CLOSER 2022 Abstracts

Area 1 - Cloud Computing Fundamentals
Short Papers
Area 2 - Cloud Operations
Full Papers Short Papers
Area 3 - Edge Cloud and Fog Computing
Full Papers Short Papers
Area 4 - Service Modelling and Analytics
Short Papers
Area 5 - Services Science
Full Papers Short Papers
Area 6 - Cloud Computing Platforms and Applications
Full Papers Short Papers
Area 7 - Cloud Computing Enabling Technology
Full Papers Short Papers

Area 1 - Cloud Computing Fundamentals

Short Papers
Paper Nr: 10
Title:

Cloud Function Lifecycle Considerations for Portability in Function as a Service
Authors:

Robin Hartauer, Johannes Manner and Guido Wirtz
Abstract: Portability is an important property to assess the quality of software. In cloud environments, where functions and other services are hosted by providers with proprietary interfaces, vendor lock-in is a typical problem. In this paper, we investigated the portability situation for public Function as a Service (FaaS) offerings based on six metrics. We designed our research to address the software lifecycle of a cloud function during implementation, packaging and deployment. For a small use case, we derived a portability-optimized implementation. Via this empirical investigation and a prototypical migration from AWS Lambda to Azure Function and from AWS Lambda to Google Cloud Function respectively, we were able to reduce writing source code in the latter case by a factor of 17 measured on a Lines of Code (LOC) basis. We found that the default zip packaging option is still the favored approach at Function as a Service (FaaS) platforms. For deploying our functions to the cloud service provider, we used Infrastructure as Code (IaC) tools. For cloud function only deployments the Serverless Framework is the best option whereas Terraform supports developers for mixed deployments where cloud functions and dependent services like databases are deployed at once.
Download
Paper Nr: 41
Title:

Latency Assessment of Blockchain-based SSI Applications Utilizing Hyperledger Indy
Authors:

Hamza Baniata, Tamas Pflanzner, Zoltan Feher and Attila Kertesz
Abstract: Blockchain is the core technology behind several revolutionary applications that require consistent and immutable Distributed Ledgers, maintained by multi-party authorities. Examples of such applications include cryptocurrencies, smart contracts, Self Sovereign Identity (SSI) and Edge/Fog-enabled smart systems (eHealth, IIoT, IoV, etc.). Hyperledger Indy and Aries are suitable open-source tools for permissioned blockchain utilization in SSI projects. Those two frameworks have gained much attraction by researchers interested in the topic, while continuously maintained under the umbrella of the Linux Foundation. However, some SSI applications require specific upper bound of response time depending on their business model. In this paper we aim at presenting a detailed latency analysis of Indy, on top of which Aries is typically built. With such an architecture, researchers and practitioners of SSI applications can decide whether this framework fulfills their application requirements. To realize our proposed architecture, we have developed a Python application with containerized Indy and Aries components. Its scripts use and build on the official open-source codes of the Hyperledger repositories. We have deployed our application on multiple virtual machines in the Google Cloud Platform, and evaluated it with various scenarios. Depending on the transaction rate, we found that the writing response latency of an Indy-based blockchain containing 4 and 8 nodes, ranges between 1–16 seconds, while the reading response latency with similar settings ranges between 0.01–5 seconds.
Download
Paper Nr: 42
Title:

Porting Non Cloud-native Applications across Linux Distributions: A Practical Approach
Authors:

Sanjeet Kumar and Suvrojit Das
Abstract: Researchers are trying to solve the problem of application portability ever since we had multiple types of Unix operating systems. The need for a solution to port application between different operating environments has only grown in recent years after the adoption of the cloud. The consumers of IT applications want the freedom of choice when it comes to operating environment. They don’t want to be put in a vendor lock-in where they are compelled to use one particular platform inspite of their dissatisfaction of service or availability of a better, economic and more competitive alternative. Sometimes the provider of the operating environment goes out of business or there are other reasons where porting of applications becomes a necessity. Linux being an open source operating system gives the liberty to consumers to first extensively test and use the applications on a free distribution and later port it to a distribution with enterprise support. Portability of cloud native apps are more or less solved by use of containerization tools and their ecosystem however the portability of non cloud-native applications is still an open research problem. This paper provides the solution to the problem of porting an application from one distribution of linux to another using an approach which can further be generalized for porting of application between wider range of operating environments.
Download
Paper Nr: 47
Title:

