System Architecture - CRASH COURSE

System architecture

A document that defines the structure of an information system provided to support the business process required to achieve the objective of an enterprise. It is divided into TA (Technical Architecture), DA (Data Architecture), and AA (Application Architecture).

System architecture

The structure of an information system, which includes principles and guidelines that define the components, such as hardware, software, security, interactions, and constraints; typically refers to Technical Architecture (TA).

Information system

A system that uses information technology (IT).

Server

Provides the computing power of an information system.

Network

Connects the information system components for communication.

Storage

Holds the data of an information system.

Block storage

Saves data in fixed block units.

File storage

Saves data in file units.

Object storage

Saves data in object units.

Direct access storage (DAS)

Storage mounted in each server.

Network attached storage (NAS)

Storage connected through a network.

Storage area network (SAN)

Storage connected through the network used exclusively for storage.

Centralized architecture

Stores and operates the system and data from an integrated center.

Multi-region distributed system architecture

Operates the systems, and the application systems are distributed regionally.

Client-server architecture

Places system functions in the servers and clients and configures them to use the service, depending on the business size or environment.

Web system architecture

Runs the server's application programs, and the client uses the service with a web browser.

Presentation layer

The highest level of an application program; provides service information and the interface to users.

Business logic layer

Runs the program information processing rules that fulfill business requirements (business logic).

Data layer

Accesses data resources, such as the database, and reads or writes data.

Two-tier architecture

Stores and processes data in the server and processes business logic and presentation for the client.

Three-tier architecture

A structure with an additional tier between the presentation tier and the data tier to process business logic, making it flexible and scalable.

Mainframe

A large steel enclosure, like a cabinet, that houses the CPU and the main memory.

Downsizing (server context)

Replacing a mainframe server with a super Unix server.

Protocol

A standardized communication rule for sending and receiving data through a network.

IETF (Internet Engineering Task Force)

Responsible for most Internet-related standard protocols.

3GPP (Third Generation Partnership Project)

Responsible for wireless communication protocols such as GSM, CDMA, UMTS, LTE, and LTE-A.

ITU-T (International Telecommunication Union Telecommunication Standardization Sector)

Responsible for protocols related to telephones.

OSI (Open Systems Interconnection) reference model

A conceptual model that characterizes and standardizes the communication functions of a telecommunication or computing system.

Application layer (OSI)

Provides services, such as a user interface, email, and database management.

Presentation layer (OSI)

Supports the syntax of data exchanged between two systems and the encoding translation and encryption related to the semantic.

Session layer (OSI)

For communication sessions; establishes, maintains, and synchronizes the interaction between communication devices.

Transport layer (OSI & TCP/IP)

Provides functions like reliable message transfer between end-to-end processes, as well as error control. Also called the host-to-host transport layer in TCP/IP.

Network layer (OSI & TCP/IP)

Supports packet transmission between networks from the sending point to the destination. In TCP/IP, also called the Internet layer, responsible for addressing and routing functions.

Data link layer (OSI)

Provides the function of transferring frames between hops without error.

Physical layer (OSI)

Transfers the actual bits through physical media.

Application layer (TCP/IP)

Allows access to services on other layers using application programs like web (HTTP), DNS, telnet, FTP, and email. Includes concepts of OSI's presentation and session layers.

Network interface layer (TCP/IP)

Has the role of actually sending and receiving TCP/IP packets through physical media. Includes OSI's data link layer and physical layer.

MAC (Medium Access Control) address

The address system used by the data link layer to transfer frames between physically connected nodes. Also called physical address or Ethernet Hardware Address (EHA).

IP address

The address system of the network layer to transfer datagrams between two hosts/routers. A logical address.

Port number

Used to transfer messages between two processes (running applications).

Well-known ports

Port numbers from 0 to 1023, allocated by IANA.

Registration ports

Port numbers from 1024 to 49,151, mostly used by servers.

Dynamic ports (Temporary ports)

Port numbers from 49,152 to 65,535, mostly used as a temporary number by clients.

Operating System (OS)

A system software that provides a computer resource interface to users or application programs by efficiently managing the limited resources of computer hardware.

Process

A running program; a system work unit that allocates, uses, and recovers various resources to perform its tasks.

Process Control Block (PCB)

Stores information necessary for process management.

Parent process

A process that creates other processes.

Child process

A newly created process by a parent process.

Thread

A basic unit for using a CPU; also called a lightweight process.

