Software Aging and Rejuvenation is a research topic that is being observed increasingly in the mo... more Software Aging and Rejuvenation is a research topic that is being observed increasingly in the most diverse computing scenarios since the demand for reliability and availability of services has increased significantly. With the increasing use of containerized systems, there is also a need for the support of an orchestrator to ease management and reduce operational costs. In this paper, the behavior of Kubernetes software is evaluated, using both Minikube and K3S tools, while they manage pods in an accelerated lifetime experiment. Operations for creating and terminating pods are executed continuously, thus enabling us to observe the utilization of computing resources (e.g., CPU, memory, and I/O), the performance of the Minikube and K3S environments, and the response time of an application hosted in those environments. Software aging effects were observed in some conditions and for specific metrics, such as virtual memory usage, indicating a memory leak that is not fully cleared when the cluster is stopped. The issues found in this work may lead to system performance degradation and eventually compromise reliability and availability when the system crashes due to memory space exhaustion as well as full utilization of swap space on the hard disk.
Distributed computing is bringing many advantages in cost, flexibility and availability. However,... more Distributed computing is bringing many advantages in cost, flexibility and availability. However, it increases the demand for performance and reliability. Resources such as CPU, memory, storage and network bandwidth, are very susceptible of presenting software aging issues. Therefore, proactive actions, also known as software rejuvenation must be performed to avoid these issues. The identification of the best moment to perform software rejuvenation is not a simple task, mostly because it may affect the system's availability and reliability. To overcome this problem, we propose an automatic forecasting strategy to support the system administrators to choose the best moment to perform software rejuvenation. Our strategy uses six time series techniques: Drift, Simple Exponential Smoothing, Holt, Holt-Winters, Linear Regression, and ARIMA. In our proposal, the most suitable one is chosen automatically as the best fit for a particular scenario. Three case studies were performed to evaluate the efficiency of our automatic strategy. Our proposal aims to increase the system's availability while decreasing the QoS violation probability. In one of our experiments, we can observe a reduction of 92.3% in the system's downtime. This research supports decision making activities and opens possibilities to foster the usage of forecasting strategies when dealing with software aging phenomenon.
The blockchain shared ledger emerged as an alternative to the bureaucratic banking system that ma... more The blockchain shared ledger emerged as an alternative to the bureaucratic banking system that may take days to confirm a payment or a transfer between clients. The blockchain concept evolved and became viable for various applications beyond the domain of financial transactions. Blockchains become a way to reach better relationships through contract validation, documents transfer, and personal and business data security. Recently, the blockchain-as-a-service has debuted on Microsoft Data Centers, and now many share an infrastructure that can change and improve their security routines. This paper evaluates the feasibility of a blockchain-as-a-service infrastructure and helps those who plan to deploy or sell blockchains. A modeling methodology based on Dynamical Reliability Block Diagrams (DRBD) is adopted to evaluate two dependability attributes: system’s reliability and availability. The proposed infrastructure contains the minimum requirements to deploy the Hyperledger Cello, a platform to create and manage blockchains. The availability results pointed out a system downtime of 121 hours per year and reliability issues that must be addressed when building blockchain-as-a-service infrastructures.
An automated hydraulic system can contribute to water and financial economics, bringing ecologica... more An automated hydraulic system can contribute to water and financial economics, bringing ecological benefits to the environment. This article presents a proposal for a system for capturing and reusing water from refrigeration appliances. An integration between hydraulics and sensors linked to an Arduino board or microcontroller is demonstrated here. The proposed system monitors the water consumption and manages in which moments the reused water or the sanitation company will be used. Measurements of the volume of water coming from refrigeration appliances, small scale prototypes and three-dimensional models were used as proof of concept, demonstrating the viability of the solution.
2018 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2018
The adoption of stochastic models has been one of the central topics in various architectures. On... more The adoption of stochastic models has been one of the central topics in various architectures. One important step to adopt it is model validation, which aims at obtaining reasonable models to represent actual behavior of services components, it has been essential to validate models against actual measurements. System-wide model simulation results can be compared with recordings from the measurement. In this paper, we accomplish the model validation to Stochastic Petri Net (SPN) models created to evaluate VoD system hosted on a private cloud system, considering MP4, MPG, Ogg and FLV formats. We proposed the performance model to represent packet transfers, and to compute performance metrics, such as throughput, packet loss, and service delivery reliability. The SPN model enables a compact representation of a large number of packets generated by video streaming. We validate the models through experimental data, using a VoD streaming service in a cloud infrastructure testbed. We demonstrate that the proposed models are accurate and can be utilized for planning the quality of service (QoS) of corporate video streaming infrastructures. A case study is presented to compare the behavior of the system. Results indicate that the model validation way adopted can be a good solution for models validation. A case study is conducted to compare the behavior of the system under distinct network scenarios. The results indicate that video streaming QoS under 3G (EVDO) networks is significantly worse than other wireless technologies, such as WiFi, 3.5G (HSPA+), and 4G (LTE).
