DFG QoE-DZ

QoE-DZ: Analysis and Optimization of the Trade-off between QoE and Energy-Efficiency in Data Centers

Head

Florian Wamser

Researchers

Summary

Energy-efficient operation of data centers is gaining more and more attention. However, saving energy usually leads to a degradation of the available computing resources, which results in reduced Quality of Service (QoS). The diminished QoS can in turn lead to poor Quality of Experience (QoE) for the end-users. Although several approaches already address the optimization of data centers either to ensure energy-efficiency or to resolve certain QoS constraints, the joint optimization of both objectives is not researched. Furthermore, QoE is supposed to enable a holistic understanding of the qualitative performance of networked communication systems and thus to complement the traditional, more technology-centric QoS perspective.

This project focuses on quantifying and adjusting the trade-off between QoE and energy-efficiency in data centers. We consider the use case of Virtual Desktop Infrastructures (VDIs) that are considered to have an increasing impact in the following years. A VDI enables users to use very lightweight systems, so-called thin clients, whereby the actual operating system runs in a data center. This approach may lead to cost and energy savings due to economies of scale by better utilization of existing resources in the data center. However, if too many services are aggregated onto very few servers, the service requirements cannot be fulfilled or result in significantly increased processing times. As a result, QoS and therefore QoE suffers. Additional challenges arise, as VDIs provide the user an entire operating system including various applications, e.g. Office products, with different requirements.

The project aims at classifying the application types running inside VDIs according to their functionality and possible requirements on the client device, transmission network, and data center. Detailed QoE models for specific application-classes in a VDI environment will be derived based on subjective user studies e.g. via crowdsourcing. Such a QoE model quantifies the relation between the end-user QoE and different influence factors on the entire transmission chain including the client, network, and data center side. In parallel, corresponding models for the energy-consumption in data centers will be defined based on existing studies. Assuming that all information on the entire transmission chain is available, this allows for deriving the optimal trade-off between QoE and energy-efficiency by adjusting, e.g., scheduling of service requests, hot and cold stand-by of servers, or virtual machine (VM) placement approaches. The methodology used in the proposed project combines literature studies, measurements based on dedicated testbeds, and subjective user studies with optimization methods using discrete-event simulations and analytical methods, like queuing theory.

Publications

Collecting Subjective Ratings in Enterprise Environments. Borchert, Kathrin; Hirth, Matthias; Zinner, Thomas; Göritz, Anja. In 9th International Conference on Quality of Multimedia Experience (QoMEX). Erfurt, Germany, 2017.
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
@inproceedings{info3-inproceedings-2017-14, address = {Erfurt, Germany}, author = {Borchert, Kathrin and Hirth, Matthias and Zinner, Thomas and Göritz, Anja}, booktitle = {9th International Conference on Quality of Multimedia Experience (QoMEX)}, keywords = {ua-net}, month = {05}, title = {Collecting Subjective Ratings in Enterprise Environments}, year = 2017 }

%0 Conference Paper %1 info3-inproceedings-2017-14 %A Borchert, Kathrin %A Hirth, Matthias %A Zinner, Thomas %A Göritz, Anja %B 9th International Conference on Quality of Multimedia Experience (QoMEX) %C Erfurt, Germany %D 2017 %T Collecting Subjective Ratings in Enterprise Environments
Performance Analysis of Hierarchical Caching Systems with Bandwidth Constraints. Burger, Valentin; Zinner, Thomas. In Proceedings of the International Telecommunication Networks and Applications Conference. Dunedin, New Zealand, 2016.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
The vast majority of Internet traffic is carried by content delivery networks. A high potential to bring content even closer to consumers and to reduce energy cost is achieved by the nano data center (NaDa) concept that has been proposed in recent work. In this approach the capacity available on small devices such as home gateways is used to support content delivery in an hierarchical caching system. In this work, we present a method to determine the efficiency of hierarchical content delivery networks with limited bandwidth, such as in the NaDa approach, analytically. We analyze the Internet Census Dataset to assess the number of available devices on autonomous system level. We evaluate the potential of hierarchical cache networks with bandwidth constraints to assist content delivery and determine its limits.

@inproceedings{info3-inproceedings-2016-22, abstract = {The vast majority of Internet traffic is carried by content delivery networks. A high potential to bring content even closer to consumers and to reduce energy cost is achieved by the nano data center (NaDa) concept that has been proposed in recent work. In this approach the capacity available on small devices such as home gateways is used to support content delivery in an hierarchical caching system. In this work, we present a method to determine the efficiency of hierarchical content delivery networks with limited bandwidth, such as in the NaDa approach, analytically. We analyze the Internet Census Dataset to assess the number of available devices on autonomous system level. We evaluate the potential of hierarchical cache networks with bandwidth constraints to assist content delivery and determine its limits.}, address = {Dunedin, New Zealand}, author = {Burger, Valentin and Zinner, Thomas}, booktitle = {Proceedings of the International Telecommunication Networks and Applications Conference}, keywords = {can}, month = 12, title = {Performance Analysis of Hierarchical Caching Systems with Bandwidth Constraints}, year = 2016 }