Multi-cloud Load Distribution for Three-tier Applications
Authors:

Adekunbi A. Adewojo and Julian M. Bass
Abstract: Web-based business applications commonly experience user request spikes called flash crowds. Flash crowds in web applications might result in resource failure and/or performance degradation. To alleviate these challenges, this class of applications would benefit from a targeted load balancer and deployment architecture of a multi-cloud environment. We propose a decentralised system that effectively distributes the workload of three-tier web-based business applications using geographical dynamic load balancing to minimise performance degradation and improve response time. Our approach improves a dynamic load distribution algorithm that utilises five carefully selected server metrics to determine the capacity of a server before distributing requests. Our first experiments compared our algorithm with multi-cloud benchmarks. Secondly, we experimentally evaluated our solution on a multi-cloud test-bed that comprises one private cloud, and two public clouds. Our experimental evaluation imitated flash crowds by sending varying requests using a standard exponential benchmark. It simulated resource failure by shutting down virtual machines in some of our chosen data centres. Then, we carefully measured response times of these various scenarios. Our experimental results showed that our solution improved application performance by 6.7% during resource failure periods, 4.08% and 20.05% during flash crowd situations when compared to Admission Control and Request Queuing benchmarks.
Download

Area 2 - Cloud Operations

Full Papers
Paper Nr: 49
Title:

Automated Generation of Application Management Workflows using TOSCA Policies
Authors:

Domenico Calcaterra and Orazio Tomarchio
Abstract: Cloud orchestration frameworks are recognised as a key driver to tackle the complexity of deploying and managing cloud resources and services. Although automated orchestration approaches have been widely adopted in industry and academia, most of them only provide limited management capabilities, such as monitoring or scaling individual components. Holistic management features concerning multiple application components are mostly not covered by existing solutions. Supporting such features typically involves a custom implementation of management logic, which can be error-prone, time-consuming and frequently outdated. In this work, we present an approach for the automated generation of holistic management workflows based on TOSCA application models, reusable node types and policy types. A prototype implementation and a case study are also discussed to prove the feasibility of the proposed idea.
Download
Short Papers
Paper Nr: 32
Title:

Real-time Statistical Log Anomaly Detection with Continuous AIOps Learning
Authors:

Lu An, An-Jie Tu, Xiaotong Liu and Rama Akkiraju
Abstract: Anomaly detection from logs is a fundamental Information Technology Operations (ITOps) management task. It aims to detect anomalous system behaviours and find signals that can provide clues to the reasons and the anatomy of a system’s failure. Applying advanced, explainable Artificial Intelligence (AI) models throughout the entire ITOps is critical to confidently assess, diagnose and resolve such system failures. In this paper, we describe a new online log anomaly detection algorithm which helps significantly reduce the time-to-value of Log Anomaly Detection. This algorithm is able to continuously update the Log Anomaly Detection model at run-time and automatically avoid potential biased model caused by contaminated log data. The methods described here have shown 60% improvement on average F1-scores from experiments for multiple datasets comparing to the existing method in the product pipeline, which demonstrates the efficacy of our proposed methods.
Download

Area 3 - Edge Cloud and Fog Computing

Full Papers
Paper Nr: 27
Title:

Live Migration of Containers in the Edge
Authors:

Rohit Das and Subhajit Sidhanta
Abstract: Being comprised of resource-constrained edge devices, live migration is a necessary feature in edge clusters for migrating the state of an entire machine in case of machine failures and network partitions without disrupting the continued availability of services. While most of the prior work in this area has provided solutions for live migration on clusters comprised of resource-rich servers or fog servers with high computing power, there is a general lack of research on live migration on the low-end ARM devices comprised in edge clusters. To that end, we propose a lightweight algorithm for performing live migration on resource-constrained edge clusters. We provide an open-source implementation of the above algorithm for migrating Linux containers. We demonstrate, using simulations as well as benchmark experiments, that our algorithm outperforms state-of-the-art live migration algorithms on resource-constrained edge clusters with network partitions and device failures.
Download
Short Papers
Paper Nr: 22
Title:

Communication Aware Scheduling of Microservices-based Applications on Kubernetes Clusters
Authors:

