ICS 408 Exam A.O.

C
1. Characteristics of Distributed Systems
2. Compare and contrast Client-Server and Peer-to-Peer (P2P)
3. Microservices architecture
4. Event-Driven architectural style
5. Communication models in distributed systems
6. Google’s Distributed File System (GFS)
7. CAP theorem
8. Concepts, use cases, advantages, and challenges of distributed databases, distributed file systems,
and object stores
9. Load balancing strategies and their advantages and disadvantages
10. Differences and similarities between concurrency and parallelism
11. Security challenges faced by distributed systems and the common
12. security measures implemented to mitigate these challenges
13. Concepts and use cases of leader election and distributed consensus, distributed coordination and
provide examples of algorithms used for each
14. Concepts of scalability and load balancing in distributed systems
15. Scalability patterns, and their pros and cons
Answer
1. Characteristics of Distributed Systems
a. Heterogeneity: Distributed systems consist of a variety of hardware, software, and network
components. These systems can run on different platforms and use different programming languages,
which requires standard protocols for communication and interaction.
b. Scalability: Distributed systems are designed to scale horizontally by adding more nodes
(machines) to handle increased load. This helps in accommodating growing data and user demands.
c. Fault Tolerance: These systems continue to function even when parts of the system fail.
Redundancy and replication are often used to ensure reliability and availability.
d. Concurrency: Multiple processes can run simultaneously across different nodes, allowing efficient
resource utilization and improved performance.
e. Transparency: The complexity of the system is hidden from users and developers, making the
system appear as a single unified entity. This includes transparency in access, location, migration,
replication, and failure.
f. Resource Sharing: Resources such as files, printers, and data can be shared among multiple users
and applications across the distributed system.

2. Compare and Contrast Client-Server and Peer-to-Peer (P2P)

a. Client-Server:
i. Structure: Centralized architecture with dedicated servers serving multiple clients.
ii. Communication: Clients request services/resources, servers respond with the requested
services/resources.
iii. Scalability: Limited by server capacity; scaling often requires upgrading server hardware or adding
more servers.
iv. Management: Easier to manage and secure due to central control.
v. Examples: Web servers, email servers, and database servers.
b. Peer-to-Peer (P2P):
i. Structure: Decentralized architecture where each node (peer) acts as both a client and a server.
ii. Communication: Peers communicate directly with each other to share resources and services.
iii. Scalability: Highly scalable; adding more peers increases system capacity.
iv. Management: More complex to manage and secure due to lack of central control.
v. Examples: File-sharing networks like BitTorrent, blockchain networks.
3. Microservices Architecture
a. Definition: An architectural style that structures an application as a collection of small, autonomous
services, each responsible for a specific business capability.
b. Characteristics:
i. Independently Deployable: Each microservice can be developed, deployed, and scaled
independently.
ii. Loosely Coupled: Services interact through well-defined APIs, minimizing dependencies between
them.
iii. Organized Around Business Capabilities: Each service corresponds to a specific business function.
c. Advantages:
i. Scalability: Services can be scaled independently based on demand.
ii. Flexibility: Technology stack can vary for different services.
iii. Resilience: Failure in one service does not impact the entire system.
d. Challenges:
i. Complexity: Managing multiple services increases overall system complexity.
ii. Communication Overhead: Inter-service communication can introduce latency and require robust
protocols.

4. Event-Driven Architectural Style

a. Definition: An architectural style in which the flow of the program is determined by events, such as
user actions, sensor outputs, or message passing.
b. Components:
i. Event Producers: Generate events based on certain actions or changes in state.
ii. Event Consumers: React to events by performing specific actions.
iii. Event Channels: Transport events from producers to consumers.
c. Advantages:
i. High Responsiveness: Systems can react quickly to changes and events.
ii. Decoupled Components: Producers and consumers do not need to be aware of each other, leading
to more modular and flexible systems.
d. Challenges:
i. Complexity: Designing and managing event flows can be complex.
ii. State Management: Keeping track of state across events can be challenging.

