Module 4 (Chapter 2)

CLOUD SECURITY AND TRUST MANAGEMENT

Prepared by,
Swathi Jk & MahaLakshmi R
Dept of CSE,CBIT-KOLAR
4.6 CLOUD SECURITY AND TRUST MANAGEMENT

Why people don’t fully trust cloud computing

•In cloud computing, users give their data and applications to service providers (like companies running cloud
servers).
•Many users feel uncomfortable because:
• They don’t know how their data is handled
• They fear privacy loss
• They worry about security attacks
• They are unsure about copyright/data ownership
So, lack of trust is a big reason why cloud computing is not accepted by everyone.

• Technology can help people trust online services.

• In cloud computing, since everything is online and virtual, it’s easier for hackers to attack than on regular
computers.
• To make cloud computing safer:
•Use new ways to protect data (data-protection models)
•Use ideas from P2P networks and grid systems to increase trust and security in cloud services.
4.6.1 Cloud Security Defense Strategies,(methods and measures used to protect
cloud systems, data, and users from threats and attacks).
• A healthy cloud ecosystem means a cloud environment where users are protected from bad things like:
• Hacking and viruses
• Cheating or fraud
• Spam, rumors, and pornography
• Privacy and copyright violations
• Different types of cloud services—IaaS (Infrastructure), PaaS (Platform), and SaaS (Software)—have their own
security needs.
• These are usually defined in agreements (SLAs) between the cloud provider and the user.

Examples of cloud security strategies:

[Link] – block unauthorized access
[Link] – hide data so only authorized users can read it
[Link] control – decide who can use what
[Link] & recovery – save copies of data in case of loss
[Link] & alerts – watch for unusual activity
[Link] Basic Cloud Security

Cloud security has three main levels:

[Link] Security (Data Center Security)
1. Protects the physical location of servers
2. Uses things like biometric scanners, CCTV cameras, motion detectors, and man traps
[Link] Security
1. Protects the cloud network from attacks
2. Uses firewalls, intrusion detection systems (IDS), and third-party vulnerability checks
[Link] Security
1. Protects the applications and data in the cloud
2. Uses SSL encryption, strict passwords, and system trust certifications
Figure 4.31 shows how different cloud service models (like IaaS, PaaS, SaaS) have different security measures applied
at each level.

Left side: Cloud Service Models

The left part shows the three main cloud service models:

[Link] (Infrastructure as a Service) – innermost layer

1. Includes: Hardware, Virtualization, APIs,
Connectivity
2. Provides the basic infrastructure (servers, storage,
network)
[Link] (Platform as a Service) – middle layer
1. Includes: Integration, OS, Middleware
2. Provides a platform for developers to build and deploy
apps
[Link] (Software as a Service) – outermost layer
1. Includes: Applications, APIs, Data, Content,
Metadata
2. Provides ready-to-use applications to end-users
2️.Right side: Security, Privacy, and Copyright Measures
The right part shows what security measures are needed at each
level: Working:
•Applications (SaaS level) Each cloud service layer has its specific security needs.
• Binary analysis, WebApp firewalls, transactional security, • IaaS focuses on hardware and infrastructure security.
copyright protection • PaaS focuses on platform and management security.
•Data/Information • SaaS focuses on application and data security.
• Data loss protection, encryption, monitoring, watermarking,
database logs
•Management
• Risk management, identity & access control, virtual machine
Note:
security, patch management
The diagram shows which security measures are needed at
•Networking
each cloud layer—from the physical servers (IaaS) to the
• Firewalls, intrusion detection/prevention (IDS/IPS), Anti-
applications users use (SaaS). Each layer has different
DDoS, DNSSEC, Quality of Service
security responsibilities.
•Trusted Computing
• Root of trust (RoT), secure APIs, trust and reputation systems
•Compute and Storage (IaaS level)
• Host-based firewalls, IDS/IPS, encryption, integrity checks,
masking
In the cloud:
•Servers can be physical machines or virtual machines (VMs).
•Users request services through interfaces, and provisioning tools allocate resources to fulfill those requests.

