Clustered Systems vs. Multiprocessor Systems

Clustered Systems vs Multiprocessor Systems

• Clustered Systems: multiple machines work together to achieve a common goal, each machine has its own memory and processing power.
• Multiprocessor Systems: multiple processors share a common memory and processing power.

Cluster Cooperation

• Key requirement for two machines in a cluster to cooperate effectively: they must be able to communicate with each other.

Shared Disk Architecture

• Major disadvantage of allowing all nodes access to the same disk storage: it can lead to data inconsistencies and conflicts.

CPU Interrupts

• Purpose of interrupts in a CPU: to temporarily pause the execution of the current task to handle a higher-priority task or event.

Shared Nothing Architecture

• Requires more complex data distribution and querying logic: because each node has its own memory and processing power, data distribution and querying logic must be more complex to ensure data consistency.

Linux Kernel

• Function of the 'jiffies' variable: a measure of time, used to track the time elapsed since the system was booted.

Cluster Software

• Critical aspect of Cluster Software in Clustered Systems: managing the communication and cooperation between nodes.

CPU DMA

• When a CPU sets up a DMA transfer: the CPU transfers control of the system bus to the DMA controller, allowing it to transfer data directly to or from memory.

Interrupts vs Traps

• Difference between interrupts and traps: interrupts are external events, such as I/O completions, while traps are internal events, such as divide by zero errors.

Secure Operating System

• Challenge of constructing a secure operating system on hardware without privileged mode: it is difficult to ensure the operating system has sufficient control over the system to enforce security policies.

Clustered Systems vs Multiprocessor Systems

• Main difference in terms of data architecture: Clustered Systems have multiple machines with their own memory, while Multiprocessor Systems have multiple processors sharing a common memory.

Shared Disk Architecture

• Critical disadvantage: it can lead to data inconsistencies and conflicts, and can be a single point of failure.

Clustered Systems

• Suitable for quick failover but with increased complexity for managing access and data integrity: Shared Disk Architecture.
• Key challenge associated with Shared Nothing Architecture: managing data distribution and querying logic.

State Management

• Essential mechanism for state management and decision-making among nodes in a clustered system: communication and coordination mechanisms.

Scalability

• Difference between Clustered Systems and Multiprocessor Systems: Clustered Systems are more scalable, as they can add more machines to the cluster as needed.