Proceedings of the 11th EAI International Conference on Performance Evaluation Methodologies and Tools, 2017
Formal models are widely used in performance and dependability studies of computational systems. ... more Formal models are widely used in performance and dependability studies of computational systems. Graphical modeling tools allow users to compose such models with ease, but they complicate the creation of models with a dynamic/complex structure, the hierarchical arrangement of different models, and the automatic execution of models with different parameter configurations. To overcome this problem, we created a scripting language for the Mercury tool that supports the combination of different modeling approaches (e.g., Stochastic Petri Nets and Reliability Block Diagrams) in a single project. In this paper, we focus on the extensions developed to improve the capabilities of Generalized Stochastic Petri net Modeling: substitution transitions, phase-type delays for timed transitions, support for nets with a variable structure, and event-based programming for simulation.
2018 13th Iberian Conference on Information Systems and Technologies (CISTI), 2018
The general approach of the work described is to develop a Systematic Review to assist health pro... more The general approach of the work described is to develop a Systematic Review to assist health professionals, specifically Physiotherapists, collecting data about interactive environments used to assist in the treatment of injuries from stroke. Although the area is well studied, it is necessary to take into account the lack of studies that mix the development of the environments, as well as the interactivity with the patient. In this work a Systematic Review was conducted in which 13 scientific papers published on virtual and interactive environments were studied until June 2017. The main essence of this study is to provide technologists, physicians and physiotherapists a better and continuous approach to offer the patient, ways in which one does not have to sacrifice oneelf or consider such monotonous procedures.
2019 15th International Conference on Network and Service Management (CNSM), 2019
There is an increasing number of Internet of Things (IoT) devices in the border of computer netwo... more There is an increasing number of Internet of Things (IoT) devices in the border of computer networks, requiring local processing and lightweight virtualization to deal with issues such as heterogeneity, Quality of Service (QoS) management, scalability, mobility, federation, and interoperability. Fog computing can provide the computational resources required by IoT devices to process their data. Low energy consumption and total cost of ownership are among the desirable properties for auxiliar infrastructures such as those deployed for fog computing, which do not require large computational power though. There is a noteworthy trend of undergoing research efforts towards the definition of software and hardware architectures for fog computing in this context. In this sense, this paper presents a Systematic Literature Mapping with the purpose of understanding and identifying metrics and gaps in current literature about orchestration of container-based applications, especially those hosted in clusters of Single Board Computer (SBC) platforms, such as Raspberry Pi, which have been used for deploying fog computing environments.
2021 16th Iberian Conference on Information Systems and Technologies (CISTI), 2021
Lightweight virtualization, and specifically containers, has become widespread in the information... more Lightweight virtualization, and specifically containers, has become widespread in the information technology industry to provide an efficient operational environment for the execution of scalable services on the Internet. Containers rely on a set of technologies different from the features that enable hardware virtualization (i.e., hypervisor-based virtual machines). However, both types of virtualized environments are employed to host applications that will be accessible anytime, anywhere. Thus, they are prone to software aging, which usually affects systems that run for long time intervals. Researchers have identified software aging effects in distinct types of cloud computing environments and hypervisors over recent years. However, a few works have dealt with this phenomenon in container-based platforms. This paper presents an experimental analysis of the software aging effects observed on Docker platforms while also evaluating a time- series model's fitness to predict resource consumption’s progress caused by software aging. We employ a stress test workload tailored for the scenario of containers arranged in a cluster managed by Docker Swarm. The obtained results indicate an increasing usage of resident memory, virtual memory, and CPU usage, as the system undergoes subsequent scale-up and scale-down operations. The quadratic trend model was the best fitting approach for predicting resident and virtual memory usage, with less than 5% of prediction error. The experimental approach presented here may help systems administrators detect evidence of software aging in container- based environments, allowing them to choose a proper method and time for deploying rejuvenation actions to mitigate the dependability issues raised in similar scenarios described here.
