Distributed Systems: Client-Server Example

Organizing Distributed Systems

• The primary focus when organizing a distributed system is to ensure efficient communication, data consistency, and fault tolerance.

Deployment

• An instantiation of a software architecture on real machines is deploying the software to the chosen hardware.

Client-Server Architecture

• A client-server system architecture involves a single centralized server and multiple clients.
• A benefit of this architecture is its ease of management.

Components in Distributed Systems

• A component in a distributed system is a self-contained unit that performs a specific function, such as a database server or a web server.
• The common file used by both the client and server in a client-server system is a shared configuration file.

Single Server Architecture Problems

• A single point of failure is one of the main problems with using a single server.
• In a single client and server architecture, the server is responsible for managing all resources and data for the client.

Multiple Servers

• Using multiple servers improves performance and scalability by distributing workloads.
• The server's primary role in a multiple-client/multiple-server architecture is to handle requests from multiple clients concurrently.
• A single client and server architecture limits the system to only one client at a time.

Remote Procedure Calls

• When a client invokes a server, it initiates a remote procedure call (RPC).
• Email and FTP are examples of client-server applications utilizing RPC.
• The caller stub in an RPC acts as an intermediary, translating the call from the client to a format that can be understood by the server.
• From the caller's point of view, the semantics of an RPC are identical to calling a local procedure.
• When a caller makes a remote procedure call, the stub on the caller side sends the call to the stub on the server side, where it is processed.
• The callee stub on the receiving end translates the request from the client, executes the procedure on the server, and sends the result back to the client.

Multitier Systems

• A three-tier system is a common example of a multitier system.
• Middleware in a distributed system provides a layer of abstraction, facilitating communication between different components.
• The primary difference between a multitier system and a peer-to-peer system lies in the centralized functionality of the multitier system.

Platform Layer

• The platform layer in a distributed system handles low-level communication and resource management.
• RPC, RMI, and CORBA are used in distributed systems to enable communication and data transfer between different components.
• The key characteristic that distinguishes a multitier system from other architectures is its layered structure.