%0 Conference Paper %1 info3-inproceedings-2016-22 %A Burger, Valentin %A Zinner, Thomas %B Proceedings of the International Telecommunication Networks and Applications Conference %C Dunedin, New Zealand %D 2016 %T Performance Analysis of Hierarchical Caching Systems with Bandwidth Constraints %X The vast majority of Internet traffic is carried by content delivery networks. A high potential to bring content even closer to consumers and to reduce energy cost is achieved by the nano data center (NaDa) concept that has been proposed in recent work. In this approach the capacity available on small devices such as home gateways is used to support content delivery in an hierarchical caching system. In this work, we present a method to determine the efficiency of hierarchical content delivery networks with limited bandwidth, such as in the NaDa approach, analytically. We analyze the Internet Census Dataset to assess the number of available devices on autonomous system level. We evaluate the potential of hierarchical cache networks with bandwidth constraints to assist content delivery and determine its limits.
Modeling the YouTube Stack: from Packets to Quality of Experience. Wamser, Florian; Casas, Pedro; Seufert, Michael; Moldovan, Christian; Tran-Gia, Phuoc; Hoßfeld, Tobias. In Computer Networks, 109(2), pp. 211–224. 2016.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
YouTube is one of the most popular and volume-dominant services in today’s Internet, and has changed the Web for ever. Consequently, network operators are forced to consider it in the design, deployment, and optimization of their networks. Taming YouTube requires a good understanding of the complete YouTube stack, from the network streaming service to the application itself. Understanding the interplays between individual YouTube functionalities and their implications for traffic and user Quality of Experience (QoE) becomes paramount nowadays. In this paper we characterize and model the YouTube stack at different layers, going from the generated network traffic to the QoE perceived by the users watching YouTube videos. Firstly, we present a network traffic model for the YouTube flow control mechanism, which permits to understand how YouTube provisions video traffic flows to users. Secondly, we investigate how traffic is consumed at the client side, deriving a simple model for the YouTube application. Thirdly, we analyze the implications for the end user, and present a model for the quality as perceived by them. This model is finally integrated into a system for real time QoE-based YouTube monitoring, highly useful to operators to assess the performance of their networks for provisioning YouTube videos. The central parameter for all the presented models is the buffer level at the YouTube application layer. This paper provides an extensive compendium of objective tools and models for network operators to better understand the YouTube traffic in their networks, to predict the playback behavior of the video player, and to assess how well they do it in the practice with the satisfaction of their customers watching YouTube.

@article{info3-article-2016-6, abstract = {YouTube is one of the most popular and volume-dominant services in today’s Internet, and has changed the Web for ever. Consequently, network operators are forced to consider it in the design, deployment, and optimization of their networks. Taming YouTube requires a good understanding of the complete YouTube stack, from the network streaming service to the application itself. Understanding the interplays between individual YouTube functionalities and their implications for traffic and user Quality of Experience (QoE) becomes paramount nowadays. In this paper we characterize and model the YouTube stack at different layers, going from the generated network traffic to the QoE perceived by the users watching YouTube videos. Firstly, we present a network traffic model for the YouTube flow control mechanism, which permits to understand how YouTube provisions video traffic flows to users. Secondly, we investigate how traffic is consumed at the client side, deriving a simple model for the YouTube application. Thirdly, we analyze the implications for the end user, and present a model for the quality as perceived by them. This model is finally integrated into a system for real time QoE-based YouTube monitoring, highly useful to operators to assess the performance of their networks for provisioning YouTube videos. The central parameter for all the presented models is the buffer level at the YouTube application layer. This paper provides an extensive compendium of objective tools and models for network operators to better understand the YouTube traffic in their networks, to predict the playback behavior of the video player, and to assess how well they do it in the practice with the satisfaction of their customers watching YouTube.}, author = {Wamser, Florian and Casas, Pedro and Seufert, Michael and Moldovan, Christian and Tran-Gia, Phuoc and Hoßfeld, Tobias}, journal = {Computer Networks}, keywords = {can}, month = 11, number = 2, pages = {211-224}, title = {Modeling the YouTube Stack: from Packets to Quality of Experience}, volume = 109, year = 2016 }

