LoadRunner Performance Test indicator TPS (Transaction per Second) summary

The content is reproduced, the specific circumstances should also be specific reference, practice can out the truth
TPS (Transaction per Second) definition

TPS is the abbreviation for transactions per Second, which is the number of transactions per second. It is the unit of measurement for software test results. A transaction is the process by which a client sends a request to the server and the server responds. The client starts the timing when the request is sent and ends the timer after receiving the server response, which calculates the time spent and the number of transactions completed, and ultimately uses that information to estimate the score. 

TPS (Transaction per Second) role

Reflects the maximum capacity of the system to process the business at the same time, and the higher the data, the better the system processing power. Description (see that the system's TPS is getting bigger over time, and how many transactions per second the system can handle in less than a minute). The highest value here does not necessarily represent the maximum processing power of the system, the TPS is affected by the load and increases gradually as the load increases, and the TPS decreases when the system enters the peak period. And in a few minutes, a small number of failed transactions begin to occur.TPS (Transaction per Second) LimitationsTPS (Transaction per Second) is a client-side view of server processing power, not the extent to which TPS can support much concurrency (such as a business 100 transactions, another 10 transactions). Tps= the total number of transactions/scripts run during the script run, if you use the aggregate point policy, the server does not process transactions during the wait time before the script executes, then the TPS and the ideal result are inconsistent at this time. Reasons for restricting TPS: server performance, code structure, client-imposed stress, NIC, and so on.

the relationship between TPS (Transaction per Second) and response time

TPS (Transaction per Second) and response times are rated in the ideal state. If there are 20 entrances, and the number of concurrent numbers is only 10, the TPS is 10, and the response time is always 1, indicating that concurrency is not enough and that the concurrency is increased to the peak of TPs. If it increases to 100 concurrency, it causes the thread to wait, causing the average response time to change from 1 seconds to 3 seconds, and the TPS from 20 down to 9;TPS and response time are calculated separately, and the two are not computed from each other. Response time and TPS are inversely proportional to each other, but there is no direct relationship between the two.

The role of TPS (Transaction per Second) in performance testing

The throughput (load capacity) of a system is closely associated with the request for CPU consumption, external interfaces, Io, and so on. The higher the CPU consumption of a single request, the slower the external system interface and IO impact, and the lower the system throughput capacity, the higher the reverse. Several important parameters of system throughput: TPS, concurrency, Response time (tps= concurrency/Average response time) using TPS to calculate the maximum daily throughput of the system: to find the highest TPS and PV, the two elements have a relatively stable relationship. The maximum TPS is obtained by the stress test or experience evaluation, and then the 1 relationship is followed to calculate the maximum daily throughput of the system. For example: to business Chinese and Taobao to the customer base is not the same, the two customer base network behavior is not applied, their TPS and PV relationship ratio is not the same. Taobao Taobao the relationship between TPs and PV is usually, the maximum TPS:PV is about 1:11*3600 (equivalent to the maximum TPS access for 11 hours, this is the scene of the commodity details, different scenarios will be some different)

B) Business-to-business

Chinese station

The relationship between business-to-business TPS and PV systems varies from system to application, with a rough estimate of 1:8 hours of relationship (09 traffic analysis data for offerdetail). Wang Pu and offerdetail the two proportions of the difference is very large, may be due to the high percentage of reptile temporary cause.

In the Taobao environment, assuming that we stress test the TPS is 100, then the system's daily throughput =100*11*3600=396 million

This is in the case of simple (single URL), some pages, a page has multiple request, the actual throughput of the system is also small. The relationship between TPS and other performance indicators

The following relationship is between TPS and concurrent virtual users (u_concurrent), LoadRunner read transaction response time (T_response) (Stable operation): Tps=u_concurrent/(T_response+t_think )

 TPs (Transaction per Second) summarizes

the number of concurrent users, the expected response time, and the TPS can be calculated. TPS is only used to calculate the expected value, and TPS in the performance test process cannot be used as a performance indicator alone. The theoretical value of the TPS data range should be between 10-100, below 10 and above 100, indicating that the system has a bottleneck point. The effect of transaction digital on execution time can be analyzed by comparing TPS with average transaction response time. Example: When the pressure increases, click-through/tps curve if the change is slow or has a flat trend, it is likely that the server began to bottleneck. TPS is a client-side view of server processing capabilities, can not prove to what extent TPS could support how much concurrency, the two are not necessarily linked.