Security Models
What Is Cloud Computing?
Cloud computing is the delivery of computing services — including servers, storage, databases, networking, software, analytics, and intelligence — over the internet (‘the cloud’) to offer faster innovation, flexible resources, and economies of scale. Rather than owning and maintaining physical data centers and servers, organizations can access technology services on an as-needed basis from a cloud provider.
Core Characteristics (NIST SP 800-145)
- On-demand self-service — consumers provision computing capabilities without requiring human interaction with each service provider.
- Broad network access — capabilities are available over the network and accessed through standard mechanisms.
- Resource pooling — provider resources serve multiple consumers using a multi-tenant model.
- Rapid elasticity — capabilities can be elastically provisioned and released to scale with demand.
- Measured service — cloud systems automatically control and optimize resource use by leveraging metering capabilities.
Why Cloud Security Matters
Organizations are migrating critical workloads, intellectual property, and sensitive customer data to cloud environments at an accelerating pace. With this migration comes a fundamental shift in security responsibilities and threat exposure. The 2024 Cloud Security Alliance (CSA) State of Cloud Security report found that 81% of organizations experienced at least one cloud security incident in the prior 12 months, highlighting the critical importance of understanding cloud security models.
Key Security Differences: Cloud vs. On-Premises
| Dimension | Key Distinction |
|---|---|
| Perimeter | Traditional network perimeter dissolves; identity becomes the new perimeter in cloud. |
| Control | Physical hardware control is relinquished; logical controls and APIs become primary security mechanisms. |
| Visibility | Log aggregation and monitoring require deliberate architecture; native cloud tools essential. |
| Elasticity | Attack surface expands and contracts dynamically; security must scale automatically. |
| Multi-tenancy | Shared underlying infrastructure introduces data isolation and side-channel attack concerns. |
| Configuration | Misconfiguration is the leading cause of cloud breaches; security posture management critical. |
Major Cloud Service Providers
Three hyperscalers dominate the global cloud market and each maintains extensive security programs, certifications, and tools:
| Provider | Security Platform / Tools | Key Security Certifications |
|---|---|---|
| AWS | AWS Security Hub, GuardDuty, Inspector, Macie, IAM, CloudTrail, Shield | FedRAMP, ISO 27001, SOC 1/2/3, PCI DSS, HIPAA, DoD IL2–IL6 |
| Microsoft Azure | Microsoft Defender for Cloud, Sentinel, Azure Policy, Entra ID, Monitor | FedRAMP High, ISO 27001/27018, SOC 1/2/3, HIPAA, HITRUST |
| Google Cloud | Security Command Center, Chronicle SIEM, Cloud Armor, BeyondCorp, IAP | FedRAMP High, ISO 27001, SOC 1/2/3, PCI DSS, HIPAA |
Infrastructure as a Service (IaaS)
IaaS provides virtualized computing resources over the internet. The provider manages physical hardware, networking, and virtualization. The customer controls operating systems, middleware, runtime, data, and applications.
IaaS Security Responsibilities (Customer)
- Operating system hardening, patching, and lifecycle management
- Network security groups, firewall rules, and virtual network configuration
- Identity and access management for workloads and administrators
- Endpoint protection and anti-malware on virtual machines
- Data encryption at rest and in transit
- Backup, recovery, and business continuity planning
Platform as a Service (PaaS)
PaaS provides a platform allowing customers to develop, run, and manage applications without managing the underlying infrastructure. The provider manages the runtime, middleware, OS, virtualization, servers, storage, and networking.
PaaS Security Responsibilities (Customer)
- Application code security and secure development lifecycle (SDLC)
- Application-level access controls and authentication mechanisms
- Input validation and protection against OWASP Top 10 vulnerabilities
- Secrets management (API keys, connection strings, certificates)
- Data classification and protection within the application
PaaS Security Considerations (Shared)
- Platform API security and service endpoint protection
- Dependency and library vulnerability management
- Logging, monitoring, and alerting integration
Software as a Service (SaaS)
SaaS delivers software applications over the internet, on-demand and typically on a subscription basis. The provider manages the entire stack from infrastructure through the application; the customer primarily manages data, user access, and configuration.