2017 IEEE 22nd Pacific Rim International Symposium on Dependable Computing (PRDC), 2017
The evaluation of dependability or performance of general systems usually relies on the assistanc... more The evaluation of dependability or performance of general systems usually relies on the assistance of stochastic modeling and simulation tools. Those software packages enables the creation of models and computation of metrics quickly and accurately. This paper introduces the Mercury tool, which is an integrated software that enables creating and evaluating Reliability Block Diagrams, Stochastic Petri Nets, Continuous Time Markov Chains, and Energy Flow Models. Mercury provides a graphical user interface, a script language for command-line interface, and also an API (Application Programming Interface) that enables interaction through external applications. The evaluation of models is not restricted to the assumption of Exponential distributions, which is a common constraint in other similar tools. Mercury implements a simulation framework that allows more than 25 probability distributions, as well as a moment matching method that enables expolynomial —phase-type —distributions for models solved through numerical analysis. This paper presents the main features and methods available in Mercury to aid the dependability and performance evaluation of various systems, for both academy and industry. The accuracy and applicability of the tool is illustrated by a case study of packet loss and throughput for a Video on Demand (VoD) system.
International Journal of Grid and Utility Computing, 2020
The Data Centre concept has evolved, mainly due to the need to reduce expenses with the required ... more The Data Centre concept has evolved, mainly due to the need to reduce expenses with the required physical space to store, provide and maintain large computational infrastructures. The Software-Defined Data Centre (SDDC) is a result of this evolution. Through SDDC, any service can be hosted by virtualising more reliable and easier-to-keep hardware resources. Nowadays, many services and resources can be provided in a single rack, or even a single machine, with similar availability, considering the deployment costs of silo-based environments. One of the ways to apply the SDDC into a data centre is through hyper-convergence. Among the main contributions of this paper are the behavioural models developed for availability and capacity-oriented availability evaluation of silo-based, converged and hyper-converged cloud computing infrastructures. The obtained results may help stakeholders to select between converged and hyper-converged environments, due to their similar availability but with the particularity of having lower deployment costs.
Mobile computing devices have improved substantially. Such a development fosters the usage of tho... more Mobile computing devices have improved substantially. Such a development fosters the usage of those devices for various purposes, including health care as a prominent example. Mobile health (mHealth) is a subfield of electronic health (eHealth) that refers to the practice of medicine and public health supported by mobile devices. A specific concern for mHealth is the limitation of power supply for the mobile devices, since interruptions on power supply can result in critical situations. This paper proposes analytical models to identify the mean time of one life cycle of the battery, and thus identify the average battery lifetime considering some device utilization characteristics. We evaluate the impact of both network connectivity (local and wide areas) and communication protocols (Polling, Long polling, WebSockets, and XMPP) on the discharging rate, and subsequently on the lifetime of batteries with different power capacities. We observed a noteworthy impact of these factors on battery usage and, thus, on mHealth applications. We use a hierarchical modeling approach that enables efficient representation of system complexity and provides accurate results for planning the autonomy of mobile devices in an mHealth context.
2016 46th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshop (DSN-W), 2016
Cloud computing environments provide storage capacity, processing power, and other computational ... more Cloud computing environments provide storage capacity, processing power, and other computational resources in a flexible way, enabling fast adaptation to highly dynamic workloads. Multimedia services, such as video streaming, are examples of applications that can use cloud computing to leverage their provisioning capacity. This way, it is possible to offer a large variety of multimedia content in many formats, so the users will be able to watch videos as they wish, with a proper resolution and quality, according to his preferences and connection speed. Private infrastructures for Video on Demand (VoD) and live video streaming are especially useful for e-learning on large corporations, universities, and governments. Analytical models are effective tools to evaluate the availability of software, hardware, and other computational resources. In this paper, we study a VoD service hosted in a private cloud computing environment. We present availability models considering the VoD streaming server components that are necessary for viewers access. Hierarchical modeling techniques are used to deal with the complexity of representing such system. Sensitivity analysis is used to determine the parameters that cause greatest impact on the availability, identifying which components require attention when attempting to achieve increased availability in a system. The proposed models are useful for planning private cloud infrastructures for VoD services.
Proceedings of the 30th Annual ACM Symposium on Applied Computing, 2015
Resource scarcity is a major obstacle for many mobile applications, since devices have limited en... more Resource scarcity is a major obstacle for many mobile applications, since devices have limited energy power and processing potential. As an example, there are applications that seamlessly augment human cognition and typically require resources that far outstrip mobile hardware's capabilities, such as language translation, speech recognition, and face recognition. The use of cloud computing may tackle this problem. This study presents SmartRank, a scheduling framework to perform load partitioning and offloading for mobile applications using cloud computing to increase performance in terms of response time. We first explore a benchmarking of face recognition application using mobile cloud and confirms its suitability to be used as case study with SmartRank. We have applied the approach to a face recognition process based on two strategies: cloudlet federation and resource ranking through balanced metrics (level of CPU utilization and round-trip time). Second, using a full factorial experimental design we tuned the SmartRank with the most suitable partitioning decision calibrating scheduling parameters. Nevertheless, SmartRank uses an equation that is extensible to include new parameters and make it applicable to other scenarios.