%0 Journal Article %1 info3-article-2016-6 %A Wamser, Florian %A Casas, Pedro %A Seufert, Michael %A Moldovan, Christian %A Tran-Gia, Phuoc %A Hoßfeld, Tobias %D 2016 %J Computer Networks %N 2 %P 211-224 %T Modeling the YouTube Stack: from Packets to Quality of Experience %V 109 %X YouTube is one of the most popular and volume-dominant services in today’s Internet, and has changed the Web for ever. Consequently, network operators are forced to consider it in the design, deployment, and optimization of their networks. Taming YouTube requires a good understanding of the complete YouTube stack, from the network streaming service to the application itself. Understanding the interplays between individual YouTube functionalities and their implications for traffic and user Quality of Experience (QoE) becomes paramount nowadays. In this paper we characterize and model the YouTube stack at different layers, going from the generated network traffic to the QoE perceived by the users watching YouTube videos. Firstly, we present a network traffic model for the YouTube flow control mechanism, which permits to understand how YouTube provisions video traffic flows to users. Secondly, we investigate how traffic is consumed at the client side, deriving a simple model for the YouTube application. Thirdly, we analyze the implications for the end user, and present a model for the quality as perceived by them. This model is finally integrated into a system for real time QoE-based YouTube monitoring, highly useful to operators to assess the performance of their networks for provisioning YouTube videos. The central parameter for all the presented models is the buffer level at the YouTube application layer. This paper provides an extensive compendium of objective tools and models for network operators to better understand the YouTube traffic in their networks, to predict the playback behavior of the video player, and to assess how well they do it in the practice with the satisfaction of their customers watching YouTube.
Correlating QoE and Technical Parameters of an SAP System in an Enterprise Environment. Borchert, Kathrin; Hirth, Matthias; Zinner, Thomas; Mocanu, Decebal Constantin. In QCMan IEEE Workshop on QoE centric Management. 2016.
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
@inproceedings{info3-inproceedings-2016-15, author = {Borchert, Kathrin and Hirth, Matthias and Zinner, Thomas and Mocanu, Decebal Constantin}, booktitle = {QCMan IEEE Workshop on QoE centric Management}, keywords = {oekonet}, month = {09}, title = {Correlating QoE and Technical Parameters of an SAP System in an Enterprise Environment}, year = 2016 }

%0 Conference Paper %1 info3-inproceedings-2016-15 %A Borchert, Kathrin %A Hirth, Matthias %A Zinner, Thomas %A Mocanu, Decebal Constantin %B QCMan IEEE Workshop on QoE centric Management %D 2016 %T Correlating QoE and Technical Parameters of an SAP System in an Enterprise Environment
Assessing the Accuracy of Network Estimations in the DOTA 2 Game Client. Hirth, Matthias; Allendorf, Fabian; Metzger, Florian; Schwartz, Christian. In Workshop on Perceptual Quality of Systems (PQS). Berlin, Germany, 2016.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Online video games and the subjective quality of player inter- actions with them rely on good network conditions. Almost equally important is a good and timely knowledge of such conditions in order to take proper countermeasures to worsening conditions. To this end, many online video games include components for estimating the current network status. In this paper, we examine the accuracy of those estimations for the popular competitive multiplayer game DOTA 2. Our results show that the game client is capable of performing a good estimation of the delay and packet loss, but only after a rather large initial delay.

@inproceedings{info3-inproceedings-2016-13, abstract = {Online video games and the subjective quality of player inter- actions with them rely on good network conditions. Almost equally important is a good and timely knowledge of such conditions in order to take proper countermeasures to worsening conditions. To this end, many online video games include components for estimating the current network status. In this paper, we examine the accuracy of those estimations for the popular competitive multiplayer game DOTA 2. Our results show that the game client is capable of performing a good estimation of the delay and packet loss, but only after a rather large initial delay.}, address = {Berlin, Germany}, author = {Hirth, Matthias and Allendorf, Fabian and Metzger, Florian and Schwartz, Christian}, booktitle = {Workshop on Perceptual Quality of Systems (PQS)}, keywords = {sys:hidden:inproceedings}, month = {08}, title = {Assessing the Accuracy of Network Estimations in the DOTA 2 Game Client}, year = 2016 }

%0 Conference Paper %1 info3-inproceedings-2016-13 %A Hirth, Matthias %A Allendorf, Fabian %A Metzger, Florian %A Schwartz, Christian %B Workshop on Perceptual Quality of Systems (PQS) %C Berlin, Germany %D 2016 %T Assessing the Accuracy of Network Estimations in the DOTA 2 Game Client %X Online video games and the subjective quality of player inter- actions with them rely on good network conditions. Almost equally important is a good and timely knowledge of such conditions in order to take proper countermeasures to worsening conditions. To this end, many online video games include components for estimating the current network status. In this paper, we examine the accuracy of those estimations for the popular competitive multiplayer game DOTA 2. Our results show that the game client is capable of performing a good estimation of the delay and packet loss, but only after a rather large initial delay.
Concept for Client-initiated Selection of Cloud Instances for Improving QoE of Distributed Cloud Services. Wamser, Florian; Seufert, Michael; Höfner, Steffen; Tran-Gia, Phuoc. In ACM SIGCOMM Workshop on QoE-based Analysis and Management of Data Communication Networks (Internet-QoE). Florianópolis, Brazil, 2016.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
We introduce a concept for client-initiated selection of service location and service quality for improving the Quality of Experience (QoE) of general cloud services. It is loosely based on the HTTP adaptive streaming approach (e.g., MPEG DASH). A manifest file compiled by the cloud service provider specifies the available service locations and qualities, from which the user selects the optimal service instance based on contextual information obtained from client measurements and user preferences. The proposed concept is defined and is implemented in two client-based decision algorithms for improving the QoE of a simple picture gallery cloud service. These decision algorithms are evaluated and their impact on the service delivery is discussed. The evaluation shows that it is possible to improve the service location and quality selection by light-weight client-based algorithms.