5. Communication Models in Distributed Systems

a. Synchronous Communication: The sender waits for the receiver to acknowledge receipt of the
message before proceeding. It ensures reliable and ordered delivery but can lead to delays if the
receiver is slow.
b. Asynchronous Communication: The sender continues its operation without waiting for the
receiver's acknowledgment. It allows for more efficient use of resources but may require mechanisms
to handle message ordering and delivery guarantees.
c. Message Passing: Processes exchange messages to communicate. It is a flexible and general model
suitable for various distributed applications.
d. Remote Procedure Call (RPC): Allows a program to cause a procedure to execute on another
address space, typically on another physical machine. It abstracts the communication details, making
distributed computing more straightforward.

6. Google’s Distributed File System (GFS)

a. Definition: A scalable distributed file system designed to handle large data-intensive applications,
used internally at Google.
b. Characteristics:
i. Fault Tolerance: Uses replication to ensure data availability even in case of node failures.
ii. High Performance: Optimized for large files and high throughput operations.
iii. Simplified Consistency Model: Provides a relaxed consistency model that is suitable for many c. c.
c. Google applications.
Components:
i. Master Server: Manages metadata, namespace, and coordinates activities.
ii. Chunk Servers: Store the actual data chunks and handle read/write requests.
iii. Clients: Communicate with the master to locate data and directly with chunk servers to access
data.
7. CAP Theorem
Definition: States that a distributed system can provide only two out of the following three guarantees
at the same time:
i. Consistency: Every read receives the most recent write or an error.
ii. Availability: Every request receives a response, without guarantee that it contains the most recent
write.
iii. Partition Tolerance: The system continues to operate despite arbitrary partitioning due to network
failures.
iv. Implication: In the presence of a network partition, a system must choose between consistency and
availability. This trade-off guides the design of distributed systems.

8. Concepts, Use Cases, Advantages, and Challenges of Distributed Databases, Distributed File
Systems, and Object Stores
a. Distributed Databases:
i. Concepts: Data is distributed across multiple physical locations, and each location can function
independently.
ii. Use Cases: Financial services, e-commerce platforms, global enterprises.
iii. Advantages: Improved reliability, availability, and scalability.
iv. Challenges: Ensuring data consistency, managing complex transactions, and handling network
partitions.
b. Distributed File Systems:
i. Concepts: Files are distributed across multiple servers to improve access speed and reliability.
ii. Use Cases: Cloud storage services, big data applications.
iii. Advantages: Scalability, fault tolerance, and high availability.
iv. Challenges: Data synchronization, replication management, and consistency issues.
c. Object Stores:
i. Concepts: Data is stored as objects, which include the data itself, metadata, and a unique identifier.
ii. Use Cases: Cloud storage (e.g., Amazon S3), multimedia storage, backup and archival.
iii. Advantages: Scalability, ease of management for unstructured data, and cost efficiency.
iv. Challenges: Handling consistency and latency, managing large volumes of data efficiently.

9. Load Balancing Strategies and Their Advantages and Disadvantages

a. Round Robin:
i. Advantages: Simple to implement, evenly distributes requests.
ii. Disadvantages: Does not account for server load differences, potentially leading to uneven load
distribution.
b. Least Connections:
i. Advantages: Balances load based on current connection counts, more effective for long-lived
connections.
ii. Disadvantages: Monitoring overhead, may not account for varying connection complexities.
c. IP Hashing:
i. Advantages: Consistent routing for the same client IP, useful for session persistence.
ii. Disadvantages: May result in uneven distribution if client IPs are not uniformly distributed.
d. Weighted Round Robin:
i. Advantages: Distributes load based on server capacity, more efficient resource utilization.
ii. Disadvantages: More complex to configure, requires accurate weight assignment.

