You are on page 1of 5

TYPES OF TESTS

BASELINE TEST:- Test the behavior of an application with minimum amount of load when the client
doesn’t define the specific requirements.

Objectives:

a) Validated the minimum user load supported by the application (Ex : 100 users)
b) Validated all transaction response times and make sure response times are within the SLA
c) Validate all performance parameters are within the minimum requirements like response times,
hits/sec, throughput, TPS

Users load - 100


Test Duration – 60 mins.

LOAD TEST:- Test an application or server behavior with an expected peak load.

Objectives:

a) Make sure the server will accept the expected peak loads.
b) Validate the response time’s behavior with the expected SLA’s.
c) Make sure the server will generate required transaction rate (TPS).

Ramp up = ramp up duration in seconds / load


Iterations = Expected transaction per hour / load =1000/50 ( 20 iterations per hour )
WORK LOAD MODEL FOR 1000USERS -30 MINUTS RAMPUP-2HOURS DURATION
SCENARIO LOA EXPECTED RAMP UP DURATIO ITERATION PACE TIME THINK TIME
NAME D TRANSACTI NS S
ON PER
HOUR
FT 400 - 2Users/9sec 2 Hrs - FIXED PT OF Random percentage of
60 SECONDS 75% - 125%
DD REQQUEST 100 - 1User/18sec 2 Hrs - FIXED PT OF Random percentage of
60 SECONDS 75% - 125%
STATEMENTS 200 - 1User/9Sec 2 Hrs - FIXED PT OF Random percentage of
60 SECONDS 75% - 125%
ACCOUNT 200 - 1User/9Sec 2 Hrs - FIXED PT OF Random percentage of
SUMMARY 60 SECONDS 75% - 125%
LOANS APPLY 50 1000 1User/36sec - 40 Fixed Interval Random percentage of
time of 180 75% - 125%
sec
CARDS APPLY 50 1000 1User/36sec - 40 Fixed Interval Random percentage of
time of 180 75% - 125%
sec
ENDURANCE TEST (OR) SOAK TEST:- Test the behavior of an application or a server with an average load
for a longer duration.

Objectives: -

a) Validate the stability or reliability of an application or a server for longer duration.


b) Identify any application memory leaks.

WORK LOAD MODEL FOR 500USERS -30 MINUTS RAMPUP-10HOURS DURATION

SCENARIO LOAD EXPECTED RAMP UP DURATIO ITERATIO PACE TIME THINK TIME
NAME TRANSAC NS NS
TION PER
HOUR
FT 200 - 2Users/9sec 10 Hrs - FIXED PT OF 60 Random percentage of
SECONDS 75% - 125%
DD REQQUEST 50 - 1User/18sec 10 Hrs - FIXED PT OF 60 Random percentage of
SECONDS 75% - 125%
STATEMENTS 100 - 1User/9Sec 10 Hrs - FIXED PT OF 60 Random percentage of
SECONDS 75% - 125%
ACCOUNT 100 - 1User/9Sec 10 Hrs - FIXED PT OF 60 Random percentage of
SUMMARY SECONDS 75% - 125%
LOANS APPLY 25 500 1User/36sec - 200 Fixed Interval Random percentage of
time of 180 sec 75% - 125%
CARDS APPLY 25 500 1User/36sec - 200 Fixed Interval Random percentage of
time of 180 sec 75% - 125%
STRESS TEST:-
Test the behavior of a server by gradually increasing the load until the server is crashed.

Objectives:-

a) At what amount of load server as crashed So that we can identify max capacity of given server.
b) We can observer sudden changes in response time behavior.

Note:- Normally we do the stress test, when client does not define the load

Page Level Stress:-

Apply the stress on a particular page by using a “rendezvous point”.

lr_ rendezvous(“Home”);

Top of the transaction.

Note: - We can apply this type of stress for the application where the load is suddenly increase on a
particular page.

Rendezvous point will be using to wait all the vusers and release on to the next transaction at the same
time.

SPIKE TEST: - Test the server behavior with the abnormal conditions of load either by increasing or
decreasing suddenly on the server.

Objectives:-

a) Checks the server behavior in abnormal situations.

b) Validate the response time’s behavior when the load is suddenly increased.

Note: - we can apply the spike test by using “Run or stop users”.
FAILOVER TEST:-

Test the behavior of “LOAD BALANCER (LB)” if one server fails in the middle of the execution.

Objectives:-

a) When the application in a live there is possibility some server may fail So we need to simulate
the same while doing the load test execution.

LG1/ 5 H/W Configuration


2500
Web-1
LG2/
2500 Web-2 App 1
LB Web-3
LB
App2
LG3/ Date Base (DB)
2500 S Web-4 App3

Round Robin Web-5


LG4/
2500 (Load is distributed)

Web Server App Server DB Server

CAPACITY (OR) VOLUMETRIC:-

Test the behavior of response time by keeping different volume of records in DB server. We do capacity
plan analysis with the future expected growth in DB records.

Note:- The DB team use the database “Bulk up loaders” to populate the dummy data in DB tables.
Controller:

1. What type of tests have you experienced?

Ans: Load/Stress/Endurance/Spike

2. What is the difference between thinktime and pacetime?

Ans:

Thinktime: It is a user waittime on any webpage for filling text fields or choose an items before
proceeding to the next webpage request

Pacetime: It is a user waittime between the iterations executions to controller the flow of
transactions (Static pacetime or dynamic pacetime)

3. Is it possible to increase the load while test is running?

Ans: Yes, It is possible to increase or decrease the load while test is running using Run/Stop vusers
functionalities in controller

4. What is rendezvous in loadrunner?

And: Rendezvous is used to hit all vusers simultaneously on a specific page at the same time.

lr_rendezvous("findflights");

5. I have a scenario like 1st hour 200 users, 2nd hour 400 users, and 3rd hour 500 user’s load to run
the test so how will you design the test?

You might also like