Clouds must have security built into their design because attacks like worms, viruses, malware, and DDoS can
exploit weaknesses and disrupt services or steal data.
Here are some cloud components that demand special security protection:
• Protection of servers from malicious software attacks such as worms, viruses, and malware
• Protection of hypervisors or VM monitors from software-based attacks and vulnerabilities
• Protection of VMs and monitors from service disruption and DoS attack side intruders unauthorized access to
critical information.
• Protection of data and information from theft, corruption, and natural disasters
• Providing authenticated and authorized access to critical data and services
[Link] Security Challenges in VMs
1. Traditional vs. Cloud-Specific Attacks
•Traditional network attacks:
• Buffer overflows, DoS (Denial of Service), spyware, malware, rootkits, Trojans, worms.
• These are threats that target computers and networks in general.
•Cloud/VM-specific attacks:
• Hypervisor malware: Malicious code targeting the hypervisor (the layer managing VMs).
• Guest hopping: One VM attacking another VM on the same host.
• VM hijacking & rootkits: Attackers take control of a VM or install stealth malware.
• Man-in-the-middle during VM migration: Intercepting or manipulating VM data while it moves between hosts.

2. Attack Types
•Passive attacks:
• Steal data or credentials without changing system behavior.
• Example: Sniffing passwords from VM network traffic.
•Active attacks:
• Directly manipulate system/kernel data.
• Can cause crashes, data corruption, or compromise multiple VMs.
3. Defense Mechani[Link] mechanisms are the strategies, tools, and techniques used to protect VMs and the cloud
environment from attacks).

•Intrusion Detection Systems (IDS):

• NIDS (Network-based IDS): Monitors network traffic.
• HIDS (Host-based IDS): Monitors VM or host system activity.

•Program Shepherding:
• Controls and verifies execution of code to prevent unauthorized actions.

•Dynamic Optimization & Virtualization Tools:

• RIO infrastructure: Monitors and optimizes code execution for security.
• VMware vSafe & vShield: VMware tools to enhance VM security.

•Hardware/Hypervisor-level security:
• Intel vPro technology: Hardware-assisted security for VMs.
• Security compliance and hardening of hypervisors.

•OS-level defenses:
• Hardened operating systems, sandboxing, and isolated execution environments to contain threats.
[Link] Cloud Defense Methods

Table 4.9 lists eight protection schemes to secure public clouds and data centers.

1. Secure data centers and computer buildings:

•Means: The physical location and structure of the data center
must be safe.
•How to deploy:
• Build in hazard-free zones (e.g., away from floods or
earthquakes).
• Enforce building safety, install cameras, bomb
detection, and earthquake-proof structures.
• Avoid unnecessary windows, maintain buffer zones
around the facility.

2. Use redundant utilities at multiple sites

•Means: Don’t rely on a single power source or network; have
backups.
•How to deploy:
• Multiple power sources and network connections.
• Store copies of databases at separate sites.
• Ensure data consistency, watermark data, and enforce user
authentication.
3. Trust delegation and negotiation
6. Fine-grained file access control
•Means: Share trust across different organizations or data centers securely.
•Means: Control who can access each file or object.
•How to deploy:
•How to deploy:
• Use cross-certificates to manage trust across PKI (Public Key
• Implement file-level access permissions.
Infrastructure) domains.
• Enhances security beyond firewalls and IDS
• Ensure certificate authorities (CAs) negotiate properly to resolve
systems.
policy conflicts.

4. Worm containment and DDoS defense

7. Copyright protection and piracy prevention
•Means: Protect the data center from network-based attacks.
•Means: Prevent illegal copying or tampering with digital
•How to deploy:
content.
• Contain internet worms.
•How to deploy:
• Implement distributed defenses against DDoS attacks.
• Peer collusion prevention (prevent groups from
• Protect all cloud platforms and data centers proactively.
bypassing copyright).
• Filter poisoned content, detect unauthorized reads or
5. Reputation system for data centers alterations.
•Means: Track and rate data center reliability and security.
•How to deploy:
8. Privacy protection
• Use P2P technology to create a reputation system.
•Means: Protect sensitive user or organizational data.
• Build hierarchies from individual data centers to distributed
•How to deploy:
file systems.
• Use double authentication and biometric identification.
• Monitor intrusions and have disaster recovery plans.
• Enforce privacy through data watermarking and
classification.
 [Link] Defense with Virtualization
1. Decoupling from Physical Hardware
•In virtualization, a VM (Virtual Machine) is separated from 3. Distributed Intrusion Detection with VMs
the physical hardware. •Live migration: Moving a running VM from one physical server to
•It acts like a software version of a computer, which can be another.
treated as binary/digital data. •Researchers suggest using live-migrated VMs for Distributed
•Why it helps: You’re not tied to one physical machine, so you Intrusion Detection Systems (DIDSes).
can manage, move, or protect VMs more flexibly. • Multiple IDS VMs can be deployed at different data centers
or resource sites.
• These monitor and detect attacks collaboratively.
2. VM Operations for Security
•VMs can be:
4. Trust and Policy Management
• Saved: Keep snapshots of the current state.
•DIDS involves multiple VMs across different PKI (Public Key
• Cloned: Make exact copies for backup or testing.
Infrastructure) domains.
• Encrypted: Protect the VM data from unauthorized access.
•Challenges:
• Moved: Shift VMs between hosts or data centers.
• Trust negotiation is needed among domains.
• Restored: Recover quickly after an attack or failure.
• Security policy conflicts must be:
•Benefits:
• Resolved at design time (before deployment).
• Supports High Availability (HA) — services remain
• Updated periodically to adapt to new threats.
operational even if a host fails.
• Enables faster disaster recovery — VMs can be restored or
migrated quickly.
[Link] Privacy and Copyright Protection,(users’ personal data and files are safe in the cloud, and that copyrighted
content is not misused or stolen.)

