Defining and Measuring Efficiency gains through Infrastructure as Code Adoption: A MultiVocal Literature Review and a case study on a company
Author
Petrovics, Alexandru-Constantin
Term
4. term
Education
Publication year
2024
Pages
83
Abstract
Dette speciale undersøger, hvordan man vurderer effektiviteten af Infrastructure as Code (IaC) – at styre IT-infrastruktur med kode. Første del er en multi-vokal litteraturgennemgang, der samler viden fra forskning og praksis. Den identificerer fire centrale resultatområder: højere hastighed og lavere omkostninger; skalerbarhed og standardisering; stærkere sikkerhed og bedre dokumentation; samt katastrofehåndtering/beredskab. Derudover kortlægges 22 nøgletal. De mest nævnte er: gennemløbstid for ændringer (hvor hurtigt ændringer kommer fra idé til produktion), tid til genopretning af service (hvor hurtigt systemer er oppe igen efter fejl), fejlrate for ændringer (andelen af ændringer der giver problemer), udrulningsfrekvens (hvor ofte man deployer), og varighed af udrulninger (hvor længe en udrulning tager). Anden del er et casestudie hos Jumia, der belyser virksomhedens tilgang til IaC-effektivitet. Interviews fremhævede især disse metrikker: infrastruktur-kodedækning (hvor stor en del af infrastrukturen der er beskrevet i kode), tid til genopretning af service, antal driftsafvigelser (forskelle mellem det, koden beskriver, og det, der faktisk kører), udrulningsfrekvens, gennemløbstid for ændringer og fejlrate for ændringer. Disse metrikker ligger tæt på dem i litteraturen, hvilket peger på en fælles forståelse af IaC-effektivitet mellem akademia og industri. Ligeledes stemmer de fordele og ulemper, Jumia oplever, overens med litteraturen. Samlet peger resultaterne på, at selv om der er bred enighed, bør evaluering af IaC-effektivitet tilpasses den enkelte organisations behov.
This thesis examines how to evaluate the efficiency of Infrastructure as Code (IaC)—managing IT infrastructure with code. The first part is a multi-vocal literature review that synthesizes insights from research and practice. It identifies four outcome areas: faster delivery and lower costs; scalability and standardization; stronger security and better documentation; and disaster recovery/management. It also compiles 22 key metrics. The most frequently cited are: lead time for changes (how quickly changes move from idea to production), time to restore service (how fast systems recover after an incident), change failure rate (the share of changes that cause problems), deployment frequency (how often releases happen), and deployment time duration (how long each release takes). The second part is a case study at Jumia that explores the company’s approach to IaC efficiency. Interviews highlighted these metrics in particular: infrastructure code coverage (how much of the infrastructure is defined in code), time to restore service, number of drifts (configuration drift—differences between what code specifies and what is actually running), deployment frequency, lead time for changes, and change failure rate. These priorities closely match the literature, suggesting a shared understanding of IaC efficiency between academia and industry. The benefits and drawbacks reported by Jumia also mirror published sources. Overall, while there is broad consensus, IaC efficiency should be assessed using metrics tailored to each organization’s needs.
[This summary has been rewritten with the help of AI based on the project's original abstract]
Documents