Page 37 - Methodology for measurement of Quality of Service (QoS) Key Performance Performance
Indicators (KPIs) for Digital Financial Services
P. 37
E.3.2 Notification SMS data table structure The remaining orphans are again sorted into catego-
TABLE E-2: Notification SMS data table structure ries:
ELEMENT TYPE – A or B side SMS which have matching DFS transac-
tions. This indicates transactions where such SMS are
Device ID Varchar(64)
missing and shall be notified along accordingly.
Import date and time datetime
– A or B side SMS which have no matching DFS trans-
SMS content Mirroring of XML structure
action. An investigation shall be conducted to clarify
the circumstances.
E.3.3 Assignment of primary test data and SMS
Each successful DFS transaction is assumed to produce E.3.4 Storage and deletion aspects of SMS on devices
a set of primary data (timestamp information according The process of SMS back-up is based on periodic copies
to the definitions elsewhere in the present document) of all SMS on a particular device.
and two confirmation SMS on the A and B device, re- In the course of the test campaign, locally stored
spectively. SMSD will accumulate unless they are deleted. Dele-
By processing SMS back-up copies uploaded to the tion procedures carry the risk of unwanted deletion of
data base, these SMS are assigned. There are two basic meaningful data. A hard cause to delete SMS would be
types (A side and B side SMS). There can be other SMS capacity issue. Unless this is given, it is assumed to be
on the device. Therefore, the classification and assign- better to handle SMS duplicates — which is technically
ment process has the following stages: quite simple in data processing — than to run a deletion
i) Identify if a SMS is of type A-side, B-side or other. process.
In case a deletion process is required after all, it is per-
ii) If A-side, attempt to find the matching B side SMS formed along the following process:
from another device.
i) There are regular device maintenance cycles (e.g.
iii) If B-side, attempt to find the matching A-side SMS once per week) where all devices used in the test
(actually steps 2 and 3 are symmetrical). campaign are participating.
iv) Attempt to find the matching primary transaction ii) From processing of previously uploaded data, a ref-
for the A-side and B-side SMS, respectively, using erence point in time for DFS confirmation SMS is
device/team allocation and timestamp calculated (SMS-RP, type time/date). It is assumed
Ideally, the process assigns all A-side and B-side SMS. that up to this RT, all uploaded SMS are checked
It is expected that “orphans” exist which do not have a and assigned (see Assignment of primary test data
counterpart. For such orphans, the first step is to check and SMS) and that any hints on missing SMS (de-
if SMS exist on devices which have not been covered by tected as missing in uploaded data, to be checked
the backup process. If this check finds previously miss- for on devices) are clarified.
ing SMS, they shall be processed. iii) In the maintenance process, all locally stored SMS
older than the SMS-RP are deleted.
Methodology for measurement of Quality of Service (QoS) Key Performance Indicators (KPIs) for Digital Financial Services • 35Methodology for measurement of Quality of Service (QoS) Key Performance Indicators (KPIs) for Digital Financial Services • 35