• Cloud users can test predictable setups before full deployment.

• Lightweight platforms like Yahoo! Pipes make building apps easier, but sharing files in the cloud can risk privacy,
security, and copyright. Users want tools for working with large datasets.
• Major platforms secure resources:
• Google uses in-house software, Amazon EC2 uses HMEC and X.509 certificates. Browser-based cloud apps also
need protection.

Here are several security features desired in a secure cloud:

• Dynamic web services with full support from secure web technologies
- Cloud apps should use modern web tools that keep data and communication safe.
• Established trust between users and providers through SLAs and reputation systems
- Users should be able to rely on the cloud provider, often ensured by agreements (SLAs) or the provider’s good reputation.
• Effective user identity management and data-access management
- Only the right people can access the right data.
• Single sign-on and single sign-off to reduce security enforcement overhead
- Users can log in once to access many services, and log out once to stop access everywhere, making security easier.
• Auditing and copyright compliance through proactive enforcement
- The system should check and enforce rules to prevent misuse of data or copyrighted content.
• Shifting of control of data operations from the client environment to cloud providers
- The cloud provider handles data storage and operations, so the client doesn’t have to worry about technical details.
• Protection of sensitive and regulated information in a shared environment
- Important data (like personal or regulated data) must be kept safe even when shared with others in the cloud.
Example 4.7 Cloud Security Safeguarded by Gateway and Firewalls

•Figure 4.32 shows how a private cloud keeps data

safe.
•A gateway controls and protects access to the cloud,
especially from public networks.

•A firewall acts like an external shield, blocking

unwanted access.

•The gateway also protects the application server,

database, message queue, web services, and browsers
using security protocols like HTTP, JMS, SQL, XML, and
SSL.

•This defense system makes sure user data is safe and

not shared with anyone without permission.

Note: The system uses firewalls and security protocols to

keep private cloud data safe from attacks and leaks.
 4.6.2 Distributed Intrusion/Anomaly Detection:
-system that monitors multiple parts of a cloud or network to catch attacks or unusual activity before they cause harm

• Data security is the weakest point in cloud computing – Cloud systems are convenient, but protecting data is still the biggest challenge.
• Need for new security standards – We need common tools and APIs to prevent data lock-in (getting stuck with one cloud provider) and
to defend against network attacks.
• IaaS (like Amazon EC2) is most vulnerable – Infrastructure-as-a-Service clouds are more exposed to external attacks.
• Role-based tools help – Tools that assign roles and permissions make it easier to manage who can do what. Example: IBM Blue Cloud
uses a role-based web portal.
• SaaS examples – Software-as-a-Service platforms can provide services (like secretarial work) from a shared cloud, but many cloud
providers don’t guarantee security.
• Threats target VMs and software – Attacks can focus on virtual machines, operating systems, or applications in the cloud.
• Intrusion Detection Systems (IDS) – These try to stop attacks before they happen.
• Signature-based IDS: Looks for known attack patterns; needs frequent updates.
• Anomaly-based IDS: Detects unusual network activity that might indicate an attack.
• Distributed IDS: Needed to protect multiple VMs and catch both known and unknown attacks.
[Link] Distributed Defense against DDoS Flooding Attacks:
- protection system for cloud networks designed to stop DDoS attacks
• DDoS defense system – A system designed to protect cloud servers from large-scale attacks.
• Multiple network domains – The system must protect all parts of the cloud network, including the edges where
users connect.
• DDoS attacks – These attacks flood a server with too much traffic, often using worms, to overload it.
• Effects of flooding – The attack can crash the server by filling memory (buffer overflow), using up disk space, or
saturating connections.

Figure 4.33(a) shows how a flooding (DDoS) attack happens.

