Mitigating vendor lock-in in serverless cloud computing

Abstract

Vendor lock-in is a major problem in serverless computing because users rely heavily on the unique services, APIs and runtime environments offered by major cloud providers. This thesis uses a Systematic Literature Review (SLR) to examine how vendor lock-in occurs, why it remains difficult to avoid and which mitigation strategies are discussed in current research. The review identifies several contributing factors including provider-specific APIs, unique event models, tightly integrated managed services, pricing structures and the absence of common open standards. It also analyses proposed mitigation ways such as abstraction layers, adherence to open standards, multi-cloud deployment practices and open source serverless frameworks. These methods are evaluated with respect to their feasibility, implementation effort and influence on portability and architectural flexibility. The findings show that no single method can fully remove the risk of lock-in in serverless environments. Based on the literature, this thesis provides a summary and offers recommendations that can be applied to the organizations to reduce reliance to one provider and to increase long-term sustainability of architectures. The thesis findings conclude by offering recommendations to support more portable, resilient and sustainable serverless architectures. These findings address the central research question of how vendor lock-in in serverless computing can be effectively mitigated.