TTCN

1
Methods for telecommunication protocol
development and conformance testing
Extracts for TTCN
Authors Róbert Horváth, Zoltán Pap, Zoltán
Rétháti Technical University of Budapest,
Electrical Engineering rob_at_, pap_at_,
rethati_at_ttt-atm.ttt.bme.hu
Consultants Katalin Tarnay, NOKIA WSS Gyula
Csopaki, Gusztáv Adamis Departement of
Telecommunication and Telematics
2
The Tree and Tabular Combined NotationTTCN
ITU-T Recommendation - Standard
TTCN 3 is coming!

• Complete description of dynamic real time system
  test suites and testing methods
• based on black box model
• hierarchical structure of test suites
• communication signals
• TTCN -MP and TTCN -GR

3
TTCN
Black Box Model
Approach of testing the structure of the system
is not known, there are interfaces (PCOs) to
communicate
1. Stimulate the system with a signal 2. Wait for
the answer signal and observe it 3. Compare with
the standard Verdicts Pass Fail Inco
nclusive
4
TTCN the Test Suite of the protocol
The goal of the work was to create the Test Suite
of the WTP protocol, and try to simplify the test
process.
5
TTCN
6
TTCN
TTCN - MP for computer, TTCN - GR (Tables) for
human development
7
TTCN Behaviour Description
8
PICS
9
PIXIT
10
The WAP and the WAP WTP Protocol
11
The WTP
Message and transaction oriented Services

• Class 0 Unreliable invoke message with no result
  message (Unreliable one-way requests)
• Class 1 Reliable invoke message with no result
  message (Reliable one-way requests)
• Class 2 Reliable invoke message with exactly one
  reliable result message (Reliable two-way
  request-reply)

12
The test generation process
Conformance testing idealized requirements

• The test should cover the whole protocol (check
  all the possible functional behaviors coverage
  value).
• The test should check abnormal situations,
  observe reactions to improper events.
• The test suite should be so created that the the
  test cases, could be executed separately.

13
The test generation process The traditional
methods

• Highly qualified experienced engineers
• Tests directly in C or in formal test language
  (e.g. Tree and Tabular Combined Notation (TTCN)
• No inner step, directly from the standard
  (informal textual description)
• Knowledge-based, time consuming work
• High costs, lower reliability and not always
  complete
• Long learning phase (to get deeply acquainted)

14
The test generation process The CATG methods

• Time gain, but formal unambiguous specification
  as input
• The bigger problem selecting the algorithm
  (NP-complete state space exploration)
• Ratio of pointless test cases increases.
• They have different length, there are no formal
  rules, etc.
• Unstructured stack overview, execution of
  subsets according to separable protocol functions
  is almost impossible.
• No proper and reasonable handling of inopportune
  and invalid test cases.

15
The test generation process

•   What are the test purposes? Translate into
  unambiguous, well-located concrete tasks!

•   What are the formal limits of the test cases?
  How to choose them, how to match them to the
  original test purposes? Long test cases? Shorter
  ones?

• How to guarantee the satisfactory coverage?

• How to collect these tests into groups? Invalid
  and inopportune test cases?

16
The test generation processStandardized Event
Pair Based Test Generation Method Using TSSTP

• Each test case has to verify one transition of
  the SDL diagram event pairs

• The formal appearance and the derivation
  algorithm have to be common for all test cases.

17
Test Suite Structure and Test Purposes
Test Suite Structure Class 0 Class 1 Class
2 With UACK Without UACK Capability Vali
d Invalid Inopportune
18
Test Suite Structure and Test Purposes
Test Purposes
19
Derivation of Test Purposes
Initial information
Informal specification
Table description
20
Derivation of Test Purposes
First step
Complete SDL specification
21
Derivation of Test Purposes
Second step
Test Suite Structure Test Purposes
22
Thank you for your attention!