@inproceedings{info3-inproceedings-2016-14, abstract = {We introduce a concept for client-initiated selection of service location and service quality for improving the Quality of Experience (QoE) of general cloud services. It is loosely based on the HTTP adaptive streaming approach (e.g., MPEG DASH). A manifest file compiled by the cloud service provider specifies the available service locations and qualities, from which the user selects the optimal service instance based on contextual information obtained from client measurements and user preferences. The proposed concept is defined and is implemented in two client-based decision algorithms for improving the QoE of a simple picture gallery cloud service. These decision algorithms are evaluated and their impact on the service delivery is discussed. The evaluation shows that it is possible to improve the service location and quality selection by light-weight client-based algorithms.}, address = {Florianópolis, Brazil}, author = {Wamser, Florian and Seufert, Michael and Höfner, Steffen and Tran-Gia, Phuoc}, booktitle = {ACM SIGCOMM Workshop on QoE-based Analysis and Management of Data Communication Networks (Internet-QoE)}, keywords = {can}, month = {08}, title = {Concept for Client-initiated Selection of Cloud Instances for Improving QoE of Distributed Cloud Services}, year = 2016 }

%0 Conference Paper %1 info3-inproceedings-2016-14 %A Wamser, Florian %A Seufert, Michael %A Höfner, Steffen %A Tran-Gia, Phuoc %B ACM SIGCOMM Workshop on QoE-based Analysis and Management of Data Communication Networks (Internet-QoE) %C Florianópolis, Brazil %D 2016 %T Concept for Client-initiated Selection of Cloud Instances for Improving QoE of Distributed Cloud Services %X We introduce a concept for client-initiated selection of service location and service quality for improving the Quality of Experience (QoE) of general cloud services. It is loosely based on the HTTP adaptive streaming approach (e.g., MPEG DASH). A manifest file compiled by the cloud service provider specifies the available service locations and qualities, from which the user selects the optimal service instance based on contextual information obtained from client measurements and user preferences. The proposed concept is defined and is implemented in two client-based decision algorithms for improving the QoE of a simple picture gallery cloud service. These decision algorithms are evaluated and their impact on the service delivery is discussed. The evaluation shows that it is possible to improve the service location and quality selection by light-weight client-based algorithms.
ERWIN - Enabling the Reproducible Investigation of WaItiNg Times for Arbitrary Workflows. Zinner, Thomas; Hirth, Matthias; Fischer, Valentin; Hohlfeld, Oliver. In 8th International Conference on Quality of Multimedia Experiene (QoMEX). Lisbon, Portugal, 2016.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Delay effects can impact the Quality of Experience of interactive systems, which motivates research assessing delay impairments, mostly for web based systems. Current studies follow individual methodologies and typically assesses individual and custom-made web pages, whose construction requires expert knowledge in web technologies. A range of native, non-web applications cannot be easily modified for delay studies. Thus, a generalized methodology for assessing delay impacts for a broad range of applications that is assessable to researchers without (web) development expertise is still missing. This paper con- tributes to this open problem by i) presenting a new methodology for reproducible delay assessments in a broad class of (web-based) systems and ii) presenting an open-source implementation to be used by the community. This methodology particularly aims at making delay assessment available to a broad range of researchers by avoiding programming skills and thus by lowering the barrier for setting-up delay assessments.

@inproceedings{info3-inproceedings-2016-6, abstract = {Delay effects can impact the Quality of Experience of interactive systems, which motivates research assessing delay impairments, mostly for web based systems. Current studies follow individual methodologies and typically assesses individual and custom-made web pages, whose construction requires expert knowledge in web technologies. A range of native, non-web applications cannot be easily modified for delay studies. Thus, a generalized methodology for assessing delay impacts for a broad range of applications that is assessable to researchers without (web) development expertise is still missing. This paper con- tributes to this open problem by i) presenting a new methodology for reproducible delay assessments in a broad class of (web-based) systems and ii) presenting an open-source implementation to be used by the community. This methodology particularly aims at making delay assessment available to a broad range of researchers by avoiding programming skills and thus by lowering the barrier for setting-up delay assessments.}, address = {Lisbon, Portugal}, author = {Zinner, Thomas and Hirth, Matthias and Fischer, Valentin and Hohlfeld, Oliver}, booktitle = {8th International Conference on Quality of Multimedia Experiene (QoMEX)}, keywords = {sys:hidden:inproceedings}, month = {06}, title = {ERWIN - Enabling the Reproducible Investigation of WaItiNg Times for Arbitrary Workflows}, year = 2016 }

