A Middleware for Managing Non-Functional Requirements in Cloud PaaS

Abstract

A key aspect of distributed applications in platformas-a-service clouds revolves around the definition of requirements for quality of service and the most efficient use of the available resources to achieve them. Matching both the requirements and resources represents a unique challenge: While requirements are generally defined in relatively abstract form and defined as system-wide parameters, the resources are concrete, low-level entities that, by themselves, are only of marginal influence on the requirements. In order to establish a relationship between these two requires both the agglomeration of properties into more abstract ones as well as refining requirements in more detailed, concrete ones, both of which can often only be done to a certain degree. This paper presents a vision how both aspects can be brought to the same level and introduces an approach enhancing component-based cloud applications with non-functional properties in the context of an integrated programming model. This allows developers of such applications to formally declare properties that may influence the application during runtime and document the expectations regarding non-functional aspects of the overall system. This can be used to ensure the health of the application and, in case requirements are violated, actions like the deployment of additional resources can be performed in order to alleviate any degradation of its non-functional state, either by evaluating, choosing and performing pre-defined actions automatically or by manual administrative intervention.

Designing service-based applications in the presence of non-functional properties: A mapping study

“…Model [57], [36], [90], [62], [83], [169], [7], [99], [95], [21], [32], [204], [184], [150], [175], [23], [209], [156], [191], [9].…”

MAS-based self-adaptive architecture for controlling and monitoring Cloud platforms

“…The number of studies that can be found on the state of the art relating CC with agent technology is actually low, but this tendency is changing and it is becoming increasingly common to ind studies and applications focused on this ield. Despite the limited number of studies on the matter, Agent-based Cloud platform is becoming a common concept, mentioned by various authors in recent years (Talia 2011(Talia , 2012Kang et al 2010;Venticinque et al 2011;Braubach et al 2014;Cao et al 2009).…”

Elastic component‐based applications in PaaS clouds

“…Currently, Jadex already supports a large number of predefined NFPs including, for example, CPU, memory, timing, and service measurements. More details about the available NFPs can be found in . Using, for example, the LatencyProperty, the average latency (using an exponential moving average) of service calls can be measured and the WaitqueueProperty keeps track of the number of incoming but not yet served service invocations.…”