2018 IEEE 32nd International Conference on Advanced Information Networking and Applications (AINA), 2018
The Information Technology (IT) systems of most organizations must support their operations 24 ho... more The Information Technology (IT) systems of most organizations must support their operations 24 hours a day, 7 days a week. Systems unavailability may have serious consequences such as data loss, customer dissatisfaction, and subsequent revenue loss. With the popularity of cloud computing, the adoption of cloud-based disaster recovery (DR) solutions has gained more space to prevent data loss and ensure business continuity. However, disaster recovery solutions are not cheap and do not exist as a single solution that suits every requirement (e.g., availability and costs). In this paper, we present an integrated model-experiment approach to evaluate cloud-based disaster recovery solutions. We use Stochastic Petri Nets (SPNs) and fault-injection experiments to evaluate availability related metrics like steady-state availability and downtime. To demonstrate the feasibility of our approach, distinct real-world cloud-based DR solutions (e.g., active/active and active/standby) are modeled and analyzed. The results revealed that disaster recovery solution significantly improves system availability and minimizes the downtime costs. In addition, our numerical analysis shows the statistical correspondence between the results of the experiments and models.
2021 Annual Reliability and Maintainability Symposium (RAMS)
SUMMARY & CONCLUSIONSSoftware aging is a topic in the literature that has matured over the pa... more SUMMARY & CONCLUSIONSSoftware aging is a topic in the literature that has matured over the past 30 years and can be defined as software degradation during its operational life. On the other hand, the software rejuvenation technique is a proactive measure that includes the application stop, removal of error accumulation conditions, and restart in a clean environment. This paper provides a systematic mapping of the contemporary literature on Software Aging and Rejuvenation (SAR), observing the research trends in this theme in the last decade, from 2010 to 2019. The results highlight which strategic approaches of rejuvenation or aging detection have been approached when designing solutions both in the commercial and academic fields, and thus, researchers can use them when dealing with new software aging phenomena. Besides, it can be of great relevance for the academic community that studies the SAR to provide a current view on state of the art in this topic.
International Journal of Web and Grid Services, 2016
The task of designing and tuning computer systems for optimal performance and dependability is co... more The task of designing and tuning computer systems for optimal performance and dependability is complex. This usually involves the choice of manufacturers, service providers, types of components, and possible architectures. Analytical modelling is helpful for planning and predicting the quality-of-service QoS measures of a system, and can be especially useful when integrated with optimisation algorithms. This study proposes a sensitive greedy randomised adaptive search procedure Sensitive GRASP, which uses sensitivity analysis of an analytical model to improve the solutions of a GRASP-based optimisation. Our approach is applied for optimising the reliability and performance of a composite web service. For scenarios with thousands of possible combinations, the proposed approach finds configurations that are closer to the optimal solution than those found by the algorithm without sensitivity analysis with equivalent computational effort.
ABSTRACT High availability in cloud computing services is essential for maintaining customer conf... more ABSTRACT High availability in cloud computing services is essential for maintaining customer confidence and avoiding revenue losses due to SLA violation penalties. Since the software and hardware components of cloud infrastructures may have limited reliability, the use of redundant components and multiple clusters may be required to achieve the expected level of dependability while also increasing the computational capacity. A drawback of such improvements is the respective impact on the capital and increase in acquisition and operational costs. This paper presents availability models for private cloud architectures based on Eucalyptus platform, and presents a comparison of costs between these architectures and similar infrastructure rented from a public cloud provider. Metrics for capacity-oriented availability and system steady-state availability are used to compare architectures with distinct numbers of clusters. A heterogeneous hierarchical modeling approach is employed to represent the systems considering both hardware and software failures. The results highlight that improvements on the availability are not significant when increasing the system to more than two clusters. The analysis also shows that the average available capacity is close to the maximum possible capacity in all architectures, and that it takes 18 months, in average, for these private cloud architectures to pay off the cost equivalent to the computational capacity rented from a public cloud. J. Dantas (B) · R. Matos · J. Araujo · P.
