Hyppää sisältöön
    • Suomeksi
    • På svenska
    • In English
  • Suomeksi
  • In English
  • Kirjaudu
Näytä aineisto 
  •   Etusivu
  • LUTPub
  • Diplomityöt ja Pro gradu -tutkielmat
  • Näytä aineisto
  •   Etusivu
  • LUTPub
  • Diplomityöt ja Pro gradu -tutkielmat
  • Näytä aineisto
JavaScript is disabled for your browser. Some features of this site may not work without it.

Data consistency challenges in SaaS enterprise integrations

Ahammed, Sabbir (2026)

Katso/Avaa
Mastersthesis_Ahammed_Sabbir.pdf (1.117Mb)
Lataukset: 


Diplomityö

Ahammed, Sabbir
2026

School of Engineering Science, Tietotekniikka

Kaikki oikeudet pidätetään.
Näytä kaikki kuvailutiedot
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20261021136005

Tiivistelmä

Software-as-a-service (SaaS) products increasingly depend on seamless and reliable data exchange with external systems used by their customers. This thesis investigates data consistency issues occurring when hired candidates data is transferred to external human resource information system (HRIS) from an applicant tracking system (ATS), that the ATS provider neither owns nor controls. The research was carried out within a small Finnish SaaS company integrated with more than 15 HRIS platforms, where, during the study period, most documented failures were first reported by customers.

The following research was conducted using the Design Science Research Methodology. Seven documented transfer-failure cases and one recurring structural condition, identified from the organization’s issue tracker between 2023 and 2026, were analyzed with a hybrid inductive–deductive coding procedure and compared with the literature on distributed-systems consistency, enterprise integration, application programming interface (API) evolution and data quality. A taxonomy was designed that classifies failures by cause, outcome, visibility and detection, and a mitigation framework was also proposed that maps them to detection, prevention and recovery measures suited to a small, consumer-side engineering team. The artifacts were demonstrated retrospectively against the documented cases and assessed through a written questionnaire completed by two engineers who work on the integrations.

The failures were grouped into four patterns including required data absent or unmapped, invalid values, external interface and validation change, and stale value mappings. The most consequential root cause was unannounced change on the vendor side, where incorrect values silently accepted by the receiving system demonstrated the most damaging class of failure. The evaluation validated the structure of the taxonomy, identified the coverage is limited to the integration boundary, and shifted the framework more toward automation. The refined framework starts with alerts based on failure records that are already maintained by the organization, followed by stops for unrecognized values from passing through silently, and validates payloads against vendor schemas where automated access is possible. The results demonstrated that when a receiving system cannot be directly controlled, prioritizing failures visibility can offer greater benefit than prevention mechanisms that depend on vendor cooperation.
Kokoelmat
  • Diplomityöt ja Pro gradu -tutkielmat [15768]
LUT-yliopisto
PL 20
53851 Lappeenranta
Ota yhteyttä | Tietosuoja | Saavutettavuusseloste
 

 

Tämä kokoelma

JulkaisuajatTekijätNimekkeetKoulutusohjelmaAvainsanatSyöttöajatYhteisöt ja kokoelmat

Omat tiedot

Kirjaudu sisäänRekisteröidy
LUT-yliopisto
PL 20
53851 Lappeenranta
Ota yhteyttä | Tietosuoja | Saavutettavuusseloste