SaaS Security Responsibilities (Customer)
- User provisioning, deprovisioning, and access governance
- Data governance — what data is stored, how it is classified, and retention policies
- Single Sign-On (SSO) and Multi-Factor Authentication (MFA) configuration
- Third-party integration and OAuth permission scoping
- Data Loss Prevention (DLP) policy configuration where available
- Vendor risk assessment and supply chain security review
Shared Responsibility Model
The shared responsibility model is the foundational security framework in cloud computing. It delineates which security obligations belong to the cloud provider and which belong to the customer. Failure to understand this boundary is one of the most significant causes of cloud security incidents.
| Responsibility Area | On-Premises | IaaS | PaaS | SaaS |
|---|---|---|---|---|
| Data & Content | Customer | Customer | Customer | Customer |
| Applications | Customer | Customer | Customer / Shared | Provider |
| Runtime / Middleware | Customer | Customer | Provider | Provider |
| Operating System | Customer | Customer | Provider | Provider |
| Virtualization | Customer | Provider | Provider | Provider |
| Servers / Storage / Net. | Customer | Provider | Provider | Provider |
| Physical Security | Customer | Provider | Provider | Provider |
Identity as the New Perimeter
In traditional IT, the network perimeter (firewalls, DMZs) provided a clear security boundary. Cloud computing dissolves this perimeter — resources are accessed from anywhere, by any device, over the public internet. Identity becomes the primary control plane. Compromised credentials are now the leading initial attack vector in cloud breaches.
Core IAM Concepts
| Concept | Definition |
|---|---|
| Authentication (AuthN) | Verifying the identity of a user or system. Answers: ‘Who are you?’ |
| Authorization (AuthZ) | Determining what an authenticated identity is permitted to do. Answers: ‘What can you do?’ |
| Principle of Least Privilege | Grant only the minimum permissions required to perform a specific task. |
| Zero Trust | Never implicitly trust; always verify — regardless of network location or prior authentication. |
| Federation | Allowing external identity providers (e.g., Azure AD, Okta) to authenticate users for cloud services. |
| Service Account / Role | Non-human identity assigned to applications or services to access cloud resources. |
Multi-Factor Authentication (MFA)
MFA requires users to provide two or more verification factors to gain access. It is one of the single most effective controls against credential-based attacks. Industry data consistently shows MFA blocks over 99% of automated account attacks.
MFA Factor Types
- Something you know — password, PIN, security question
- Something you have — hardware token (YubiKey), TOTP app (Google Authenticator, Authy), SMS code
- Something you are — fingerprint, facial recognition, iris scan (biometrics)
Common IAM Misconfigurations
IAM misconfigurations are consistently the leading source of cloud security incidents. The following are the most prevalent:
-
Overly permissive policies — use of wildcards (*) in IAM policies granting excessive permissions
Mitigation: Regular access reviews, AWS IAM Access Analyzer, Azure Permissions Management
-
Unused credentials — dormant accounts, old API keys, unused roles still attached to resources
Mitigation: Credential rotation enforcement, lifecycle management, automated deprovisioning
-
Missing MFA on root/administrator accounts
Mitigation: Enforce MFA via Conditional Access policies or Service Control Policies (SCPs)
-
Publicly exposed cloud storage — S3 buckets, Azure Blob containers without ACL controls
Mitigation: Block public access at the account/organization level, enable CSPM scanning
-
Hard-coded credentials in source code or container images
Mitigation: Secrets management tools (AWS Secrets Manager, Azure Key Vault, HashiCorp Vault)
Top Cloud Security Threats
The Cloud Security Alliance (CSA) Pandemic Eleven identifies the most critical threats facing cloud environments:
| # | Threat | Description | Severity |
|---|---|---|---|
| 1 | Insufficient IAM | Weak or misconfigured identity controls allowing unauthorized access to cloud resources. | Critical |
| 2 | Insecure Interfaces/APIs | Exposed or poorly secured APIs that can be exploited to access or manipulate cloud resources. | Critical |
| 3 | Misconfiguration | Improperly configured cloud services exposing data or enabling lateral movement. | Critical |
| 4 | Lack of Cloud Architecture Strategy | Ad-hoc cloud adoption without a governance or security architecture framework. | High |
| 5 | Account Hijacking | Credential theft enabling attackers to take over cloud accounts and access all associated resources. | Critical |
| 6 | Insider Threats | Malicious or negligent employees with cloud access causing data theft or sabotage. | High |
| 7 | Advanced Persistent Threats | Sophisticated attackers establishing persistent footholds within cloud environments. | High |
| 8 | Data Breaches | Unauthorized access to sensitive data stored in cloud services. | Critical |
| 9 | Limited Cloud Visibility | Insufficient monitoring and logging preventing detection and response to incidents. | High |
| 10 | Abuse of Cloud Services | Criminals using cloud resources for cryptojacking, phishing infrastructure, malware hosting. | Medium |
| 11 | Supply Chain Vulnerabilities | Compromised third-party services, libraries, or providers used within cloud environments. | High |
Cloud-Specific Attack Techniques
Attackers compromise cloud accounts to deploy cryptocurrency mining software, running at the victim’s expense. This is often the first malicious action after IAM credential theft due to ease of execution and financial gain.
