SLA-Driven Governance of RESTful Systems

The Software as a Service (SaaS) paradigm has become entrenched in the industry as a deployment model, bringing flexibility to the customers and a recurring revenue to the business. The main architectural paradigm of SaaS systems is the service-oriented one since it provides numerous advantages in terms of elasticity, fault tolerance, and flexible architectural design. Currently, the RESTful paradigm, a layer of abstraction on the server created by defining resources and entities that can be accessed by means of a URI, is the preferred choice for the construction of SaaS, as it promotes the deployment, isolation and integration of microservices through APIs. Nowadays, APIs are regarded as a new form of business product and ever more organizations are publicly opening up access to their APIs as a way to create new business opportunities. In the same way, other organizations also consume a number of third-party APIs as part of their business. We henceforth define the concept of a RESTful System as an information system following the RESTful paradigm to shape the integration model between both its own components as well as other information systems. Furthermore, understanding governance as the way in which a component is directed and controlled, in RESTful Systems, those components will be the RESTful APIs and what we aim to control or regulate is their behavior (i.e., how an API is being consumed or provided). As APIs are increasingly regarded as business products, a crucial activity is to describe the set of plans (i.e., the pricing) that depicts the functionality and performance being offered to clients. API providers usually define certain limitations in each instance of a plan (e.g., quotas and rates); for example, a free plan might be limited to having one hundred monthly requests, and a professional plan to have five hundred monthly requests. However, although API providers use the Service Level Agreement (SLA) concept to delimit the functionality and guarantees to which they commit to their customers, there is no standard model used by API providers for modeling API pricing (including the plans and limitations). Although some providers do model the information regarding the API pricing and API limitations with an ad hoc approach, there is no widely accepted model in the industry. Wherefore answering questions regarding API limitations (e.g., determining whether or not a certain pricing is valid) is still a manual or non-interoperable process coming along with some inconveniences (being tedious, time-consuming, error-prone, etc.). Understating governance as to how a system is directed and controlled, we translate this concept to meet the SLA-driven approach: we consider the SLA (i.e., API pricing) as the element that will drive the directions, policies and rules to deliver and maintain the RESTful System. Adding the SLA to the idea of governance of RESTful systems leads to the main hypothesis of this dissertation: there is no well-established model for describing API pricings)in RESTful systems, which is hindering the automatic SLA-Driven governance. We claim the main goal of this thesis to be: the creation of an expressive, fully-fledged specification of SLAs for RESTful APIs endorsed with an open ecosystem of tools aimed at the SLA-Driven Governance of RESTful systems. The results of this endeavor are twofold: (I) Creation of a sufficiently expressive specification for the description of API pricings and the analysis of their validity. This comprises: (i) conducting an analysis of real-world APIs to evaluate the characteristics of the API pricings and limitations; (ii) identifying the relevance of SLAs in APIs in both academic and industrial scenarios; (iii) proposing a comprehensive model for describing API pricings; (iv) defining analysis operations for common questions regarding the validity in API pricings and limitations; (v) performing an evaluation of the model in real-world APIs. (II) Implementation of an ecosystem of tools to support the SLA-Driven governance of RESTful APIs. This includes: (i) developing a set of API governance tools; (ii) implementing a validity analysis operation; (iii) performing a validation of the tools and operations in realistic scenarios. In this thesis, we present the Governify4APIs ecosystem as the set comprised of (i) a model aimed at describing API pricings that is closely aligned with industry standards in APIs (OpenAPI Specification) and (ii) a set of companion tools for enacting the automatic governance using our specification, ranging from low-level validation tasks to SaaS solutions based on our model. Governify4APIs is, therefore, a fully-fledged specification, aligned with the mainstream standards and intended to enable an SLA-Driven Governance of RESTful Systems.

Author dc:creator

• Gámez Díaz, Antonio

Advisors dc:contributor.advisor

• Fernández Montes, Pablo
• Ruiz Cortés, Antonio

Statement dc:rights

• Attribution-NonCommercial-NoDerivatives 4.0 Internacional

Licence dc:rights.uri

• http://creativecommons.org/licenses/by-nc-nd/4.0/

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/11441/130656

OAI identifier oai:identifier

oai:idus.us.es:11441/130656