401 Lecture 5

• Architecture styles are broad design principles that guide the overall structure of a software system.
• Choosing the right architecture style is critical.
• It influences the design, implementation, costs, and schedule of software development.
• Architecture characteristics include scalability, elasticity, maintainability, consistency, teamwork, and portability.
• Additional architecture characteristics are modularity, flexibility, reusability, interoperability, simplicity, and deployability.
• Further characteristics include performance, security, reliability, testability, fault tolerance, and update capability.

Monolithic Architecture

• This is a traditional software development model.
• A single code base performs multiple business functions.
• Key characteristics include a single application block and unified deployment.
• There is no component isolation, interconnection and interdependence as well as tight coupling.
• The architecture is a self-contained unit that is simple to develop and deploy.
• It can be difficult to modify without impacting the whole system.

Layered Architecture

• Each layer provides specific functionalities and services to the layers above it.
• Key characteristics include simplicity, modularity, maintainability, and separation of concerns.
• More characteristics include scalability, reusability, testability, interoperability, and flexibility.
• The presentation layer is the top layer, followed by the business layer, persistence layer, and database layer.
• The database is another layer.
• A main class instantiates everything that is required.
• The architecture is good for demand maintainability and budget contrains.
• Ideal for small and simple applications.
• Such architecture is unsuitable for Real time or max performance systems.
• Edsger Dijkstra, popularized Layered Architecture in the 1960s while working on the theoretical basis for operating systems.
• His work on structured programming and the abstraction layer concept made software easier to understand, debug, and maintain.
• These ideas are foundational to how systems are designed today.

Pipeline Architecture

• This structures the processing of data or tasks as a series of connected processing stages.
• Key characteristics include sequential processing and modularity.
• Also: Parallelism, isolation, scalability, reusability, flexibility, simplicity, and testability
• Pipes form the communication channel between filters.
• Each pipe is unidirectional and point-to-point between filters.
• The payload in the pipes may contain any data structure, but architects usually like the idea of small amounts data.
• Filters are self-contained, independent and stateless.
• Filters perform one task only.
• Composite tasks are handled by a sequence of filters.
• A producer is the starting point of a process.
• It is outbound only, and sometimes called the source.
• A transformer accepts input, and optionally performs transformation, then forwards it to the outbound pipe.
• A tester accepts input, and tests one or more criteria, and optionally produces the output.
• The consumer is the termination point for the pipeline flow.
• Consumers may persist the final result of the pipeline process to a database, or display final results on a user interface screen.