Multi-threaded process

A process having multiple threads.

Race condition

Two or more parallel processes simultaneously accessing and changing the same data, or the order of manipulating the data affecting the execution result.

Critical section

A part of the code in a process where shared data is changed.

Mutual exclusion

If a process runs in its critical section, other processes cannot run their own critical section.

Progress (critical section)

Only processes that are not running in the remainder section can enter a critical section without a running process.

Bounded waiting (critical section)

After a process makes a request to enter its critical section, the time allowed for other processes to enter the critical section should be limited.

Semaphore

An integer variable used as a synchronization tool, with P (wait) and V (signal) operations.

Deadlock

A state where a process in "standby" is never changed because the requested resource is allocated to other standby processes.

Hold & wait (deadlock condition)

Processes are requesting other resources while occupying the current resource.

Non-preemption (deadlock condition)

The resource allocated to each process cannot be forcibly released until its use is complete.

Circular wait (deadlock condition)

Requests for a resource by different processes are continuously repeated.

Virtual memory

The technique of running processes, even if they are not present in the memory unit.

First-fit (memory allocation)

Allocates the first searched space that is sufficiently large.

Best-fit (memory allocation)

Allocates the smallest available space that can meet the required memory unit capacity.

Worst-fit (memory allocation)

Allocates the largest available space.

External fragmentation

Wasted memory space between allocated blocks.

Internal fragmentation

Wasted memory space within an allocated block because the block is larger than the requested space.

Compaction technique

Merges small-sized available memory to remove external fragmentation.

Coalescing technique

Merges spaces with adjacent addresses in the unused empty space list to create a larger space.

Scheduling

The OS process of allocating processes to the CPU efficiently.

Preemptive scheduling

A process can take CPU resources while another process is occupying them.

Non-preemptive scheduling

If a CPU resource is allocated to a process, it cannot be allocated to another process until the task has been completed.

FIFO (First In First Out) scheduling

Allocates CPU resources according to the order in which they are requested.

SJF (Shortest Job First) scheduling

Allocates CPU resources to the process with the shortest expected job time.

Round Robin (R-R) scheduling

Allocates CPU resources according to request order, but each process can use the CPU only for a specified period.

Paging technique

Partitions main memory into fixed-size frames and partitions tasks into fixed-size pages to load into frames.

Segmentation technique

Partitions tasks into variable-size segments (logical units) and loads them into main memory.

Optimal technique (page replacement)

Replaces a page that is not likely to be used for the longest time.

FIFO (page replacement)

Replaces the first loaded page.

LRU (Least Recently Used) technique (page replacement)

Replaces the page that has been unused for the longest time.

LFU (Least Frequently Used) technique (page replacement)

Replaces the least used or least intensively used page.

NUR (Not Used Recently) technique (page replacement)

Replaces the page not used recently, assuming it's less likely to be used soon.

Locality

A tendency to intensively refer to only some pages while a process is running.

Working Set

A set of frequently referred to page lists for a certain period to execute a process efficiently.

Thrashing

A phenomenon in which the CPU utilization rate decreases because page replacement takes longer than processing time.

File system

Stores related data to find and access OS data and programs quickly; a directory structure providing information on an aggregation of files.

File

A set of data with a name, stored in an auxiliary memory unit.

Directory

A logical structure to manage files in a file system.

Contiguous allocation (file system)

Arranges files as a set of contiguous addresses on the physical disk.

Linked allocation (file system)

Stores files in block units and links each block to a list.

Indexed allocation (file system)

Manages all pointer data of a file in an index block.

i-node (UNIX)

Has the role of allocation, application, creation, link, and deleting through the file/directory data.

Computer architecture

A conceptual design and the operational structure that is the basis for computer systems.

CPU (Central Processing Unit)

Interprets instructions and handles arithmetic or logical operations and data processing.

Main memory

Memory semiconductor chips closely located to the CPU, accessible at high speed for temporary storage.

Auxiliary storage device

Secondary storage, accessible at low speed, with high storage density for permanent storage.

I/O device

Consists of input and output devices for interaction between users and computers.

Von Neumann architecture

CPU can read commands and read/write data from the same memory using the same signal bus.

Harvard architecture

Separates instruction memory and data memory, allowing parallel access.

Control unit (CPU)

Hardware module that sequentially generates control signals to interpret and run program codes.

ALU (Arithmetic Logic Unit)

CPU element for arithmetic and logical operations.