and enforcement. The proposed direction is an adaptive model-based approach that captures requirements, maps them to controls, and reports security status.
- IX. CONCLUSION: Cloud security inherits problems from technologies such as virtualization and service-oriented architecture.
- IX. CONCLUSION: Multi-tenancy and isolation require a vertical security solution spanning the SaaS layer through physical infrastructure.The paper calls for physical-like tenant boundaries rather than only virtual boundaries.
- IX. CONCLUSION: Security management is critical because the cloud model combines many security requirements and controls.
- IX. CONCLUSION: The cloud model should use a holistic security wrapper so access to cloud-platform objects passes through security components first.
- IX. CONCLUSION: The proposed adaptive model-based approach captures stakeholder requirements, maps them to architecture, patterns, and enforcement mechanisms, and provides current security-status feedback.