Miguel Angel López Peña, Isabel Muñoz Fernández

Current new IoT standards do not detail enough some important and emergent aspects as the Fog/Edge computing support, the IoT computation topology management or the IoT visualization systems. This work defines three new concepts: a) the paradigm of edge/cloud computing transparency that lets the computation nodes change dynamically without administrator intervention; b) the IoT computing topology management that gives an IoT system global view, from the hardware and communication infrastructures to the software deployed on them, and c) the automation and integration of IoT visualization systems for real time data visualization, current IoT topology and current paths of data flows. It is also defined a new architectural model that includes these concepts and covers other IoT demands, like security safeguard services based on Blockchain. This architectural model definition is taken as the basis for developing a new advanced IoT platform referred as SAT-IoT.

Daniel Seybold, Moritz Keppler, Daniel Gründler, Jörg Domaschka

Big Data and IoT applications require highly-scalable database management system (DBMS), preferably operated in the cloud to ensure scalability also on the resource level. As the number of existing distributed DBMS is extensive, the selection and operation of a distributed DBMS in the cloud is a challenging task. While DBMS benchmarking is a supportive approach, existing frameworks do not cope with the runtime constraints of distributed DBMS and the volatility of cloud environments. Hence, DBMS evaluation frameworks need to consider DBMS runtime and cloud resource constraints to enable portable and reproducible results. In this paper we present Mowgli, a novel evaluation framework that enables the evaluation of non-functional DBMS features in correlation with DBMS runtime and cloud resource constraints. Mowgli fully automates the execution of cloud and DBMS agnostic evaluation scenarios, including DBMS cluster adaptations. The evaluation of Mowgli is based on two IoT-driven scenarios, comprising the DBMSs Apache Cassandra and Couchbase, nine DBMS runtime configurations, two cloud providers with two different storage backends. Mowgli automates the execution of the resulting 102 evaluation scenarios, verifying its support for portable and reproducible DBMS evaluations. The results provide extensive insights into the DBMS scalability and the impact of different cloud resources. The significance of the results is validated by the correlation with existing DBMS evaluation results.

Benjamin Schanzel, Mark Leznik, Simon Volpert, Jörg Domaschka, Stefan Wesner

Providing runtime dependencies for computational workflows in shared environments, like HPC clusters, requires appropriate management efforts from users and administrators. Users of a cluster define the software stack required for a workflow to execute successfully, while administrators maintain the mechanisms to offer libraries and applications in different versions and combinations for the users to have maximum flexibility. The Environment Modules system is the tool of choice on bwForCluster BinAC for this purpose. In this paper, we present a solution to execute a workflow which relies on a software stack not available via Environment Modules on BinAC. The paper describes the usage of a containerized, user-defined software stack for this particular problem using the Singularity and Docker container platforms. Additionally, we present a solution for the reproducible provisioning of identical software stacks across HPC and non-HPC environments. The approach uses a Docker image as the basis for a Singularity container. This allows users to define arbitrary software stacks giving them the ability to execute their workflows across different environments, from local workstations to HPC clusters. This approach provides identical versions of software and libraries across all environments.

Constantine Ayimba, Paolo Casari, Vincenzo Mancuso

Infrastructure providers employing Virtual Network Functions (VNFs) in a cloud computing context need to find a balance between optimal resource utilization and adherence to agreed Service Level Agreements (SLAs). Tenants should be allocated as much computing, storage and network capacity as they need in order not to violate SLAs, but not more so that the infrastructure provider can accommodate more tenants to increase revenue. This paper presents an optimizer VNF that ensures that a given virtual machine (VM) is sufficiently utilized before directing traffic to another VM, and an orchestrator VNF that scales the number of VMs up or down as needed when workloads change, thereby limiting the number of active VMs to a minimum that can deliver the service. We setup a testbed to transcode and stream Video on Demand (VoD) as a service. We present experimental results which show that when the optimizer and orchestrator are used together they outperform static provisioning in terms of both resource utilization and service response times.

Radhika Loomba, Thijs Metsch, Leonard Feehan, Joe Butler

Optimal placement of service chains (composed of multiple connected service components) on heterogeneous cloud/fog infrastructure is a challenging problem. Even when deploying a small set of service components, the search-space of all possible solutions is quite large. Furthermore, current approaches sacrifice either precision or scalability when presented with heterogeneous platform configurations. In this scenario, the goal is to support orchestration that optimally selects solutions that meet performance constraints, leverages differentiating platform features and delivers these solutions within reasonable run-time. This paper presents the novel Hybrid Fitness-Utility Algorithm that addresses these issues by utilizing concepts derived from evolutionary algorithms to dimensionally reduce complexity, and incorporates the utility of placing each individual service component on an infrastructure resource for optimized selection between potential solutions. Our results show that the algorithm succeeds in determining optimal service chain placement onto distributed infrastructure for a simulated cloud/fog scenario and for a live multi point-of-presence OpenStack-based testbed with over 90 percent confidence.

