| Summary: | With the increase of iOS applications on the market the demand and use of Backend as a Service (BaaS) providers also increase. In an early phase of the development it is beneficial for a potential application publisher to use a BaaS to quickly reach the market. Over time the provided services may be inadequate which make many BaaS users migrate to a custom developed backend. This paper intends to investigate which BaaS provider gives the least dismissed effort when making a transition to a custom developed backend with the purpose of providing basis for potential application publishers in the selection of a provider, given that a future transition to a custom backend will occur. From a population of ten providers, five were randomly selected – Firebase, Kinvey, CloudMine, Kumulos and Kii. In order to measure required effort for each provider, code that is tightly coupled to each provider’s SDK was implemented, according to provider guidelines and documentation. The implementations were measured with the COCOMO II model which gives a result in terms of required person months (PM). The measured PM of each implementation was compared. The hypothesis of the study could be rejected if the resulting PM of two implementations were disjointed. The result and analysis show difference in PM which lead to a rejection of the hypothesis. Whether the assumptions of the organization, product and project affected the results were analysed and the hypothesis was rejected regardless of these assumptions. If the organization of a potential application publisher resembles the one in the research Firebase is the recommended choice of BaaS provider.
|
|---|