• A hidden attacker uses many “zombie” computers (infected machines under
their control) to send huge amounts of traffic.
• All the attack traffic targets one victim server at the bottom (connected to
router R0), trying to overload and crash it.
• Flooding traffic moves like a tree – The attack spreads from many sources
through the network toward the victim, forming a tree-like flow.

• Routers detect unusual traffic – Each router along the path can notice
abnormal spikes in traffic.

• Change-point detection – Routers look for sudden changes in traffic patterns to

spot a DDoS attack early.

• Prevents server overload – The system can detect the attack before the victim
server crashes.

• Cloud core protection – This method is designed to protect the main cloud
network, not just the edges.

• Provider-level cooperation – Cloud providers work together, so edge networks

don’t need to manually intervene.
Example 4.8 Man-in-the-Middle Attacks

Figure 4.34, VM migration – Moving a virtual machine (VM) from one host (VMM A) to another host (VMM B) over a
network.
Man-in-the-Middle (MITM) attack – An attacker intercepts the VM while it’s moving through the network.
What the attacker can do:
• View or steal sensitive data inside the VM.
• Modify the VM, including its operating system or applications.
• Insert a VM-based rootkit, which is a hidden malicious program that can take control of the VM without the user
knowing.
Risk: The attacker can compromise the VM and migration process completely while remaining undetected.
 4.6.3 Data and Software Protection Techniques

• Data coloring technique – A method used to track and protect data, making sure it stays accurate and private
while being used in the cloud.
• Watermarking scheme – A way to mark software files so that if they are copied or shared illegally, the owner
can trace and protect them.

[Link] Data Integrity and Privacy Protection

• Users want a cloud platform with lots of tools – This helps them build applications that work with big data.
• Examples of tools – Programs like MapReduce, BigTable, EC2, Hadoop, AWS, Google App Engine (GAE), WebSphere help manage and
process large datasets.
• Need for security software – Users also need security and privacy tools to keep their data safe while using the cloud.

Such software should offer the following features:

• Special APIs for authenticating users and sending e-mail using commercial accounts
• Fine-grained access control to protect data integrity and deter intruders or hackers
• Shared data sets protected from malicious alteration, deletion, or copyright violation
• Ability to secure the ISP or cloud service provider from invading users’ privacy
• Personal firewalls at user ends to keep shared data sets from Java, JavaScript, and ActiveX
• applets
• A privacy policy consistent with the cloud service provider’s policy, to protect against identity
• theft, spyware, and web bugs
• VPN channels between resource sites to secure transmission of critical data objects
 [Link] Data Coloring and Cloud Watermarking

• Sharing data in the cloud can be risky. Users want safe platforms with tools to work on protected data. Watermarking helps protect
digital files.

As shown in Figure 4.35,

• Data coloring – Giving each piece of data a unique “color” or label.
• Purpose – Makes it easy to tell data objects apart and track them.

• User and data color matching – Each user is given a color

that matches their data, helping track and manage trust.

• Watermarking in cloud storage – Watermarks can be added,

stored, and retrieved from colored data to protect it.

• Why data coloring is useful – Traditional

encryption/decryption is slow and heavy, but coloring uses
very few calculations, making it faster.

• Combined protection – Cryptography, watermarking, and

coloring can be used together to keep cloud data safe and
private.
[Link] Data Lock-in Problem and Proactive Solutions

Data lock-in problem – When you put your data and apps in a cloud, it’s hard to move them to another cloud.
Why it happens:
1. Proprietary APIs – Each cloud has its own tools, so you can’t easily extract data.
2. App incompatibility – You may need to rewrite apps from scratch when switching clouds.

Solution: Use standardized cloud APIs and OVF (Open Virtualization Format).
• OVF is a universal format for virtual machines that works across different clouds.
• It allows data and apps to move easily, improving cross-cloud use and workload migration.
4.6.4 Reputation-Guided Protection of Data Centers

• Trust – A personal feeling about someone; it’s subjective and can be biased. It can spread indirectly (transitive)
but isn’t always mutual (not symmetric).
• Trust can be passed along through other people.
• Example: If Alice trusts Bob and Bob trusts Charlie, then Alice might also trust Charlie even if she hasn’t met
him directly.
• Trust can flow through connections, not just directly between two people.

• Reputation – A public view about someone; it’s more objective because it’s based on many opinions.
• Example :Reputation is like a public rating based on many people’s experiences, not just one person’s
opinion.

• Changes over time – Reputation can change or fade, and recent behavior matters more than old history.
• Example: A restaurant had 5-star ratings last year, but recently several customers complained. Its
reputation drops, and new reviews matter more than old ones.