Matheus Felipe Ferreira da Silva Lisboa Tigre, Guto Leoni Santos, Theo Lynn, Djamel Sadok, Judith Kelner, Patricia Takako Endo

The Internet of Things has the potential of transforming health systems through the collection and analysis of patient physiological data via wearable devices and sensor networks. Such systems can offer assisted living services in realtime and offer a range of multimedia-based health services. However, lack of service availability, particularly in the cases of emergencies, can lead to adverse outcomes and in the worst case, death. In this paper, we propose an e-health monitoring architecture based on sensors and cloud and fog infrastructure scenarios. Further, we propose stochastic models to analyze how failures impact on the e-health system availability. We analyze four different scenarios and from results, we identify that the sensors and fog devices are the components that have the most significant impact on the availability of the entire e-health system in the scenarios analyzed.

Cleber Matos de Morais, Patricia Endo, Sergej Svorobej, Theo Lynn

RECAP is a European Union funded project that seeks to develop a next-generation resource management solution, from both technical and business perspectives, when adopting technological solutions spanning across cloud, fog, and edge layers. The RECAP project is composed of a set of use cases that present highly complex and scenario-specific requirements that should be modelled and simulated in order to find optimal solutions for resource management. Due use cases characteristics, configuring simulation scenarios is a high time consuming task and requires staff with specialist expertise.

Jakub Krzywda, Ahmed Ali-Eldin, Eddie Wadbro, Per-Olov Östberg, Erik Elmroth

Server power capping limits the power consumption of a server to not exceed a specific power budget. This allows data center operators to reduce the peak power consumption at the cost of performance degradation of hosted applications. Previous work on server power capping rarely considers Quality-of-Service (QoS) requirements of consolidated services when enforcing the power budget. In this paper, we introduce ALPACA, a framework to reduce QoS violations and overall application performance degradation for consolidated services. ALPACA reduces unnecessary high power consumption when there is no performance gain, and divides the power among the running services in a way that reduces the overall QoS degradation when the power is scarce. We evaluate ALPACA using four applications: MediaWiki, SysBench, Sock Shop, and CloudSuite’s Web Search benchmark. Our experiments show that ALPACA reduces the operational costs of QoS penalties and electricity by up to 40% compared to a non optimized system.

Thang Le Duc, Per-Olov Ostberg

Content Delivery Networks (CDNs) are handling a large part of the traffic over the Internet and are of growing importance for management and operation of coming generations of data intensive applications. This paper addresses modeling and scaling of content-oriented applications, and presents workload, application, and infrastructure models developed in collaboration with a large-scale CDN operating infrastructure provider aimed to improve the performance of content delivery subsystems deployed in wide area networks. It has been shown that leveraging edge resources for the deployment of caches of content greatly benefits CDNs. Therefore, the models are described from an edge computing perspective and intended to be integrated in network topology aware application orchestration and resource management systems.

Demetris Trihinas, Luis F. Chiroque, George Pallis, Antonio Fernandez Anta, Marios D. Dikaiakos

With the widespread adoption of temporal graphs to study fast evolving interactions in dynamic networks, attention is needed to provide graph metrics in time and at scale. In this paper, we introduce ATMoN, an open-source library developed to computationally offload graph processing engines and ease the communication overhead in dynamic networks over an unprecedented wealth of data. This is achieved, by efficiently adapting, in place and inexpensively, the temporal granularity at which graph metrics are computed based on runtime knowledge captured by a low-cost probabilistic learning model capable of approximating both the metric stream evolution and the volatility of the graph topology. After a thorough evaluation with real-world data from mobile, face-to-face and vehicular networks, results show that ATMoN is able to reduce the compute overhead by at least 76%, data volume by 60% and overall cloud costs by at least 54%, while always maintaining accuracy above 88%.

Mark Leznik, Simon Volpert, Frank Griesinger, Daniel Seybold, Jörg Domaschka

With the rapid rise of the cloud computing paradigm, the manual maintenance and provisioning of the technological layers behind it, both in their hardware and virtualized form, became cumbersome and error-prone. This has opened up the need for automated capacity planning strategies in heterogeneous cloud computing environments. However, even with mechanisms to fully accommodate customers and fulfill service-level agreements, providers often tend to over-provision their hardware and virtual resources. A proliferation of unused capacity leads to higher energy costs, and correspondingly, the price for cloud technology services. Capacity planning algorithms rely on data collected from the utilized resources. Yet, the amount of data aggregated through the monitoring of hardware and virtual instances does not allow for a manual supervision, much less data analysis or a correlation and anomaly detection. Current data science advancements enable the assistance of efficient automation, scheduling and provisioning of cloud computing resources based on supervised and unsupervised machine learning techniques. In this work, we present the current state of the art in monitoring, storage, analysis and adaptation approaches for the data produced by cloud computing environments, to enable proactive, dynamic resource provisioning.