Software Aging and Rejuvenation is a research topic that is being observed increasingly in the mo... more Software Aging and Rejuvenation is a research topic that is being observed increasingly in the most diverse computing scenarios since the demand for reliability and availability of services has increased significantly. With the increasing use of containerized systems, there is also a need for the support of an orchestrator to ease management and reduce operational costs. In this paper, the behavior of Kubernetes software is evaluated, using both Minikube and K3S tools, while they manage pods in an accelerated lifetime experiment. Operations for creating and terminating pods are executed continuously, thus enabling us to observe the utilization of computing resources (e.g., CPU, memory, and I/O), the performance of the Minikube and K3S environments, and the response time of an application hosted in those environments. Software aging effects were observed in some conditions and for specific metrics, such as virtual memory usage, indicating a memory leak that is not fully cleared when the cluster is stopped. The issues found in this work may lead to system performance degradation and eventually compromise reliability and availability when the system crashes due to memory space exhaustion as well as full utilization of swap space on the hard disk.
Distributed computing is bringing many advantages in cost, flexibility and availability. However,... more Distributed computing is bringing many advantages in cost, flexibility and availability. However, it increases the demand for performance and reliability. Resources such as CPU, memory, storage and network bandwidth, are very susceptible of presenting software aging issues. Therefore, proactive actions, also known as software rejuvenation must be performed to avoid these issues. The identification of the best moment to perform software rejuvenation is not a simple task, mostly because it may affect the system's availability and reliability. To overcome this problem, we propose an automatic forecasting strategy to support the system administrators to choose the best moment to perform software rejuvenation. Our strategy uses six time series techniques: Drift, Simple Exponential Smoothing, Holt, Holt-Winters, Linear Regression, and ARIMA. In our proposal, the most suitable one is chosen automatically as the best fit for a particular scenario. Three case studies were performed to evaluate the efficiency of our automatic strategy. Our proposal aims to increase the system's availability while decreasing the QoS violation probability. In one of our experiments, we can observe a reduction of 92.3% in the system's downtime. This research supports decision making activities and opens possibilities to foster the usage of forecasting strategies when dealing with software aging phenomenon.
The blockchain shared ledger emerged as an alternative to the bureaucratic banking system that ma... more The blockchain shared ledger emerged as an alternative to the bureaucratic banking system that may take days to confirm a payment or a transfer between clients. The blockchain concept evolved and became viable for various applications beyond the domain of financial transactions. Blockchains become a way to reach better relationships through contract validation, documents transfer, and personal and business data security. Recently, the blockchain-as-a-service has debuted on Microsoft Data Centers, and now many share an infrastructure that can change and improve their security routines. This paper evaluates the feasibility of a blockchain-as-a-service infrastructure and helps those who plan to deploy or sell blockchains. A modeling methodology based on Dynamical Reliability Block Diagrams (DRBD) is adopted to evaluate two dependability attributes: system’s reliability and availability. The proposed infrastructure contains the minimum requirements to deploy the Hyperledger Cello, a platform to create and manage blockchains. The availability results pointed out a system downtime of 121 hours per year and reliability issues that must be addressed when building blockchain-as-a-service infrastructures.
An automated hydraulic system can contribute to water and financial economics, bringing ecologica... more An automated hydraulic system can contribute to water and financial economics, bringing ecological benefits to the environment. This article presents a proposal for a system for capturing and reusing water from refrigeration appliances. An integration between hydraulics and sensors linked to an Arduino board or microcontroller is demonstrated here. The proposed system monitors the water consumption and manages in which moments the reused water or the sanitation company will be used. Measurements of the volume of water coming from refrigeration appliances, small scale prototypes and three-dimensional models were used as proof of concept, demonstrating the viability of the solution.
2018 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2018
The adoption of stochastic models has been one of the central topics in various architectures. On... more The adoption of stochastic models has been one of the central topics in various architectures. One important step to adopt it is model validation, which aims at obtaining reasonable models to represent actual behavior of services components, it has been essential to validate models against actual measurements. System-wide model simulation results can be compared with recordings from the measurement. In this paper, we accomplish the model validation to Stochastic Petri Net (SPN) models created to evaluate VoD system hosted on a private cloud system, considering MP4, MPG, Ogg and FLV formats. We proposed the performance model to represent packet transfers, and to compute performance metrics, such as throughput, packet loss, and service delivery reliability. The SPN model enables a compact representation of a large number of packets generated by video streaming. We validate the models through experimental data, using a VoD streaming service in a cloud infrastructure testbed. We demonstrate that the proposed models are accurate and can be utilized for planning the quality of service (QoS) of corporate video streaming infrastructures. A case study is presented to compare the behavior of the system. Results indicate that the model validation way adopted can be a good solution for models validation. A case study is conducted to compare the behavior of the system under distinct network scenarios. The results indicate that video streaming QoS under 3G (EVDO) networks is significantly worse than other wireless technologies, such as WiFi, 3.5G (HSPA+), and 4G (LTE).