- Detection: Sudden spike in compute costs, unexpected EC2/VM instances, unusual API calls (RunInstances, CreateVM)
- Prevention: Budget alerts, cost anomaly detection, least-privilege IAM for instance creation
SSRF exploits vulnerable web applications to make server-side HTTP requests to internal resources. In cloud environments, this is particularly dangerous because the Instance Metadata Service (IMDS) is accessible at 169.254.169.254 and can return temporary credentials.
- Detection: Unusual HTTP requests to metadata IPs, credential usage from unexpected locations
- Prevention: IMDSv2 enforcement (token-required requests), WAF rules blocking metadata IP access from applications
Once an attacker gains access to one cloud resource, they use it as a pivot point to access additional resources, accounts, or connected on-premises systems.
- Detection: Unusual AssumeRole calls, cross-account access from unexpected sources, CloudTrail anomalies
- Prevention: Strict trust policies for cross-account roles, AWS Organizations SCPs, conditional access policies
Key Compliance Frameworks
NIST Cybersecurity Framework (CSF) 2.0
The NIST CSF provides a voluntary framework consisting of standards, guidelines, and best practices to manage cybersecurity risk. The 2.0 update adds a Govern function:
- GOVERN — Establish and monitor organizational cybersecurity strategy, expectations, and policy
- IDENTIFY — Develop understanding of systems, assets, data, and risks
- PROTECT — Implement appropriate safeguards to ensure delivery of services
- DETECT — Implement activities to identify cybersecurity events
- RESPOND — Take action regarding detected cybersecurity incidents
- RECOVER — Maintain resilience and restore capabilities affected by incidents
ISO/IEC 27001 & 27017
ISO 27001 is the international standard for Information Security Management Systems (ISMS). ISO 27017 extends 27001 with cloud-specific controls, providing additional guidance on:
- Shared roles and responsibilities between cloud provider and customer
- Removal and return of assets at contract termination
- Protection and separation of virtual environments
- Virtual machine hardening
- Administrative operations and procedures in cloud environments
SOC 2 Type II
SOC 2 is an auditing procedure developed by the AICPA that ensures service providers securely manage data to protect the interests of the organization and the privacy of its clients. Type II reports cover operational effectiveness over a period (typically 6–12 months) across five Trust Service Criteria: Security, Availability, Processing Integrity, Confidentiality, and Privacy.
FedRAMP
The Federal Risk and Authorization Management Program (FedRAMP) provides a standardized approach to security assessment, authorization, and continuous monitoring for cloud products and services used by U.S. federal agencies. FedRAMP authorizations (Low, Moderate, High) signal the sensitivity level of federal data the system may host.
Compliance Mapping to Cloud Controls
| Control Domain | NIST CSF 2.0 | ISO 27001/27017 | SOC 2 TSC |
|---|---|---|---|
| Access Control | PR.AA (Protect: Identity Mgmt) | A.9 / CLD.9 | CC6.1 – Logical Access |
| Incident Response | RS.MA (Respond: Mgmt) | A.16 | CC7.3 – Incident Response |
| Encryption | PR.DS (Protect: Data Security) | A.10 / CLD.13 | CC6.7 – Data in Transit/Rest |
| Vulnerability Mgmt | ID.RA (Identify: Risk Assessment) | A.12.6 | CC7.1 – Vulnerability Mgmt |
| Logging & Monitoring | DE.AE (Detect: Anomalies) | A.12.4 / CLD.12.4 | CC7.2 – Monitoring |
| Change Management | PR.IP (Protect: Processes) | A.12.1 | CC8.1 – Change Management |
Defense-in-Depth for Cloud
Defense-in-depth applies multiple layers of security controls across the cloud stack, ensuring that a failure in one layer does not lead to a total compromise.
| Layer | Controls & Best Practices |
|---|---|
| Data | Encryption at rest (AES-256) and in transit (TLS 1.2+), DLP policies, data classification, key management (HSM/KMS) |
| Application | Secure SDLC, DAST/SAST scanning, WAF, API gateway with auth, dependency scanning, container image signing |
| Identity | MFA, least privilege IAM, PAM solution, just-in-time access, federated SSO, privileged identity management |
| Endpoint / Workload | EDR on VMs, container security scanning, runtime protection, patch management, hardened OS baselines |
| Network | VPC/VNET segmentation, security groups, NACLs, Private Link/Endpoints, micro-segmentation, DDoS protection |
| Infrastructure | CSPM tool, CIS Benchmarks, Infrastructure as Code (IaC) scanning, immutable infrastructure patterns |
| Physical | Provider responsibility — verify via certifications (SOC 2, ISO 27001), site audit rights, compliance attestations |
Cloud Security Posture Management (CSPM)
CSPM tools continuously monitor cloud environments for security misconfigurations, compliance violations, and risky exposure. They are a critical component of any mature cloud security program.
