Shulou(Shulou.com)06/02 Report--

Performance factors to be considered in the selection of development framework

Before the new product enters the research and development stage, the selection of technology, operating system, hardware and database is an important task that must be completed, which is a comprehensive evaluation of the non-functional requirements of the product and various elements and constraints in the architecture design. it is a comprehensive choice to verify whether the future technical framework can meet the continuous operation and maintenance expansion in the process of business expansion.

As can be seen from the above figure, technology selection is actually a process of decomposing product technology from different dimensions, and through analysis, we can reasonably decompose the technical requirements, then comprehensively evaluate the technical / product requirements and finally choose the appropriate framework, such as the performance indicators in the non-functional indicators, which are the key analytical indicators in the Internet era.

In the following years, although the company will cooperate with the company to travel to various real estate to do pre-sales POC non-functional technical support or travel to various city firms to assist local project managers in dealing with non-functional problems, and occasionally invited to some local Internet enterprises to assist them in testing and tuning work such as production performance fault handling or development framework selection, in fact, when doing these non-functional consultation or fault handling. Most of the problems encountered are due to the high probability of failure caused by immaturity such as the beginning of the framework design. Therefore, in order to guard against the future, many enterprises are willing to invest in these technical verification before the selection of the technology development framework before the new products are put on the shelves, the main purpose is to ensure the return on investment and optimize the input cost of IT. For example, non-functional test verification such as framework common class performance maintenance, capacity planning performance verification, hardware platform and software platform procurement and selection to predict performance and capacity planning, as well as whether the architecture can support high concurrency, architecture expansion, agile development and other software design capabilities and market trends when the company's business development increases in the future. for example, now many enterprises prefer to consider micro-service architecture.

As professional non-functional technicians, when we help customers select models, we cannot implement technology because of technology. Products are ultimately for actual customers to use, but products are also products of technology, so we need to consider the following four-image thinking. Consider non-functional factors in different roles:

In fact, it is the different dimensions of technical and non-technical personnel to consider, how to verify testing.

The user is concerned with the corresponding time of the user's action.

A) simplicity and agility of business operations

B) rationality and correctness of data retrieval

C) efficiency of data exchange, etc.

The second is the technical point of view, such as

From a technical administrator's point of view, consider the performance points that need to be concerned.

A) looking at database performance from a DBA perspective, such as table locks?

B) does the network administrator see if there are any transmission performance problems such as network congestion?

C) the system operation and maintenance staff to see if there is a bottleneck in resource utilization, such as disk space.

D) middleware managers, check for connections, insufficient threads, or memory recovery exceptions, etc.

E) Architecture managers, framework development and design phase to consider scalability, security, fault tolerance, portability, detachability, transport mode asynchronism, etc.

Again, consider it from the perspective of the developer (designer).

A) Application thread lock problem?

B) Index reasonableness?

C) timely release of the object?

D) the rationality of the number of data displayed? Etc.

So from the perspective of a performance test engineer, what should we focus on?

A) hierarchical problem decomposition of response time

B) the formula for calculating the number of users of the system

C) decomposition and root cause analysis of the utilization of service resources

D) estimation and calculation of TPS values and positioning analysis of corresponding problems

E) how does the throughput verify the size?

For example, the formula for calculating throughput

From a business point of view, throughput can be measured in units such as requests per second, pages per second, number of people per day, or number of business processed per hour.

From a network point of view, throughput can be measured in bytes per second

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.