Our lives are supported by information systems, including convenience-store cash registers, railway seat reservations, bank ATMs, and smartphone payments. An information system is the entire structure that combines computers and networks to collect, process, and transmit information and provide services. In this lesson, you will learn how information systems are structured and how reliability is designed to create systems that do not stop.
Examples of Familiar Information Systems
A convenience store’s POS system records “what, when, and how many items were sold” the moment a product barcode is scanned at the register and sends the information to the headquarters server. The collected data is used to replenish inventory and develop new products. A seat-reservation system centrally manages data in a database so that the same seat is not sold twice, even when accessed from any counter or app throughout the country. Remember that a database is at the center of an information system.
Three-Tier Structure of Web Systems
Many Web services are built with a three-tier structure (three-tier architecture) that combines three types of servers with different roles.
・Web server (presentation layer): The gateway that receives requests from browsers and returns screens (HTML).
・Application server (application layer): Executes service-specific processing (business logic), such as login processing and fee calculation.
・Database server (data layer): Stores member information and product data and reads and writes data in response to queries.
Separating the layers makes it easier to expand and modify systems flexibly, such as increasing only the number of Web servers when access increases or replacing only the database with higher-performance equipment.
Cloud Computing
Instead of purchasing and operating servers themselves, organizations can borrow computing resources from a provider over the Internet. This is called cloud computing. Depending on the scope of what is borrowed, services are classified as SaaS, in which an entire application is borrowed (for example, Web mail or online Office), PaaS, in which a development and execution environment is borrowed, and IaaS, in which only infrastructure such as servers and storage is borrowed. Advantages include lower initial costs and the ability to increase or decrease scale as needed.
Reliability Design — Ways to Keep Systems Running and Ways to Stop Safely
When banking and transportation systems stop, they can have a major impact on society. Therefore, the following designs are used.
・Redundancy: Prepare multiple devices or communication lines so that the entire system does not stop even if one fails.
・Fault-tolerant: A design in which the system as a whole continues to function even if part of it fails.
・Fail-safe: A design that moves to the safe side when a failure occurs. Examples: a traffic signal that displays red when it fails, and an oil heater whose flame goes out when it tips over.
・Foolproof: A design that remains safe even when humans operate it incorrectly. Example: a washing machine that does not run unless the lid is closed.
・Series (the system works only when both operate): 0.9 × 0.9 ＝ 0.81. This is lower than for one device.
・Parallel (the system works if either one operates＝redundancy): The probability that both fail at the same time is 0.1 × 0.1 ＝ 0.01, so availability is 1 − 0.01 ＝ 0.99. This is much higher than for one device.
The numbers confirm that simply placing two identical devices in parallel creates a system that is more resistant to failure.
- A database is at the center of an information system, and centralized management prevents data inconsistencies.
- Web systems are basically structured in three tiers: “Web server／application server／DB server.”
- Cloud services are divided into SaaS／PaaS／IaaS according to the scope of what is borrowed.
- Availability increases with redundancy (parallel operation), but decreases when components are added in series.
- Fail-safe＝moves to the safe side during failure; foolproof＝prevents incorrect operation.
Practice Problems
- Briefly explain the role of each layer in the three-tier structure of a Web system.
- When two devices with availability 0.8 are used, find the overall system availability when they are connected in series and when they are connected in parallel.
- Is a “railway crossing system that automatically turns all signals red when a failure occurs” based on fail-safe or foolproof thinking? Answer with a reason.
Answers・Explanations
- Answer: The Web server receives requests from browsers and returns screens. The application server performs service-specific processing such as login and calculations. The database server stores data and responds to reads and writes.
Explanation: Separating roles makes it easier to strengthen or modify parts of a system according to the load. - Answer: Series: 0.8×0.8＝0.64. Parallel: 1−(0.2×0.2)＝1−0.04＝0.96
Explanation: In series, both devices must operate, so the probabilities are multiplied. In parallel, subtract the probability that both fail at the same time from 1. Redundancy improves availability from 0.8 to 0.96. - Answer: Fail-safe. When a failure occurs, the system is designed to move to the “safe side” by stopping trains.
Explanation: Foolproof design prevents human operating errors (for example, pressing the wrong button). Fail-safe design ensures safety when equipment fails.