Angelo Marchese and Orazio Tomarchio
Abstract: Edge computing paradigm has enabled new application categories with low latency requirements. Container technologies are increasingly spreading to provide flexible and scalable services also within these dynamic environments. However, scheduling distributed microservices applications in the Cloud-to-Edge continuum is a challenging problem, considering the instability and limited network connectivity of Edge infrastructure. Existing container orchestration systems, like Kubernetes, allow to ease the deployment and scheduling of distributed applications in Cloud data centers but their scheduling strategy presents some limitations when dealing with latency critical applications, because it does not consider application communication requirements. In this work we propose an extension of the default Kubernetes scheduler that takes into account microservices communication requirements, modeled through the use of the TOSCA language, traffic history and network latency metrics in order to assign node scores when scheduling each application Pod. A qualitative analysis of the proposed scheduler is presented with a use case.
Download
Paper Nr: 23
Title:

Optimal Placement of Micro-services Chains in a Fog Infrastructure
Authors:

Claudia Canali, Giuseppe Di Modica, Riccardo Lancellotti and Domenico Scotece
Abstract: Fog computing emerged as a novel approach to deliver micro-services that support innovative applications. This paradigm is consistent with the modern approach to application development, that leverages the composition of small micro-services that can be combined to create value-added applications. These applications typically require the access from distributed data sources, such as sensors located in multiple geographic locations or mobile users. In such scenarios, the traditional cloud approach is not suitable because latency constraints may not be compatible with having time-critical computations occurring on a far away data-center; furthermore, the amount of data to exchange may cause high costs imposed by the cloud pricing model. A layer of fog nodes close to application consumers can host pre-processing and data aggregation tasks that can reduce the response time of latency-sensitive elaboration as well as the traffic to the cloud data-centers. However, the problem of smartly placing micro-services over fog nodes that can fulfill Service Level Agreements is far more complex than in the more controlled scenario of cloud computing, due to the heterogeneity of fog infrastructures in terms of performance of both the computing nodes and inter-node connectivity. In this paper, we tackle such problem proposing a mathematical model for the performance of complex applications deployed on a fog infrastructure. We adapt the proposed model to be used in a genetic algorithm to achieve optimized deployment decisions about the placement of micro-services chains. Our experiments prove the viability of our proposal with respect to meeting the SLA requirements in a wide set of operating conditions.
Download
Paper Nr: 43
Title:

Multivariate Interpolation at the Edge to Infer Faulty IoT Sensor Metrics
Authors:

Marcos P. Konzen, Patric R. Izolan, Fábio J. Griesang, Paulo S. De Souza, Tiago C. Ferreto, Arthur F. Lorenzon, Marcelo C. Luizelli, Julio C. Balzano De Mattos, Cinara Ewerling Da Rosa and Fábio D. Rossi
Abstract: Virtual sensors are software entities that allow the estimation, through models, of critical variables in a given environment. Metrics can be modeled computationally to estimate the values measured by a sensor without installing it physically in the specified location. The monitoring and control of its variables by the edge are of great importance, as they are directly related to increased productivity. This article presents the idea behind virtual sensors, discusses some challenges and trends, presents such sensors’ modeling for estimating values, and gives results based on a Smart Farming case study. The results show that the virtual sensors’ estimated values are very close to reality, which shows that our method can be used with very high confidence.
Download
Paper Nr: 16
Title:

From the Sky to the Ground: Comparing Fog Computing with Related Distributed Paradigms
Authors:

Joao Bachiega Jr., Breno Costa, Leonardo R. Carvalho, Victor C. Oliveira, William X. Santos, Maria C. S. de Castro and Aleteia Araujo
Abstract: Fog computing is a paradigm that enables provisioning resources and services at the network edge, closer to end devices, complementing cloud computing and recently it’s being embraced by sky computing. Beyond this computational paradigm, there are many other related technologies which also make use of distributed computing to improve the quality of service delivered to the end-users. The main contribution of this paper is to present a comparison between fog computing and nine other relevant related paradigms, namely Sky Computing, Cloud Computing, Edge Computing, Mobile Edge Computing, Mobile Cloud Computing, Mobile Ad hoc Cloud Computing, Mist Computing, Cloudlet Computing, and Dew Computing, highlighting the similarities and differences between them based on the main fog characteristics. A graphical characterization for each paradigm, highlighting its computational power, communication type and position in a three-layers architecture (Cloud-Fog-IoT) and some relevant challenges in this area are also presented.
Download
Paper Nr: 36
Title:

Mobile Games at the Edge: A Performance Evaluation to Guide Resource Capacity Planning
Authors:

