|
|
|
|
|
 |
SIP Performance |
|
| Juphoon SIP
Performance Test |
Test Goal
As a part of protocol application, SIP Codec
generally occupies 40% to 60% of the CPU resource ,
and the performance of codec are more comparable. So
we tested 3 third party SIP stacks' and Juphoon SIP
stack's codec module.
Test Condition
The same common SIP
signals, with the same hardware (Celeron mobile
1.3G, 384 RAM). Juphoon SIP setting was parsing all
headers and SDP body in detail.
Test Cases
13 cases: INVITE 、 180 、 183 、 200 、 ACK
、 BYE 、 INFO 、 PRACK 、 INVITE(ISUP) 、 200(BYE) 、
200(INFO)、 200(PRACK1)、 200(PRACK2)
|
 |
Tester message length: 5150 byte, header counted: 109 , repeat: 30000 times,
cost time unit: second . Codec : Decode & Encode
DV : Debug Version
RV : Release Version
|
| Juphoon has made many performance tests with 3rd-bussiness SIP stacks, our SIP stack performance is faster than 3-9 times averagely. |
| SIP Parser Precision |
| Juphoon SIP parses sytax tree to concret element, user just only use value(int, enum type) while reading sip message, and use only need assign a value(int, enum type) while formating sip message. So it is very convenient for programmer, the binary code only need store one element data while using these token strings. |
|
| OSIP 2-3.0.1 parses syntax tree to string element, user must convert string to value while reading sip message, and user must convert value to string while formating sip message. |
|
| Juphoon SIP Call Tester |
| Call Tester is used to test the SIP Stack, making
the stack remain correct and stable after
modifications and extensions. We provide Call Tester
to simulate the call flow, which could be normal
case such as value-added services and abnormal case.
The call test role could be SIP UA or just call TPT
user. The Call Tester also could be call engine for
other VoIP protocol (for example, MGCP, H.322 and
etc.) .
|
 |
|
|
|
| |
|
|
| |
|
|