Core CSPM Capabilities
- Continuous visibility across multi-cloud environments (AWS, Azure, GCP)
- Automated compliance checking against CIS Benchmarks, NIST, PCI DSS, HIPAA
- Detection and remediation of misconfigured resources (open S3 buckets, exposed ports, public snapshots)
- Security score / cloud security benchmark reporting
- Auto-remediation workflows via Lambda functions, Azure Functions, or playbooks
Data Protection & Encryption
Encryption Key Management
Who controls the encryption keys determines who ultimately controls access to your data. Three key management models exist in cloud:
Cloud Incident Response
Cloud incident response requires significant adaptations from traditional IR processes. Speed of containment is amplified by cloud’s API-driven control plane.
Cloud IR Playbook — Core Steps
- PreparationMaintain IR runbooks, pre-authorize response roles, configure CloudTrail/Activity Log retention, test playbooks quarterly
- Detection & AnalysisUse SIEM/cloud-native detection, triage alerts by severity, preserve logs immediately (logs may auto-expire)
- ContainmentIsolate compromised resources (security group deny-all), revoke IAM credentials, snapshot affected instances for forensics
- EradicationRemove malware, terminate unauthorized resources, rotate all potentially exposed credentials, patch vulnerabilities
- RecoveryRestore from known-good snapshots, verify integrity, re-enable services incrementally, monitor closely
- Post-Incident ActivityRoot cause analysis, lessons learned, update controls and runbooks, report per regulatory requirements
| Term | Definition |
|---|---|
| CASB | Cloud Access Security Broker — security policy enforcement point between users and cloud service providers. |
| CNAPP | Cloud-Native Application Protection Platform — unified platform combining CSPM, CWPP, and CIEM capabilities. |
| CSPM | Cloud Security Posture Management — tools that identify misconfiguration and compliance risks in cloud environments. |
| CWPP | Cloud Workload Protection Platform — security for workloads (VMs, containers, serverless) running in cloud. |
| IAM | Identity and Access Management — framework of policies and controls for managing digital identities and access. |
| IaC | Infrastructure as Code — provisioning infrastructure through machine-readable configuration files rather than manual processes. |
| IMDSv2 | Instance Metadata Service v2 — session-oriented method to access EC2 metadata that mitigates SSRF attacks. |
| KMS | Key Management Service — managed service for creating and controlling cryptographic keys for data encryption. |
| MFA | Multi-Factor Authentication — authentication method requiring two or more verification factors. |
| SIEM | Security Information and Event Management — tool for real-time analysis of security alerts from applications and hardware. |
| SCP | Service Control Policy — in AWS Organizations, guardrails limiting actions available to accounts within the organization. |
| SSRF | Server-Side Request Forgery — attack where an attacker causes the server to make HTTP requests to unintended locations. |
| VPC | Virtual Private Cloud — logically isolated virtual network within a cloud provider’s infrastructure. |
| Zero Trust | Security model that assumes no implicit trust; requires verification of every user and device attempting access. |
References & Further Reading
- NIST SP 800-145: The NIST Definition of Cloud Computing
- NIST SP 800-144: Guidelines on Security and Privacy in Public Cloud Computing
- NIST Cybersecurity Framework 2.0 — csrc.nist.gov/projects/cybersecurity-framework
- CSA Cloud Controls Matrix (CCM) v4 — cloudsecurityalliance.org
- CSA Top Threats to Cloud Computing: Pandemic Eleven
- ISO/IEC 27001:2022 — Information Security Management Systems
- ISO/IEC 27017:2015 — Code of Practice for Information Security Controls for Cloud Services
- CIS Benchmarks for AWS, Azure, and GCP — cisecurity.org/cis-benchmarks
- MITRE ATT&CK Cloud Matrix — attack.mitre.org/matrices/enterprise/cloud
- FedRAMP Program Documentation — fedramp.gov
- AWS Well-Architected Framework: Security Pillar
- Microsoft Azure Security Benchmark v3
- Google Cloud Security Foundations Guide