Gabriel Araújo, Carlos Brito, Leonel Correia, Tuan A. Nguyen, Jae W. Lee and Francisco A. Silva
Abstract: Mobile games are very popular among young generations, especially during the worldwide Covid-19 pandemic. The pandemic has caused an enormous increase in data transactions and computation over the Internet. Computing for games often consumes a vast amount of computational resources. Nowadays, mobile devices require heavy computing tasks. For this reason, edge computing resources are essentially needed in the game industry for non-latency data transactions. However, edge computing involves many aspects that make its architecture highly complex to evaluate. Pure performance evaluation of such computing systems is necessary for real-world mobile edge computing systems (MEC) in the game industry. This paper proposes a closed queuing network to evaluate the performance of a game execution scenario in MEC. The model permits the evaluation of the following metrics: mean response time, drop rate, and utilization level. The results indicate that the variation in the number of physical machines (PM) and virtual machines (VM) has a similar impact on the system’s overall performance. The results also show that dropped messages can be avoided by making small calibrations on the capabilities of the VM/PM resources. Finally, this study seeks to assist the development of game computing systems at MEC without the need for prior expenses with real infrastructures.
Download
Paper Nr: 44
Title:

Latency-aware Privacy-preserving Service Migration in Federated Edges
Authors:

Paulo Souza, Ângelo Crestani, Felipe Rubin, Tiago Ferreto and Fábio Rossi
Abstract: Edge computing has been in the spotlight for bringing data processing nearby data sources to support the requirements of latency-sensitive and bandwidth-hungry applications. Previous investigations have explored the coordination between multiple edge infrastructure providers to maximize profit while delivering enhanced applications’ performance and availability. However, existing solutions overlook potential conflicts between data protection policies implemented by infrastructure providers and security requirements of privacy-sensitive applications (e.g., databases and payment gateways). Therefore, this paper presents Argos, a heuristic algorithm that migrates applications according to their latency and privacy requirements in federated edges. Experimental results demonstrate that Argos can reduce latency and privacy issues by 16.98% and 7.95%, respectively, compared to state-of-the-art approaches.
Download
Paper Nr: 50
Title:

A Quality-driven Machine Learning Governance Architecture for Self-adaptive Edge Clouds
Authors:

Claus Pahl, Shelernaz Azimi, Hamid R. Barzegar and Nabil El Ioini
Abstract: Self-adaptive systems such as clouds and edge clouds are more and more using Machine Learning (ML) techniques if sufficient data is available to create respective ML models. Self-adaptive systems are built around a controller that, based on monitored system data as input, generate actions to maintain the system in question within expected quality ranges. Machine learning (ML) can help to create controllers for self-adaptive systems such as edge clouds. However, because ML-created controllers are created without a direct full control by expert software developers, quality needs to be specifically looked at, requiring a better understanding of the ML models. Here, we explore a quality-oriented management and governance architecture for self-adaptive edge controllers. The concrete objective here is the validation of a reference governance architecture for edge cloud systems that facilitates ML controller quality management in a feedback loop.
Download

Area 4 - Service Modelling and Analytics

Short Papers
Paper Nr: 21
Title:

Towards a Performance Model for Byzantine Fault Tolerant Services
Authors:

Thomas Loruenser, Benjamin Rainer and Florian Wohner
Abstract: Byzantine fault-tolerant systems have been researched for more than four decades, and although shown possible early, the solutions were impractical for a long time. With PBFT the first practical solution was proposed in 1999 and spawned new research which culminated in novel applications using it today. Although the safety and liveness properties of PBFT-type protocols have been rigorously analyzed, when it comes to practical performance only empirical results —often in artificial settings— are known and imperfections on the communication channels are not specifically considered. In this work we present the first performance model for PBFT that specifically considers the impact of unreliable channels and the use of different transport protocols over them. We also performed extensive simulations to verify the model and to gain more insight into the impact of deployment parameters on the overall transaction time. We show that the usage of UDP can lead to significant speedups for PBFT protocols compared to TCP when tuned accordingly, even over lossy channels.
Download
Paper Nr: 39
Title:

Total Cost Modeling for VNF based on Licenses and Resources
Authors:

Ghoshana Bista, Eddy Caron and Anne-Lucie Vion
Abstract: Moving to NFV (Network Function Virtualization) and SDN (Software Defined Network), Telco cloud architectures face four key challenges: interoperability, automation, reliability, and adaptability. All these challenges involve the optimization of resources; whether it is to increase the utilization of hardware resources (virtualization) or to deliver shared computing resources and functions in real-time (cloudification). Softwarization of networks is a consequence of telecom cloudification. Virtual Network Function (VNF) is protected by IPR (Intellectual Property Right) like any software, ensured by a license describing usage rights and restrictions at a given cost. Until now limited studies have happened in the economic dimension linked to softwarisation. Currently, the telco industry struggles to converge and standardize licensing and cost models. At risk: the network cloudification benefits could be swept away by poor management of resources (Hardware and Software). This article presents a preliminary model for optimizing the total cost of a VNF, based on the Resource Cost (RC) and License Cost (LC). This analysis is inspired by measurement and licensing practices commonly observed in the Telcos industries,i.e consumption and capacity.
Download

Area 5 - Services Science

Full Papers
Paper Nr: 15
Title:

Semantic Code Clone Detection Method for Distributed Enterprise Systems
Authors:

Jan Svacina, Vincent Bushong, Dipta Das and Tomas Cerny
Abstract: Conventional approaches to code clone detection consider systems from elementary construct perspectives, making it difficult to detect semantic clones. This paper argues that semantic clone detection could be improved for enterprise systems since they typically use well-established architectures and standards. Semantic clone detection is crucial for enterprises where software’s codebase grows and evolves and maintenance costs rise significantly. While researchers have proposed many code clone detection techniques, there is a lack of solutions targeted explicitly toward enterprise systems and even fewer solutions dedicated to semantic clones. Semantic clones exhibit the same behavior between clone pairs but differ in the syntactic structure. This paper proposes a novel approach to detect semantic clones for enterprise frameworks. The driving idea is to transform a particular enterprise application into a control-flow graph representation. Next, various proprietary similarity functions are applied to compare targeted enterprise metadata for each pair of the control-flow graph fragment. As a result, we achieve to detect semantic clones with high accuracy and reasonable time complexity.
Download
Paper Nr: 18
Title:

Analysis and Rewrite of Quantum Workflows: Improving the Execution of Hybrid Quantum Algorithms
Authors:

Benjamin Weder, Johanna Barzen, Martin Beisel and Frank Leymann
Abstract: With the rapid evolution of quantum computers and the emergence of different quantum cloud offerings, use cases from various application areas, such as chemistry or physics, can now be implemented and executed on real quantum computers. Thereby, the applications are typically hybrid, i.e., combine quantum and classical programs. Workflows enable the orchestration of these programs and provide advantages, such as robustness or reproducibility. However, different quantum algorithms require executing the quantum and classical programs in a loop with many iterations, leading to an inefficient orchestration through the workflow. For the efficient execution of such algorithms, hybrid runtimes are offered, combining the quantum and classical programs in a single hybrid program, optimizing the execution. However, this leads to a conceptional gap between the modeling benefits of workflow technologies, e.g., modularization, reuse, and understandability, and the efficiency improvements when using hybrid runtimes. To overcome this issue, we present a method to model all tasks explicitly in the workflow model and analyze the workflow to detect loops that can benefit from hybrid runtimes. Furthermore, corresponding hybrid programs are automatically generated, and the workflow is rewritten to use them. We validate the practical feasibility of our approach by a prototypical implementation.
Download
Short Papers
Paper Nr: 11
Title:

Saga Pattern Technologies: A Criteria-based Evaluation
Authors:

Karolin Dürr, Robin Lichtenthäler and Guido Wirtz
Abstract: One challenge in Microservices Architectures is coordinating business workflows between services, for which the Saga pattern is frequently mentioned as a solution in the literature. This work presents a criteria catalog based on which existing technological solutions that help with Saga implementations can be qualitatively evaluated to enable an informed decision between them. It considers criteria relevant for the Saga pattern, microservices characteristics, and for operating a system in production. We use it to evaluate four technological solutions by implementing an exemplary use case. Due to their different origins, the technologies come with varying strengths and weaknesses and as a result no solution is superior. The results can help developers decide which technology to use and provide insights into what to consider when implementing the Saga pattern.
Download
Paper Nr: 25
Title:

DISDA: Digital Service Design Architecture for Smart City Ecosystems
Authors:

Mouzhi Ge and Barbora Buhnova
Abstract: Due to the interoperability difficulty and development bottleneck of various services in a city, the effective design of services has become a critical concern in smart city ecosystems. Based on the interconnection concept of Management by Competencies, this paper proposes an DISDA architecture to facilitate the digital service design in a smart city ecosystem. The DISDA architecture not only can guide the users to design the services with the defined processes but also can measure the maturity of the existing services and determine possible service improvements. Based on the proposed service design architecture, we conduct a case study to validate the usability and applicability of the proposed service design architecture.
Download
Paper Nr: 35
Title:

A Research-backed Extended Taxonomy for Cloud Computing Elasticity
Authors:

Raouia Bouabdallah
Abstract: Elasticity is an important feature to characterize cloud computing from traditional Information Technology (IT) infrastructure. It refers to the ability of the cloud provider to provision and release cloud resources, with demand, appearing to be infinite in any quantity at any time.In this paper, we propose a taxonomy which is an extended elasticity classifications compared to existing ones. Then, we discuss industrial and academic research on cloud elasticity to identify the main issues and drawbacks. Finally, we propose a synthesis of the studied works based on elasticity solutions’ characteristics provided from our taxonomy.
Download

Area 6 - Cloud Computing Platforms and Applications

Full Papers
Paper Nr: 12
Title:

Performance of Databases Used in Data Stream Processing Environments
Authors:

Manuel Weißbach and Thomas Springer
Abstract: Data stream processing (DSP) is being used in more and more fields to process large amounts of data with minimal latency and high throughput. In typical setups, a stream processing engine is combined with additional components, especially database systems to implement complex use cases, which might cause a significant decrease of processing performance. In this paper we examine the specific data access patterns caused by data stream processing and benchmark database systems with typical use cases derived from a real-world application. Our tests involve popular databases in combination with Apache Flink to identify the system combinations with the highest processing performance. Our results show that the choice of a database is highly dependent on the data access pattern of the particular use case. In one of our benchmarks, we found a throughput difference of a factor of 46.2 between the best and the worst performing database. From our experience in implementing a complex real-world application, we have derived a set of performance optimization recommendations to help system developers to select an appropriate database for their use case and to find a high-performing system configuration.
Download
Paper Nr: 28
Title:

Offline Mining of Microservice-based Architectures
Authors:

Jacopo Soldani, Javad Khalili and Antonio Brogi
Abstract: Designing, implementing, and operating microservices is known to be complex and costly, mainly due to the multitude of heterogeneous software services forming a microservice-based application. Such tasks can be simpler if a specification of the microservice-based architecture (MSA) of an application is available. At the same time, due to the number of services and service interactions in a MSA, manually generating a specification of such MSA is complex and costly. For this reason, in this paper we present a novel technique for automatically mining the specification of a MSA from its Kubernetes deployment. The obtained MSA specification is in µTOSCA, a microservice-oriented profile of the human- and machine-readable OASIS standard TOSCA. We also present a prototype implementation of our technique, which we use to assess it by means of case studies based on third-party applications.
Download
Paper Nr: 46
Title:

A Novel Weight-assignment Load Balancing Algorithm for Cloud Applications
Authors:

Adekunbi A. Adewojo and Julian M. Bass
Abstract: Load balancing dynamically optimizes cloud resources and performance, and enhances the performance of applications deployed on cloud. We have chosen to investigate the class of cloud deployed web-based three-tier business applications. There is a problem with load balancing for this class of applications when they suffer from overload due to sudden flash crowds and resource failures. We propose a novel weight assignment load balancing algorithm to address this problem. Our approach utilises five carefully selected server metrics to efficiently distribute load among virtual machines. First, we validated our novel algorithm by comparing it with a baseline load-balancing algorithm and round-robin algorithm. Then, we experimentally evaluated our solution, by varying the number of user requests and carefully measuring response times and throughput. The experiments were performed on a private cloud environment testbed running OpenStack. Our experimental results show that our approach improves the response time of user requests by 5.66% compared to the baseline algorithm and 15.15% compared to round-robin algorithm in flash crowd scenario. In addition, while handling between 110% to 190% overload, our approach improved response times in all scenarios. Consequently, our novel algorithm outperforms the baseline and round-robin algorithms in overload conditions.
Download
Short Papers
Paper Nr: 20
Title:

Secure Scheduling of Scientific Workflows in Cloud
Authors:

Shubhro Roy, Arun Ramamurthy, Anand Pawar, Mangesh Gharote and Sachin Lodha
Abstract: Scheduling of tasks in scientific workflow has been challenging due to heterogeneous and interdependent tasks in workflow. The scheduling involves selection of different types of virtual machines (VM) belonging to different instance series (computing, memory, storage) to minimize the overall execution cost and time (makespan). Apart from VM selection, selection of security services (such as authentication, integrity, confidentiality) is critical. In this paper, we propose OptReUse - a workflow schedule generation algorithm for efficient reuse of VMs. Our approach of OptReUse algorithm along with combinatorial optimization approach gives lower cost of scheduling compared to the prior art without incurring delay in the makespan. Further, we enhance the security model by accurate estimation of risks. Our experiments using standard scientific workflows demonstrate that the proposed method gives lower costs compared to the prior VM resource utilization methods.
Download
Paper Nr: 40
Title:

Scalable Infrastructure for Workload Characterization of Cluster Traces
Authors:

Thomas van Loo, Anshul Jindal, Shajulin Benedict, Mohak Chadha and Michael Gerndt
Abstract: In the recent past, characterizing workloads has been attempted to gain a foothold in the emerging serverless cloud market, especially in the large production cloud clusters of Google, AWS, and so forth. While analyzing and characterizing real workloads from a large production cloud cluster benefits cloud providers, researchers, and daily users, analyzing the workload traces of these clusters has been an arduous task due to the heterogeneous nature of data. This article proposes a scalable infrastructure based on Google’s dataproc for analyzing the workload traces of cloud environments. We evaluated the functioning of the proposed infrastructure using the workload traces of Google cloud cluster-usage-traces-v3. We perform the workload characterization on this dataset, focusing on the heterogeneity of the workload, the variations in job durations, aspects of resources consumption, and the overall availability of resources provided by the cluster. The findings reported in the paper will be beneficial for cloud infrastructure providers and users while managing the cloud computing resources, especially serverless platforms.
Download
Paper Nr: 51
Title:

Orama: A Benchmark Framework for Function-as-a-Service
Authors:

Leonardo D. Carvalho and Aleteia F. Araujo
Abstract: The prominent Function-as-a-Service (FaaS) cloud service model has positioned itself as an alternative for solving several problems, and, interest in cloud-oriented architectural solutions that use FaaS has therefore grown rapidly. Consequently, the importance of knowing the behavior of FaaS-based architectures under different concurrency scenarios has also become significant, especially in implementation decision-making processes. In this work, the Orama framework is proposed, which helps in the execution of benchmarks in FaaS-based environments, orchestrating the deployment of pre-built architectures, as well as the execution of tests and statistical analysis. Experiments were carried out with architectures containing multiple cloud services in conjunction with FaaS in two public cloud providers (AWS and GCP). The results were analyzed using factorial design and t-test and showed that the use cases running on AWS obtained better results in runtime compared to their counterparts on GCP, but showed considerable error rates in competition situations. It is worth mentioning that the Orama framework was used from in the automated provisioning of use cases, execution of benchmarks, analysis of results and deprovisioning of the environment, supporting the entire process.
Download
Paper Nr: 29
Title:

Planning for Software System Recovery by Knowing Design Limitations of Cloud-native Patterns
Authors:

Alireza Hakamian, Floriment Klinaku, Sebastian Frank, André van Hoorn and Steffen Becker
Abstract: Context. Application designers use cloud-native architectural patterns such as Circuit Breaker that come with third-party implementations to improve overall system reliability. Problem. Important quality decisions are hidden in the codebase and are usually not documented by third-party implementations. Runtime changes may invalidate, e.g., pattern’s decision assumption(s) and cause the reliant service to face unacceptable quality degradation with no recovery plan. Objective. The primary goal of this study is to derive important quality decisions of patterns independent of a particular implementation. Method. To achieve our objective, we perform exploratory research on two architectural patterns, (1) Circuit Breaker and (2) Event Sourcing, which come with different third-party implementations and that application designers often use. We formally specify the design and the guarantees of each pattern using Temporal Logic of Actions (TLA) and verify the guarantees, which guide us in deriving important quality decisions. Result. To show the usefulness of our method, we systematically generate failure scenarios for third-party implementations of Circuit Breaker and Event Sourcing patterns that compromise Hystrix’ and Kafka’s guarantees on preventing further degradation of protected services and the loss of committed messages, respectively. Conclusion. The result suggests that important quality decisions derived from formal models of the patterns help application designers prepare for unacceptable system quality degradation by knowing when a third-party implementation of the architectural patterns fails to maintain its guarantees.
Download

Area 7 - Cloud Computing Enabling Technology

Full Papers
Paper Nr: 45
Title:

