Page 14 - Methodology for measurement of Quality of Service (QoS) Key Performance Performance
Indicators (KPIs) for Digital Financial Services
P. 14
TABLE 3-2: Role and activity assignments during a 4-transaction cycle
DEVICE TA1 TA2 TA3 TA4
SP1
Person 1
FP1
SP2
Person 2
FP2
OP1 operated by P2 P2 P1 P1
than qualitative, it may not practical and is not nec- 3.5 Automation of tests
essarily required to exclude frequency-dependent ef- The methodology in the present document describes
fects entirely. However, respective effects need to be testing in a generic way, i.e. service tests can be done
recorded and documented carefully as part of the manually as well as in an automated way. It is under-
reporting in order to understand their impact on the stood that automation of tests is desirable to achieve
testing conditions. a greater degree of repeatability, and less variation in
With respect to guard times, it is conceivable that the quantitative data values due to inaccuracy of e.g. man-
system has a certain “dead time” after each transaction, ual time measurements. It is likewise understood that
where the system would not accept a new transaction such automation requires a higher initial effort to ensure
or create unexpected results of a transaction attempted reliability of operation under unsupervised conditions
within this period. It is advisable to be aware of this pos- or to cover a wider range of end-user devices.
sibility and obtain respective information before actual
parameters of a test campaign are determined. Techni-
cally, it would also be possible to probe for such effects. 4 TRANSACTION MODEL
This would, however, require a sufficiently controlled ex-
ecution of testing, i.e. automated test control with the 4.1 Person to Person (P2P) Mobile Money (MoMo)
ability to systematically reduce inter-transaction time transfer
and check for related effects.
4.1.1 Transaction description
3.3 Re-initialization after unsuccessful transactions Abstract: Transfer of a known amount of M units of
If a transaction fails, in particular after a time-out con- money from account A to account B.
dition has occurred, it shall be ensured that the service Success definition: The correct amount plus applica-
and the device or application are in the typical neutral ble operators’ fees have been deducted from the A
starting state again, i.e. that no memory of previous er- party account and the correct amount (net) has been
ror states remains in the system. credited to the B party account within the defined
time window.
3.4 Disappeared money Examples for unsuccessful execution are cases
It is possible that during a transaction, the amount of • Where the system sends—at any stage of the trans-
money deducted is not correct with respect to trans- fer—an explicit response indicating failure of the
ferred amount and fees. This includes the case that the transfer.
amount is correct but sent to a third party by an error
in the system. From an end customer perspective, this • Where the transfer has been done but the amount is
is either a loss (if too much money is deducted), or an wrong.
unjustified gain (if money is credited but not deducted • Timed-out and still pending TA.
on the other side of the transaction. For simplicity, we
use the term “disappear” for both variants of this kind ■ NOTE 1: The description does not explicitly refer
of effect. to assignment of roles to devices or operators. For
instance, if a particular device is assigned to repre-
■ NOTE 1: In cases of disappeared money, insertion of sent a given account, the device may be operated
fresh money will be necessary.
as A Party or B Party. Related events occur, and
■ NOTE 2: Retrieval of lost money should be treated as related activities are performed, on the respective
a second stream of activities. device.
12 • Methodology for measurement of Quality of Service (QoS) Key Performance Indicators (KPIs) for Digital Financial Services