Information Technology Reference
In-Depth Information
Top Level:
Top Access Report (TAR)
a flat list of frequently accessed services in ranking order
(may be grouped by interconnection in customer usage scenarios)
Middle Level:
Unified Markov Models (UMMs)
for groups of TAR entries linked by CPR (call-pair report)
(may be expanded into lower-level UMMs or other models)
Bottom Level:
Detailed UMMs or other Models
associated with frequently visited or critical nodes of UMMs
(may correspond to testing models other than UMMs)
F
IG
. 4. Hierarchical implementation of an integrated web testing strategy.
approach. This implementation of our integrated strategy in a hierarchical form is
discussed below:
At the top level, TAR can be used directly as our flat OP for statistical usage-
based testing.
Entries in TAR can be grouped according to their connections via CPR, and a
UMM can be constructed for each of these groups, forming our middle-level
usage models, or our individual UMMs.
The hierarchical nature of our UMMs will allow us to have lower-level UMMs
as well as other lower-level testing models to thoroughly test selected functions
or web components.
This hierarchical implementation of our integrated strategy is graphically depicted
in
Fig. 4
. We focus on testing frequently used individual functions or services at
the top level, testing common navigation patterns and usage sequences at the mid-
dle level, and covering selected areas at the bottom level. Specific low-level UMMs
or other coverage-based testing models can be built to thoroughly test the related
features or critical components in the higher-level flat OPs or UMMs. Coverage,
criticality, and other information can also be easily used to generate test cases using
lower-level models under our OPs.
5 . 2 Te s t i n g H i g h - L e v e l C o m p o n e n t U s a g e w i t h M u s a
O p e r a t i o n a l P r o fi l e s
According to Musa
[33]
, an operational profile (OP) is a list of disjoint set of
operations and their associated probabilities of occurrence. It is a quantitative charac-
terization of the way a software system is or will be used.
Table II
gives an example
OP for the web site,
www.seas.smu.edu
,
listing the number of requests for differ-