Predicting and Avoiding SLA Violations of Containerized Applications using Machine Learning and Elasticity
Authors:

Paulo Souza, Miguel Neves, Carlos Kayser, Felipe Rubin, Conrado Boeira, João Moreira, Bernardo Bordin and Tiago Ferreto
Abstract: Container-based virtualization represents a low-overhead and easy-to-manage alternative to virtual machines. On the other hand, containers are more prone to performance interference and unpredictability. Consequently, there is growing interest in predicting and avoiding performance issues in containerized environments. Existing solutions tackle this challenge through proactive elasticity mechanisms based on workload variation predictions. Although this approach may yield satisfactory results in some scenarios, external factors such as resource contention can cause performance losses regardless of workload variations. This paper presents Flavor, a machine-learning-based system for predicting and avoiding performance issues in containerized applications. Rather than relying on workload variation prediction as existing approaches, Flavor predicts application-level metrics (e.g., query latency and throughput) through a deep neural network implemented using Tensorflow and scales applications accordingly. We evaluate Flavor by comparing it against a state-of-the-art resource scaling approach that relies solely on workload prediction. Our results show that Flavor can predict performance deviations effectively while assisting operators to wisely scale their services by increasing/decreasing the number of application containers to avoid performance issues and resource underutilization.
Download
Paper Nr: 52
Title:

Comparison of FaaS Platform Performance in Private Clouds
Authors:

Marcelo C. Motta, Leonardo Reboucas De Carvalho, Michel F. Rosa and Aleteia P. Favacho De Araujo
Abstract: Since its appearance in mid-2014, there has been notable growth in the adoption of cloud services via the Function-as-a-Service (FaaS) model, with several public cloud providers offering this model in their catalog. Various papers have mostly evaluated the performance of these models in public cloud environments. However, the implementation of this model in a private cloud environment has not been explored enough by the academic community. This article presents a two-level factorial design-based assessment of the main open-source operating platforms (Such as OpenWhisk, Fission, and OpenFaaS) currently available to serve as a comparative tool that can be used in decision-making processes. Results showed that regarding the stress test the OpenWhisk platform has greater reliability. On the other hand, regarding the processing of the Matrix, Factors, and Filesystem functions, Fission remains at similar behavior at all concurrency levels.
Download
Short Papers
Paper Nr: 8
Title:

Evaluation of Language Runtimes in Open-source Serverless Platforms
Authors:

Karim Djemame, Daniel Datsev and Vasilios Kelefouras
Abstract: Serverless computing is revolutionising cloud application development as it offers the ability to create modular, highly-scalable, fault-tolerant applications, with minimal operational management. In order to contribute to its widespread adoption of serverless platforms, the design and performance of language runtimes that are available in Function-as-a-Service (FaaS) serverless platforms is key. This paper aims to investigate the performance impact of language runtimes in open-source serverless platforms, deployable on local clusters. A suite of experiments is developed and deployed on two selected platforms: OpenWhisk and Fission. The results show a clear distinction between compiled and dynamic languages in cold starts but a pretty close overall performance in warm starts. Comparisons with similar evaluations for commercial platforms reveal that warm start performance is competitive for certain languages, while cold starts are lagging behind by a wide margin. Overall, the evaluation yielded usable results in regards to preferable choice of language runtime for each platform.
Download
Paper Nr: 14
Title:

Reconstructing the Holistic Architecture of Microservice Systems using Static Analysis
Authors:

Vincent Bushong, Dipta Das and Tomas Cerny
Abstract: Cloud-native systems, fueled by microservice architecture, provide several benefits to enterprises. These benefits include scalability, short deployment cycles, and flexibility for evolution. Most benefits come from microservices’ independence and decentralization. However, the pay-off comes as a lack of a centralized view of the overall system’s architecture. The system’s data model is separated among and partially replicated between each microservice, requiring extra effort to create a single view on the context map. Additionally, while a microservice’s API and its interaction can be statically documented (i.e., communication diagram), system evolution makes it difficult to maintain. As a result, modifications to the system can decay from the original intended design, and the changes will be obscured by the lack of an up-to-date centralized view. To address this, we propose a method of software architecture reconstruction based on static code analysis of the microservice mesh, generating a communication diagram, context map, and microservice-specific bounded contexts. This gives developers and Development and Operations engineers (DevOps) a centralized view of how the overall program works, useful for furthering system comprehension and observation.
Download