5S principles in the design of IT systems

5S principle: Simple, Solid, strong, Safe, Smart.

Simple

Simple is simply the principle, here the simple has two meanings, one is the logic is simple and easy to understand, the other is the simple coupling of the structure is weak.

According to the method of modularization design, we usually define the boundary of the system first, then divide the module within the system, then divide the function points within the module.

Usually in the top-down design, the simple principle is very easy to follow, but sometimes the function is gradually enhanced, from a function point to a system, in the late development of the early design will often be reconstructed, which is quite painful.

Give an example of a manufacturing system.

For example, the workshop has a station, the workers are required to do assembly work, after the completion of the quality inspection. Strictly speaking, the worker integrates two roles: the fitter and the quality inspector, but because of the difficulty of production technology and low demand for the production cycle, these two actions are done by the same person in order to save labor costs. The corresponding IT system functions will also unify the logic to facilitate user operation.

But later production cycle requirements more and more high, in order to improve efficiency, the workshop recruit new workers, a station into two stations, a person's work into two people's life, one of them specialized to do assembly, the other one specializes in quality inspection. This requires the IT system to split the functionality and turn a logic into two logic.

Usually in order to reduce the development time, we can add a number of if then else judgments in the original complete logic, and then let different roles execute different logical blocks. That is, load the logic first and then judge the execution.

Another design idea is that at the beginning of the first development, according to the simple principle of analysis, the logic of the Assembly and the logic of quality control are closed into logical blocks, when implemented according to the configuration of the situation to be combined. Corresponding to change the logic of the previous user to perform two roles, and two users to execute the corresponding roles. The code behind the scenes is only responsible for their own logic, and the foreground display combines the logic according to the configuration.

Solid

Solid is the logic to be solid and complete.

The integrity of the business logic can reduce system anomalies and thus improve the usability of the system.

An effective way to implement the solid principle is to have a black-box test that allows the user to write a test scenario for the user to receive tests.

Developers often think of normal business flows, while users are familiar with more unusual business flows, and testing of abnormal business flows can improve the integrity of the logic.

Strong

Strong is the system to be reliable, can support multiple concurrency, multi-user, large data volume, in various ways to reduce the time of the machine.

Hardware, you can improve the system availability through cluster, Load balance and so on.

Software, you can reduce the computational time through caching, read/write separation, user grouping, redundancy, archiving, database partitioning and other technologies.

A simple principle to maintain strong is to find and solve bottlenecks. For example, a system bottleneck is to run a particular report, if several people run this report at the same time period, it causes database queries for a long time. The solution is to improve the database query efficiency from the hardware point of view, or prioritize query statements and table structure, there is also a way to allow the user to schedule the query request, such as early in the idle time period in advance in the background to execute the query.

Safe

Safe is a security principle that avoids disruption of system services through various means.

In IT systems, Black swan events (unpredictable abnormal events) are difficult to avoid, the solution is to constantly accumulate exception handling measures, whenever there is an exception, through error, log, notification and other ways to handle the exception, so as to avoid disruption of system services.

In the design of the system, a configurable exception handling module should be planned, which avoids the timely processing of exceptions in the case of overriding the core code.

Smart

Smart is the smart principle.

Now the network of Open source code, modules, methodologies, many Internet applications can also be borrowed to enterprise systems.

Map-reduce, for example, is a methodology advocated by Google, and in fact many of the code that includes a lot of if and else can use map-reduce to simplify logic.

5S principles in the design of IT systems