%0 Conference Paper %1 info3-inproceedings-2016-6 %A Zinner, Thomas %A Hirth, Matthias %A Fischer, Valentin %A Hohlfeld, Oliver %B 8th International Conference on Quality of Multimedia Experiene (QoMEX) %C Lisbon, Portugal %D 2016 %T ERWIN - Enabling the Reproducible Investigation of WaItiNg Times for Arbitrary Workflows %X Delay effects can impact the Quality of Experience of interactive systems, which motivates research assessing delay impairments, mostly for web based systems. Current studies follow individual methodologies and typically assesses individual and custom-made web pages, whose construction requires expert knowledge in web technologies. A range of native, non-web applications cannot be easily modified for delay studies. Thus, a generalized methodology for assessing delay impacts for a broad range of applications that is assessable to researchers without (web) development expertise is still missing. This paper con- tributes to this open problem by i) presenting a new methodology for reproducible delay assessments in a broad class of (web-based) systems and ii) presenting an open-source implementation to be used by the community. This methodology particularly aims at making delay assessment available to a broad range of researchers by avoiding programming skills and thus by lowering the barrier for setting-up delay assessments.
Text Categorization for Deriving the Application Quality in Enterprises using Ticketing Systems. Zinner, Thomas; Lemmerich, Florian; Schwarzmann, Susanna; Hirth, Matthias; Karg, Peter; Hotho, Andreas. In 17th International Conference on Big Data Analytics and Knowledge Discovery (DaWaK 2015). Valencia, Spain, 2015.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Today’s enterprise services and business applications are of- ten centralized in a small number of data centers. Employees located at branches and side offices access the computing infrastructure via the internet using thin client architectures. The task to provide a good application quality to the employers using a multitude of different applications and access networks has thus become complex. Enterprises have to be able to identify resource bottlenecks and applications with a poor performance quickly to take appropriate countermeasures and enable a good application quality for their employees. Ticketing systems within an enterprise use large databases for collecting complaints and problems of the users over a long period of time and thus are an interesting starting point to identify performance problems. However, manual categorization of tickets comes with a high workload. In this paper, we analyze in a case study the applicability of supervised learning algorithms for the automatic identification of relevant tickets, i.e., tickets indicating problematic applications. In that regard, we evaluate different classification algorithms using 12,000 manually annotated tickets accumulated in July 2013 at the ticketing system of a nation- wide operating enterprise. In addition to traditional machine learning metrics, we also analyze the performance of the different classifiers on business-relevant metrics.

@inproceedings{info3-inproceedings-2015-517, abstract = {Today’s enterprise services and business applications are of- ten centralized in a small number of data centers. Employees located at branches and side offices access the computing infrastructure via the internet using thin client architectures. The task to provide a good application quality to the employers using a multitude of different applications and access networks has thus become complex. Enterprises have to be able to identify resource bottlenecks and applications with a poor performance quickly to take appropriate countermeasures and enable a good application quality for their employees. Ticketing systems within an enterprise use large databases for collecting complaints and problems of the users over a long period of time and thus are an interesting starting point to identify performance problems. However, manual categorization of tickets comes with a high workload. In this paper, we analyze in a case study the applicability of supervised learning algorithms for the automatic identification of relevant tickets, i.e., tickets indicating problematic applications. In that regard, we evaluate different classification algorithms using 12,000 manually annotated tickets accumulated in July 2013 at the ticketing system of a nation- wide operating enterprise. In addition to traditional machine learning metrics, we also analyze the performance of the different classifiers on business-relevant metrics.}, address = {Valencia, Spain}, author = {Zinner, Thomas and Lemmerich, Florian and Schwarzmann, Susanna and Hirth, Matthias and Karg, Peter and Hotho, Andreas}, booktitle = {17th International Conference on Big Data Analytics and Knowledge Discovery (DaWaK 2015)}, keywords = {sys:hidden:inproceedings}, month = {09}, title = {Text Categorization for Deriving the Application Quality in Enterprises using Ticketing Systems}, year = 2015 }