Proceedings of the 11th EAI International Conference on Performance Evaluation Methodologies and Tools, 2017
Formal models are widely used in performance and dependability studies of computational systems. ... more Formal models are widely used in performance and dependability studies of computational systems. Graphical modeling tools allow users to compose such models with ease, but they complicate the creation of models with a dynamic/complex structure, the hierarchical arrangement of different models, and the automatic execution of models with different parameter configurations. To overcome this problem, we created a scripting language for the Mercury tool that supports the combination of different modeling approaches (e.g., Stochastic Petri Nets and Reliability Block Diagrams) in a single project. In this paper, we focus on the extensions developed to improve the capabilities of Generalized Stochastic Petri net Modeling: substitution transitions, phase-type delays for timed transitions, support for nets with a variable structure, and event-based programming for simulation.
2018 13th Iberian Conference on Information Systems and Technologies (CISTI), 2018
The general approach of the work described is to develop a Systematic Review to assist health pro... more The general approach of the work described is to develop a Systematic Review to assist health professionals, specifically Physiotherapists, collecting data about interactive environments used to assist in the treatment of injuries from stroke. Although the area is well studied, it is necessary to take into account the lack of studies that mix the development of the environments, as well as the interactivity with the patient. In this work a Systematic Review was conducted in which 13 scientific papers published on virtual and interactive environments were studied until June 2017. The main essence of this study is to provide technologists, physicians and physiotherapists a better and continuous approach to offer the patient, ways in which one does not have to sacrifice oneelf or consider such monotonous procedures.
2019 15th International Conference on Network and Service Management (CNSM), 2019
There is an increasing number of Internet of Things (IoT) devices in the border of computer netwo... more There is an increasing number of Internet of Things (IoT) devices in the border of computer networks, requiring local processing and lightweight virtualization to deal with issues such as heterogeneity, Quality of Service (QoS) management, scalability, mobility, federation, and interoperability. Fog computing can provide the computational resources required by IoT devices to process their data. Low energy consumption and total cost of ownership are among the desirable properties for auxiliar infrastructures such as those deployed for fog computing, which do not require large computational power though. There is a noteworthy trend of undergoing research efforts towards the definition of software and hardware architectures for fog computing in this context. In this sense, this paper presents a Systematic Literature Mapping with the purpose of understanding and identifying metrics and gaps in current literature about orchestration of container-based applications, especially those hosted in clusters of Single Board Computer (SBC) platforms, such as Raspberry Pi, which have been used for deploying fog computing environments.
2021 16th Iberian Conference on Information Systems and Technologies (CISTI), 2021
Lightweight virtualization, and specifically containers, has become widespread in the information... more Lightweight virtualization, and specifically containers, has become widespread in the information technology industry to provide an efficient operational environment for the execution of scalable services on the Internet. Containers rely on a set of technologies different from the features that enable hardware virtualization (i.e., hypervisor-based virtual machines). However, both types of virtualized environments are employed to host applications that will be accessible anytime, anywhere. Thus, they are prone to software aging, which usually affects systems that run for long time intervals. Researchers have identified software aging effects in distinct types of cloud computing environments and hypervisors over recent years. However, a few works have dealt with this phenomenon in container-based platforms. This paper presents an experimental analysis of the software aging effects observed on Docker platforms while also evaluating a time- series model's fitness to predict resource consumption’s progress caused by software aging. We employ a stress test workload tailored for the scenario of containers arranged in a cluster managed by Docker Swarm. The obtained results indicate an increasing usage of resident memory, virtual memory, and CPU usage, as the system undergoes subsequent scale-up and scale-down operations. The quadratic trend model was the best fitting approach for predicting resident and virtual memory usage, with less than 5% of prediction error. The experimental approach presented here may help systems administrators detect evidence of software aging in container- based environments, allowing them to choose a proper method and time for deploying rejuvenation actions to mitigate the dependability issues raised in similar scenarios described here.