10. Differences and Similarities Between Concurrency and Parallelism

a. Concurrency:
i. Definition: Multiple tasks making progress, overlapping in execution, often on a single processor.
ii. Focus: Managing multiple tasks to ensure progress.
iii. Example: Multithreading where threads share the same resources.
b. Parallelism:
i. Definition: Multiple tasks running simultaneously, typically on multiple processors or cores.
ii. Focus: Performing tasks simultaneously to speed up execution.
iii. Example: Parallel processing in multi-core systems.
c. Similarity: Both deal with executing multiple tasks.
d. Difference: Concurrency is about dealing with multiple tasks at once (managing task execution),
while parallelism is about doing multiple tasks at the same time (actual simultaneous execution).

11 & 12. Security Challenges Faced by Distributed Systems and Common Security Measures
Implemented to Mitigate These Challenges
a. Challenges:
i. Data Breaches: Unauthorized access to sensitive data.
ii. DDoS Attacks: Overloading systems with a flood of requests, causing service disruptions.
iii. Man-in-the-Middle Attacks: Interception and alteration of communication between nodes.
iv. Endpoint Security: Securing each node in the distributed system.
v. Data Integrity: Ensuring data is not altered or tampered with during transmission or storage.
b. Security Measures:
i. Encryption: Protects data in transit and at rest using cryptographic techniques.
ii. Authentication and Authorization: Verifies user identity and grants appropriate access levels.
iii. Firewalls and Intrusion Detection Systems (IDS): Monitor and control incoming and outgoing
network traffic based on predetermined security rules.
iv. Secure Communication Protocols: Use of SSL/TLS to secure communications between nodes.
v. Regular Security Audits: Conducting periodic security assessments to identify and mitigate
vulnerabilities.

13. Concepts and Use Cases of Leader Election and Distributed Consensus, Distributed Coordination,
and Examples of Algorithms Used for Each
a. Leader Election:
i. Concept: A process to designate a single node as the coordinator or leader among distributed nodes.
ii. Use Cases: Distributed databases, distributed caching, and cluster management.
iii. Algorithms: Bully algorithm, Paxos.
b. Distributed Consensus:
i. Concept: Agreement among distributed processes or nodes on a single data value or decision.
ii. Use Cases: Blockchain, replicated state machines, and configuration management.
iii. Algorithms: Raft, Paxos.
c. Distributed Coordination:
i. Concept: Synchronizing and managing the activities of distributed components to achieve a
common goal.
ii. Use Cases: Resource allocation, distributed job scheduling, and configuration management.
iii. Examples: Zookeeper (coordination service for distributed systems), Chubby (distributed lock
service).

14. Concepts of Scalability and Load Balancing in Distributed Systems

a. Scalability:
i. Concept: The ability of a system to handle increased load by adding resources.
ii. Types:
I. Vertical Scaling (Scaling Up): Adding more power (CPU, RAM) to an existing node.
II. Horizontal Scaling (Scaling Out): Adding more nodes to the system.
b. Load Balancing:
i. Concept: Distributing workload across multiple resources to ensure no single resource is
overwhelmed.
ii. Goal: Improve system performance, reliability, and availability.
iii. Techniques: DNS load balancing, hardware load balancers, software load balancers.

15. Scalability Patterns, and Their Pros and Cons

a. Replication:
i. Concept: Creating multiple copies of data across different nodes.
ii. Pros: Increased availability and fault tolerance.
iii. Cons: Challenges in maintaining consistency across replicas.
b. Sharding:
i. Concept: Dividing data into smaller, more manageable pieces (shards) and distributing them across
multiple nodes.
ii. Pros: Improved performance and scalability.
iii. Cons: Complexity in data management and rebalancing shards.
c. Caching:
i. Concept: Storing frequently accessed data in fast-access storage (e.g., memory) to reduce load on
primary storage.
ii. Pros: Reduced latency, improved performance.
iii. Cons: Potential for stale data, cache invalidation challenges.