Abstract

Ontologies have become the de-facto information representation method in the semantic web domain, but recently gained popularity in other domains such as cloud computing. In this context, ontologies enable service discovery, effective comparison and selection of IaaS, PaaS and SaaS offerings and ease the application deployment process by tackling what is known as the vendor lock-in problem. In this paper we propose a novel ontology named HOCC: holistic ontology for effective cluster comparison. The ontology design process is based on four different information categories, namely Performance, SLA, cost and environmental impact. In addition we present our approach for populating, managing and taking advantage of the proposed ontology as developed in a real world Kubernetes cluster setting, as well as instantiating the ontology with example services and data (namely performance aspects of a serverless function).

Authors

• Yannis Poulakis,
• Georgios Fatouros,
• George Kousiouris
• Dimosthenis Kyriazis

The post HOCC: an ontology for the holistic description of cluster settings as you have correctly identified appeared first on PHYSICS.

Abstract

Serverless environments have attracted significant attention in recent years as a result of their agility in execution as well as inherent scaling capabilities as a cloud-native execution model. While extensive analysis has been performed in various critical performance aspects of these environments, such as cold start times, the aspect of workflow orchestration delays has been neglected. Given that this paradigm has become more mature in recent years and application complexity has started to rise from a few functions to more complex application structures, the issue of delays in orchestrating these functions may become severe. In this work, one of the main open source FaaS platforms, Openwhisk, is utilized in order to measure and investigate its orchestration delays for the main sequence operator of the platform. These are compared to delays included in orchestration of functions through two alternative means, including the execution of orchestrator logic functions in supporting runtimes based on Node-RED. The delays inserted by each different orchestration mode are measured and modeled, while boundary points of selection between each mode are presented, based on the number and expected delay of the functions that constitute the workflow. It is indicative that in certain cases, the orchestration overheads might range from 0.29% to 235% compared to the beneficial computational time needed for the workflow functions. The results can extend simulation and estimation mechanisms with information on the orchestration overheads.

Authors

• George Kousiouris
• Chris Giannakos
• Konstantinos Tserpes
• Teta Stamati

The post Measuring Baseline Overheads in Different Orchestration Mechanisms for Large FaaS Workflows appeared first on PHYSICS.

Abstract

Edge Computing (EC) performs computation at the close proximity to the end devices and reduces dependency to the Cloud and Internet. It also overcomes the Quality of Service (QoS) and latency issues that come naturally from the best- effort behaviour of the Cloud. Although reduced dependency to the Internet enables the application of the EC to the real- time applications, without a standard architecture, EC benefits cannot be incorporated. To this end, in our previous papers, we proposed a novel software reference architecture (SRA) for Edge Servers. The SRA allows receiving Edge Computing benefits without worrying about setup or configuration, but only focusing on software development. Regardless of the computing power of the Edge Servers, the architecture enables a decentralised (real- time) task execution at the edge. Once a task request arrives, the receiving server determines whether to execute the task on itself or offload it to any of the neighbouring servers, by considering several parameters, such as latency, and available resources. In this paper, we explain how servers decide where to execute the task, without requiring a centralized load balancer. Moreover, we show how they share their resources between each other to create a common knowledge of their neighbouring servers.

Index Terms—real-time computing, edge computing, task of- floading, fog computing, decentralized load balancer.

Authors

• Volkan Gezer
• Achim Wagner

The post Real-Time Edge Framework (RTEF): Decentralized Decision Making for Offloading appeared first on PHYSICS.

Modern cloud computing advances have been pressing application modernization in the last 15 years, stressing the need for application redesign towards the use of more distributed and ephemeral resources. From the initial IaaS and PaaS approaches, to microservices and now to the serverless model (and especially the Function as a Service approach), new challenges arise constantly for application developers. This paper presents a design and development environment that aims to ease application evolution and migration to the new FaaS model, based on the widely used Node-RED open source tool. The goal of the environment is to enable a more user friendly and abstract function and workflow creation for complex FaaS applications. To this end, it bypasses workflow description and function reuse limitations of the current FaaS platforms, by providing an extendable, pattern-enriched palette of ready-made, reusable functionality that can be combined in arbitrary ways. The environment embeds seamless DevOps processes for generating the deployable artefacts (i.e. functions and images) of the FaaS platform (Openwhisk). Annotation mechanisms are also available for the developer to dictate diverse execution options or management guidelines towards the deployment and operation stacks. The evaluation is based on case studies of indicative scenarios, including creating, registering and executing functions and flows based on the Node-RED runtime, embedding of existing legacy code in a FaaS environment, parallelizing a workload, collecting data at the edge and creating function orchestrators to accompany the application. For the latter, a detailed argumentation is provided as to why this process should not be constrained by the “double billing” principle of FaaS.

Authors

• George Kousiouris, 
• Szymon Ambroziak, 
• Domenico Costantino, 
• Stylianos Tsarsitalidis, 
• Evangelos Boutas, 
• Alessandro Mamelli, 
• Teta Stamati

The post Combining Node-RED and Openwhisk for Pattern-based Development and Execution of Complex FaaS Workflows appeared first on PHYSICS.