%0 Conference Paper %1 info3-inproceedings-2015-517 %A Zinner, Thomas %A Lemmerich, Florian %A Schwarzmann, Susanna %A Hirth, Matthias %A Karg, Peter %A Hotho, Andreas %B 17th International Conference on Big Data Analytics and Knowledge Discovery (DaWaK 2015) %C Valencia, Spain %D 2015 %T Text Categorization for Deriving the Application Quality in Enterprises using Ticketing Systems %X Today’s enterprise services and business applications are of- ten centralized in a small number of data centers. Employees located at branches and side offices access the computing infrastructure via the internet using thin client architectures. The task to provide a good application quality to the employers using a multitude of different applications and access networks has thus become complex. Enterprises have to be able to identify resource bottlenecks and applications with a poor performance quickly to take appropriate countermeasures and enable a good application quality for their employees. Ticketing systems within an enterprise use large databases for collecting complaints and problems of the users over a long period of time and thus are an interesting starting point to identify performance problems. However, manual categorization of tickets comes with a high workload. In this paper, we analyze in a case study the applicability of supervised learning algorithms for the automatic identification of relevant tickets, i.e., tickets indicating problematic applications. In that regard, we evaluate different classification algorithms using 12,000 manually annotated tickets accumulated in July 2013 at the ticketing system of a nation- wide operating enterprise. In addition to traditional machine learning metrics, we also analyze the performance of the different classifiers on business-relevant metrics.
Analyzing the Impact of Delay and Packet Loss on Google Docs. Dinh-Xuan, Lam; Schwartz, Christian; Hirth, Matthias; Wamser, Florian; Truong Thu, Huong. In 7th International Conference on Mobile Networks and Management. Santander, Spain, 2015.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Software as a Service allows end users to use complex soft- ware directly from their browsers, transferring heavy computation to servers in the cloud. One use of this paradigm is word processing, former a classic use cases of Thin-Client computing. Similar to Thin-Client systems, the network parameters are an important influence factor for the cloud application performance. In this paper, we study Google Docs as an example for online word processing tools. We consider a traditional single user scenario as well as a collaborative scenario with two users working on one document simultaneously. We identify multiple relevant sub-processes per scenario as performance metrics and use a testbed to automatically evaluate the performance of Google Docs under varying network parameters. The main contributions of the paper are (1) the quantification of the impact of network parameters such as delay and packet loss on application performance metrics for both scenarios and (2) linear regression models to derive the application performance for giving network parameters.

@inproceedings{info3-inproceedings-2015-522, abstract = {Software as a Service allows end users to use complex soft- ware directly from their browsers, transferring heavy computation to servers in the cloud. One use of this paradigm is word processing, former a classic use cases of Thin-Client computing. Similar to Thin-Client systems, the network parameters are an important influence factor for the cloud application performance. In this paper, we study Google Docs as an example for online word processing tools. We consider a traditional single user scenario as well as a collaborative scenario with two users working on one document simultaneously. We identify multiple relevant sub-processes per scenario as performance metrics and use a testbed to automatically evaluate the performance of Google Docs under varying network parameters. The main contributions of the paper are (1) the quantification of the impact of network parameters such as delay and packet loss on application performance metrics for both scenarios and (2) linear regression models to derive the application performance for giving network parameters.}, address = {Santander, Spain}, author = {Dinh-Xuan, Lam and Schwartz, Christian and Hirth, Matthias and Wamser, Florian and Truong Thu, Huong}, booktitle = {7th International Conference on Mobile Networks and Management}, keywords = {can}, month = {09}, title = {Analyzing the Impact of Delay and Packet Loss on Google Docs}, year = 2015 }

%0 Conference Paper %1 info3-inproceedings-2015-522 %A Dinh-Xuan, Lam %A Schwartz, Christian %A Hirth, Matthias %A Wamser, Florian %A Truong Thu, Huong %B 7th International Conference on Mobile Networks and Management %C Santander, Spain %D 2015 %T Analyzing the Impact of Delay and Packet Loss on Google Docs %X Software as a Service allows end users to use complex soft- ware directly from their browsers, transferring heavy computation to servers in the cloud. One use of this paradigm is word processing, former a classic use cases of Thin-Client computing. Similar to Thin-Client systems, the network parameters are an important influence factor for the cloud application performance. In this paper, we study Google Docs as an example for online word processing tools. We consider a traditional single user scenario as well as a collaborative scenario with two users working on one document simultaneously. We identify multiple relevant sub-processes per scenario as performance metrics and use a testbed to automatically evaluate the performance of Google Docs under varying network parameters. The main contributions of the paper are (1) the quantification of the impact of network parameters such as delay and packet loss on application performance metrics for both scenarios and (2) linear regression models to derive the application performance for giving network parameters.
Deriving the Employee-perceived Application Quality in Enterprise IT Infrastructures using Information from Ticketing Systems. Schwarzmann, Susanna; Zinner, Thomas; Hirth, Matthias. In Workshop on Knowledge Discovery, Data Mining and Machine Learning (KDML 2014). Aachen, Germany, 2014.
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
@inproceedings{info3-inproceedings-2014-502, address = {Aachen, Germany}, author = {Schwarzmann, Susanna and Zinner, Thomas and Hirth, Matthias}, booktitle = {Workshop on Knowledge Discovery, Data Mining and Machine Learning (KDML 2014)}, keywords = {sys:hidden:inproceedings}, month = {09}, title = {Deriving the Employee-perceived Application Quality in Enterprise IT Infrastructures using Information from Ticketing Systems}, year = 2014 }