• Purpose in cloud – Reputation systems are used to protect cloud data centers and users by evaluating who is
reliable.
• Reputation systems help identify trustworthy users and services.
• Example: If a cloud user consistently follows rules and behaves properly, the system gives them a good
reputation, making them more trusted to access shared resources.
[Link] Reputation System Design Options

Figure 4.36 provides an overview of reputation system design options.

• Reputation – What the public thinks about

someone or something, like honesty, reliability, or
quality.
• Applies to – A person, software agent, product, or
service.
• Based on many opinions – It’s a collective
evaluation by users, agents, or owners.
• Where used – Reputation systems have been used in
peer-to-peer (P2P) networks, multiagent systems,
and e-commerce platforms.

• Reputation systems for cloud services – These are designed to track and evaluate the trustworthiness of users or services in the cloud.
• Two-tier classification – Most reputation systems were made for P2P networks or social networks, but they can be adapted for cloud
computing.
• Centralized vs. Distributed:
• Centralized – One main authority manages all reputations.
• Distributed – Multiple centers work together to manage reputations.
• Purpose – Using reputation and trust techniques from P2P and social networks can help protect cloud data centers and platforms
from attacks.
 • Centralized vs. Distributed reputation systems:
• Centralized: Easy to set up, but needs strong, reliable servers.
• Distributed: Harder to build, but handles failures better and can scale to more users.

• Second tier – Scope of reputation,(how widely the reputation is applied or who it evaluates)
• User-oriented: Focuses on individual users or agents (like most P2P systems).
• Resource/site-oriented: Focuses on the whole data center, products, or services.

• Examples of commercial centralized reputation systems:

• eBay, Google, Amazon – They track the reputation of services or sellers.

• Distributed reputation systems – Mostly developed by researchers and universities.

Examples:
• Aberer & Despotovic – Trust model for P2P systems.
• Eigentrust (Stanford) – Uses a trust matrix to calculate reputation.
• PeerTrust (Georgia Tech) – Supports e-commerce applications.
• PowerTrust (USC) – Uses Power law of Internet traffic for P2P reputation.
• Vu et al. – Proposed a QoS-based ranking system for P2P transactions.
[Link] Reputation Systems for Clouds
1. Redesigning reputation systems for cloud - “This means taking existing reputation systems (like those used in P2P or social networks) and
modifying them so they can work effectively in cloud environments.”
2. Data consistency across multiple databases – “Ensures all copies of data are accurate and up-to-date”.
3. Copyright protection -Protects content distributed widely from illegal use or copying.
4. User data separation from SaaS programs -Even if multiple users use the same application, no one can access another user’s data.
5. Unique naming of data objects - Prevents confusion and ensures global consistency in cloud storage
6. Unauthorized updates prohibited - Users cannot tamper with others’ data, protecting integrity.
7. Trust overlay network & hierarchy - Reputation can be implemented using a trust network and hierarchical P2P systems to monitor
both site-level resources and file-level data objects.
8. Coarse-grained and fine-grained access control -Ensures both general and detailed control over shared resources.
9. Centralized vs. distributed reputation systems.
• Centralized: Easier to implement but needs strong servers.
• Distributed: More scalable and handles failures better.
10. Reputation systems help security mechanisms- They assist all five previously mentioned security measures in protecting data centers.
11. Safe cloning of VMs - Social/reputation tools can help safely clone virtual machines.
12. Snapshot control and RPO - Cloud uses defined recovery points to manage VM snapshots safely.
13. New security mechanisms users demand Includes - Secured logging , Migration over secure virtual LANs ,ECC-based encryption for safe
VM movement
14 . Sandboxes Provide a safe and controlled environment to run third-party apps and test code without risking the main system.
[Link] Trust Overlay Networks

• Reputation is a collective evaluation by users and resource owners to measure trustworthiness.

• In cloud services, a trust overlay network with a distributed hash table (DHT) is used to quickly combine local reputations
into a global score.
• A two-layer design ensures fast updates and distribution of these reputations to all users, supporting trusted cloud operations.
• Figure 4.37 shows construction of the two layers of the trust overlay network.
• Bottom layer (trust overlay):
• Manages distributed trust and reputation across multiple resource sites.
• Handles user/server authentication, access control, trust delegation, and data integrity.
• Top layer (security overlay):
• Focuses on fast virus/worm detection, signature sharing, and piracy detection.
• Helps contain malware, prevent DDoS attacks, and stop copyright violations across multiple data centers.
• Reputation, privacy, copyright protection, and adaptive security measures together create a trusted and secure cloud
environment.
• Virtualization, reputation, watermarking, and cloud security services together protect data, while common standards
ensure different clouds can work smoothly together.