2017 IEEE 22nd Pacific Rim International Symposium on Dependable Computing (PRDC), 2017
The evaluation of dependability or performance of general systems usually relies on the assistanc... more The evaluation of dependability or performance of general systems usually relies on the assistance of stochastic modeling and simulation tools. Those software packages enables the creation of models and computation of metrics quickly and accurately. This paper introduces the Mercury tool, which is an integrated software that enables creating and evaluating Reliability Block Diagrams, Stochastic Petri Nets, Continuous Time Markov Chains, and Energy Flow Models. Mercury provides a graphical user interface, a script language for command-line interface, and also an API (Application Programming Interface) that enables interaction through external applications. The evaluation of models is not restricted to the assumption of Exponential distributions, which is a common constraint in other similar tools. Mercury implements a simulation framework that allows more than 25 probability distributions, as well as a moment matching method that enables expolynomial —phase-type —distributions for models solved through numerical analysis. This paper presents the main features and methods available in Mercury to aid the dependability and performance evaluation of various systems, for both academy and industry. The accuracy and applicability of the tool is illustrated by a case study of packet loss and throughput for a Video on Demand (VoD) system.
International Journal of Grid and Utility Computing, 2020
The Data Centre concept has evolved, mainly due to the need to reduce expenses with the required ... more The Data Centre concept has evolved, mainly due to the need to reduce expenses with the required physical space to store, provide and maintain large computational infrastructures. The Software-Defined Data Centre (SDDC) is a result of this evolution. Through SDDC, any service can be hosted by virtualising more reliable and easier-to-keep hardware resources. Nowadays, many services and resources can be provided in a single rack, or even a single machine, with similar availability, considering the deployment costs of silo-based environments. One of the ways to apply the SDDC into a data centre is through hyper-convergence. Among the main contributions of this paper are the behavioural models developed for availability and capacity-oriented availability evaluation of silo-based, converged and hyper-converged cloud computing infrastructures. The obtained results may help stakeholders to select between converged and hyper-converged environments, due to their similar availability but with the particularity of having lower deployment costs.
Mobile computing devices have improved substantially. Such a development fosters the usage of tho... more Mobile computing devices have improved substantially. Such a development fosters the usage of those devices for various purposes, including health care as a prominent example. Mobile health (mHealth) is a subfield of electronic health (eHealth) that refers to the practice of medicine and public health supported by mobile devices. A specific concern for mHealth is the limitation of power supply for the mobile devices, since interruptions on power supply can result in critical situations. This paper proposes analytical models to identify the mean time of one life cycle of the battery, and thus identify the average battery lifetime considering some device utilization characteristics. We evaluate the impact of both network connectivity (local and wide areas) and communication protocols (Polling, Long polling, WebSockets, and XMPP) on the discharging rate, and subsequently on the lifetime of batteries with different power capacities. We observed a noteworthy impact of these factors on battery usage and, thus, on mHealth applications. We use a hierarchical modeling approach that enables efficient representation of system complexity and provides accurate results for planning the autonomy of mobile devices in an mHealth context.
2016 46th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshop (DSN-W), 2016
Cloud computing environments provide storage capacity, processing power, and other computational ... more Cloud computing environments provide storage capacity, processing power, and other computational resources in a flexible way, enabling fast adaptation to highly dynamic workloads. Multimedia services, such as video streaming, are examples of applications that can use cloud computing to leverage their provisioning capacity. This way, it is possible to offer a large variety of multimedia content in many formats, so the users will be able to watch videos as they wish, with a proper resolution and quality, according to his preferences and connection speed. Private infrastructures for Video on Demand (VoD) and live video streaming are especially useful for e-learning on large corporations, universities, and governments. Analytical models are effective tools to evaluate the availability of software, hardware, and other computational resources. In this paper, we study a VoD service hosted in a private cloud computing environment. We present availability models considering the VoD streaming server components that are necessary for viewers access. Hierarchical modeling techniques are used to deal with the complexity of representing such system. Sensitivity analysis is used to determine the parameters that cause greatest impact on the availability, identifying which components require attention when attempting to achieve increased availability in a system. The proposed models are useful for planning private cloud infrastructures for VoD services.
Proceedings of the 30th Annual ACM Symposium on Applied Computing, 2015
Resource scarcity is a major obstacle for many mobile applications, since devices have limited en... more Resource scarcity is a major obstacle for many mobile applications, since devices have limited energy power and processing potential. As an example, there are applications that seamlessly augment human cognition and typically require resources that far outstrip mobile hardware's capabilities, such as language translation, speech recognition, and face recognition. The use of cloud computing may tackle this problem. This study presents SmartRank, a scheduling framework to perform load partitioning and offloading for mobile applications using cloud computing to increase performance in terms of response time. We first explore a benchmarking of face recognition application using mobile cloud and confirms its suitability to be used as case study with SmartRank. We have applied the approach to a face recognition process based on two strategies: cloudlet federation and resource ranking through balanced metrics (level of CPU utilization and round-trip time). Second, using a full factorial experimental design we tuned the SmartRank with the most suitable partitioning decision calibrating scheduling parameters. Nevertheless, SmartRank uses an equation that is extensible to include new parameters and make it applicable to other scenarios.