%0 Conference Paper %1 info3-inproceedings-2014-502 %A Schwarzmann, Susanna %A Zinner, Thomas %A Hirth, Matthias %B Workshop on Knowledge Discovery, Data Mining and Machine Learning (KDML 2014) %C Aachen, Germany %D 2014 %T Deriving the Employee-perceived Application Quality in Enterprise IT Infrastructures using Information from Ticketing Systems
Performance Model for Waiting Times in Cloud File Synchronization Services. Schwartz, Christian; Hirth, Matthias; Hoßfeld, Tobias; Tran-Gia, Phuoc. In 26th International Teletraffic Congress (ITC). Karlskrona, Sweden, 2014.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Over the last few years, the importance of cloud services for file synchronization has been increasing. With the release of network enabled cameras like Google Glass the trend of synchronizing new photos with the cloud has emerged as an important new use case for mobile networks. Multiple stakeholder are interested in optimizing this synchronization process according to different orthogonal metrics. The end user is interested in a fast synchronization of images as well as a low energy consumption of the client used for upload to the network. The network operator requires a low number of network connections per synchronization of a batch of images in order to reduce the signaling load of the mobile network. This paper answers the question if the goals of the stakeholders can be achieved by selecting an appropriate synchronization scheduling mechanism. Therefore, we first present a model for the waiting time of cloud file synchronization services. Second, we perform measurements on the popular Dropbox service to obtain parameters required in the model. Finally, we perform a parameter study over the considered mechanisms and parameters, suggest a preferable algorithm and identify trade-offs favorable for all stakeholders.

@inproceedings{info3-inproceedings-2014-495, abstract = {Over the last few years, the importance of cloud services for file synchronization has been increasing. With the release of network enabled cameras like Google Glass the trend of synchronizing new photos with the cloud has emerged as an important new use case for mobile networks. Multiple stakeholder are interested in optimizing this synchronization process according to different orthogonal metrics. The end user is interested in a fast synchronization of images as well as a low energy consumption of the client used for upload to the network. The network operator requires a low number of network connections per synchronization of a batch of images in order to reduce the signaling load of the mobile network. This paper answers the question if the goals of the stakeholders can be achieved by selecting an appropriate synchronization scheduling mechanism. Therefore, we first present a model for the waiting time of cloud file synchronization services. Second, we perform measurements on the popular Dropbox service to obtain parameters required in the model. Finally, we perform a parameter study over the considered mechanisms and parameters, suggest a preferable algorithm and identify trade-offs favorable for all stakeholders.}, address = {Karlskrona, Sweden}, author = {Schwartz, Christian and Hirth, Matthias and Hoßfeld, Tobias and Tran-Gia, Phuoc}, booktitle = {26th International Teletraffic Congress (ITC)}, keywords = {smartenit}, month = {09}, title = {Performance Model for Waiting Times in Cloud File Synchronization Services}, year = 2014 }

%0 Conference Paper %1 info3-inproceedings-2014-495 %A Schwartz, Christian %A Hirth, Matthias %A Hoßfeld, Tobias %A Tran-Gia, Phuoc %B 26th International Teletraffic Congress (ITC) %C Karlskrona, Sweden %D 2014 %T Performance Model for Waiting Times in Cloud File Synchronization Services %X Over the last few years, the importance of cloud services for file synchronization has been increasing. With the release of network enabled cameras like Google Glass the trend of synchronizing new photos with the cloud has emerged as an important new use case for mobile networks. Multiple stakeholder are interested in optimizing this synchronization process according to different orthogonal metrics. The end user is interested in a fast synchronization of images as well as a low energy consumption of the client used for upload to the network. The network operator requires a low number of network connections per synchronization of a batch of images in order to reduce the signaling load of the mobile network. This paper answers the question if the goals of the stakeholders can be achieved by selecting an appropriate synchronization scheduling mechanism. Therefore, we first present a model for the waiting time of cloud file synchronization services. Second, we perform measurements on the popular Dropbox service to obtain parameters required in the model. Finally, we perform a parameter study over the considered mechanisms and parameters, suggest a preferable algorithm and identify trade-offs favorable for all stakeholders.
GTP-based Load Model and Virtualization Gain for a Mobile Network’s GGSN. Metzger, Florian; Schwartz, Christian; Hoßfeld, Tobias. In 5th International Conference on Communications and Electronics. Da Nang, Vietnam, 2014.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Multiple outages in major mobile networks have been reported in the recent past. In fixed and datacenter networks, such capacity problems are solved by scaling out, i.e. purchasing additional hardware. In mobile networks this is not easily possible as network components are usually sold as sealed middleboxes. With the advances in server performance and SDN it has been suggested to virtualize these boxes. This also opens up opportunities to dimension according to current load and save energy by switching off parts of the infrastructure. Such suggestions immediately raise the question on the cost of virtualization. To answer this question, we introduce models for both a traditional as well as a virtualized GGSN. In addition, we provide distributions for the load experienced at the GGSN based on network measurements. With this at hand, we study the influence of different dimensioning parameters on important performance metrics, with special consideration for the impact of provisioning new instances for the virtual GGSN.

