On Formalizing and Identifying Patterns in Cloud Workload Specifications
Christos Tsigkanos, Timo Kehrer · 2016
Managing, configuring and deploying complex applications in the cloud are emerging problems in contemporary cloud computing. Cloud workload specifications, as portable abstractions of cloud computations, are an important means to deal with these problems on an architectural level. They focus on defining the components of an application and their structural relations, workflows regarding their initialization and management, along with configuration and artifacts required for application operation. The prevalence of cloud computing drives demand for applications that rely on systematic engineering and smell-free architectures, thus quality assurance techniques for cloud workload specifications are strongly required. In particular, cloud workload specifications exhibit characteristics of software architectures such as patterns or anti-patterns which state desired or undesired quality aspects. To facilitate formal reasoning about latent qualities of a workload design, we propose a static bigraphical semantics for the modeling language defined by the emerging Topology and Orchestration Specification for Cloud Applications (TOSCA) standard. Thereupon, we illustrate how to check for the presence (absence) of (anti-)patterns expressed as logical formulae over bigraphical predicates.