2018 IEEE 32nd International Conference on Advanced Information Networking and Applications (AINA), 2018
The Information Technology (IT) systems of most organizations must support their operations 24 ho... more The Information Technology (IT) systems of most organizations must support their operations 24 hours a day, 7 days a week. Systems unavailability may have serious consequences such as data loss, customer dissatisfaction, and subsequent revenue loss. With the popularity of cloud computing, the adoption of cloud-based disaster recovery (DR) solutions has gained more space to prevent data loss and ensure business continuity. However, disaster recovery solutions are not cheap and do not exist as a single solution that suits every requirement (e.g., availability and costs). In this paper, we present an integrated model-experiment approach to evaluate cloud-based disaster recovery solutions. We use Stochastic Petri Nets (SPNs) and fault-injection experiments to evaluate availability related metrics like steady-state availability and downtime. To demonstrate the feasibility of our approach, distinct real-world cloud-based DR solutions (e.g., active/active and active/standby) are modeled and analyzed. The results revealed that disaster recovery solution significantly improves system availability and minimizes the downtime costs. In addition, our numerical analysis shows the statistical correspondence between the results of the experiments and models.
2021 Annual Reliability and Maintainability Symposium (RAMS)
SUMMARY & CONCLUSIONSSoftware aging is a topic in the literature that has matured over the pa... more SUMMARY & CONCLUSIONSSoftware aging is a topic in the literature that has matured over the past 30 years and can be defined as software degradation during its operational life. On the other hand, the software rejuvenation technique is a proactive measure that includes the application stop, removal of error accumulation conditions, and restart in a clean environment. This paper provides a systematic mapping of the contemporary literature on Software Aging and Rejuvenation (SAR), observing the research trends in this theme in the last decade, from 2010 to 2019. The results highlight which strategic approaches of rejuvenation or aging detection have been approached when designing solutions both in the commercial and academic fields, and thus, researchers can use them when dealing with new software aging phenomena. Besides, it can be of great relevance for the academic community that studies the SAR to provide a current view on state of the art in this topic.
International Journal of Web and Grid Services, 2016
The task of designing and tuning computer systems for optimal performance and dependability is co... more The task of designing and tuning computer systems for optimal performance and dependability is complex. This usually involves the choice of manufacturers, service providers, types of components, and possible architectures. Analytical modelling is helpful for planning and predicting the quality-of-service QoS measures of a system, and can be especially useful when integrated with optimisation algorithms. This study proposes a sensitive greedy randomised adaptive search procedure Sensitive GRASP, which uses sensitivity analysis of an analytical model to improve the solutions of a GRASP-based optimisation. Our approach is applied for optimising the reliability and performance of a composite web service. For scenarios with thousands of possible combinations, the proposed approach finds configurations that are closer to the optimal solution than those found by the algorithm without sensitivity analysis with equivalent computational effort.
ABSTRACT High availability in cloud computing services is essential for maintaining customer conf... more ABSTRACT High availability in cloud computing services is essential for maintaining customer confidence and avoiding revenue losses due to SLA violation penalties. Since the software and hardware components of cloud infrastructures may have limited reliability, the use of redundant components and multiple clusters may be required to achieve the expected level of dependability while also increasing the computational capacity. A drawback of such improvements is the respective impact on the capital and increase in acquisition and operational costs. This paper presents availability models for private cloud architectures based on Eucalyptus platform, and presents a comparison of costs between these architectures and similar infrastructure rented from a public cloud provider. Metrics for capacity-oriented availability and system steady-state availability are used to compare architectures with distinct numbers of clusters. A heterogeneous hierarchical modeling approach is employed to represent the systems considering both hardware and software failures. The results highlight that improvements on the availability are not significant when increasing the system to more than two clusters. The analysis also shows that the average available capacity is close to the maximum possible capacity in all architectures, and that it takes 18 months, in average, for these private cloud architectures to pay off the cost equivalent to the computational capacity rented from a public cloud. J. Dantas (B) · R. Matos · J. Araujo · P.
Uploads
Papers by Rubens S Matos