@inproceedings{info3-inproceedings-2014-497, abstract = {Multiple outages in major mobile networks have been reported in the recent past. In fixed and datacenter networks, such capacity problems are solved by scaling out, i.e. purchasing additional hardware. In mobile networks this is not easily possible as network components are usually sold as sealed middleboxes. With the advances in server performance and SDN it has been suggested to virtualize these boxes. This also opens up opportunities to dimension according to current load and save energy by switching off parts of the infrastructure. Such suggestions immediately raise the question on the cost of virtualization. To answer this question, we introduce models for both a traditional as well as a virtualized GGSN. In addition, we provide distributions for the load experienced at the GGSN based on network measurements. With this at hand, we study the influence of different dimensioning parameters on important performance metrics, with special consideration for the impact of provisioning new instances for the virtual GGSN.}, address = {Da Nang, Vietnam}, author = {Metzger, Florian and Schwartz, Christian and Hoßfeld, Tobias}, booktitle = {5th International Conference on Communications and Electronics}, keywords = {sys:hidden:inproceedings}, month = {07}, title = {GTP-based Load Model and Virtualization Gain for a Mobile Network's GGSN}, year = 2014 }

%0 Conference Paper %1 info3-inproceedings-2014-497 %A Metzger, Florian %A Schwartz, Christian %A Hoßfeld, Tobias %B 5th International Conference on Communications and Electronics %C Da Nang, Vietnam %D 2014 %T GTP-based Load Model and Virtualization Gain for a Mobile Network's GGSN %X Multiple outages in major mobile networks have been reported in the recent past. In fixed and datacenter networks, such capacity problems are solved by scaling out, i.e. purchasing additional hardware. In mobile networks this is not easily possible as network components are usually sold as sealed middleboxes. With the advances in server performance and SDN it has been suggested to virtualize these boxes. This also opens up opportunities to dimension according to current load and save energy by switching off parts of the infrastructure. Such suggestions immediately raise the question on the cost of virtualization. To answer this question, we introduce models for both a traditional as well as a virtualized GGSN. In addition, we provide distributions for the load experienced at the GGSN based on network measurements. With this at hand, we study the influence of different dimensioning parameters on important performance metrics, with special consideration for the impact of provisioning new instances for the virtual GGSN.
Need for Speed? On Quality of Experience for File Storage Services. Amrehn, Philipp; Vandenbroucke, Karel; Hoßfeld, Tobias; de Moor, Katrien; Hirth, Matthias; Schatz, Raimund; Casas, Pedro. In 4th International Workshop on Perceptual Quality of Systems (PQS 2013). Vienna, Austria, 2013.
- [ Abstract ]
- [ BibTeX ]
- [ EndNote ]
- [ Download ]
Cloud computing has received a lot of attention within several research domains. However, there are only a few studies that focus on Quality of Experience (QoE) from a user perspective in the context of personal cloud storage services. In order to gain insights in existing use practices and to detect possible QoE influencing factors and relevant features, we conducted an online survey amongst users of personal cloud storage services and applications in general (N=349) and one specifically targeted at Dropbox users (N=49). Both studies point to accessibility, data sharing and data back-up as key features of personal cloud storage services. In addition, performance, i.e. upload and download speed, goes hand in hand with synchro-nization speed and should be taken into account when measur-ing QoE. However, cost, privacy and security related issues are as important factors as well.

@inproceedings{info3-inproceedings-2013-469, abstract = {Cloud computing has received a lot of attention within several research domains. However, there are only a few studies that focus on Quality of Experience (QoE) from a user perspective in the context of personal cloud storage services. In order to gain insights in existing use practices and to detect possible QoE influencing factors and relevant features, we conducted an online survey amongst users of personal cloud storage services and applications in general (N=349) and one specifically targeted at Dropbox users (N=49). Both studies point to accessibility, data sharing and data back-up as key features of personal cloud storage services. In addition, performance, i.e. upload and download speed, goes hand in hand with synchro-nization speed and should be taken into account when measur-ing QoE. However, cost, privacy and security related issues are as important factors as well.}, address = {Vienna, Austria}, author = {Amrehn, Philipp and Vandenbroucke, Karel and Hoßfeld, Tobias and de Moor, Katrien and Hirth, Matthias and Schatz, Raimund and Casas, Pedro}, booktitle = {4th International Workshop on Perceptual Quality of Systems (PQS 2013)}, keywords = {smartenit}, month = {09}, title = {Need for Speed? On Quality of Experience for File Storage Services}, year = 2013 }

%0 Conference Paper %1 info3-inproceedings-2013-469 %A Amrehn, Philipp %A Vandenbroucke, Karel %A Hoßfeld, Tobias %A de Moor, Katrien %A Hirth, Matthias %A Schatz, Raimund %A Casas, Pedro %B 4th International Workshop on Perceptual Quality of Systems (PQS 2013) %C Vienna, Austria %D 2013 %T Need for Speed? On Quality of Experience for File Storage Services %X Cloud computing has received a lot of attention within several research domains. However, there are only a few studies that focus on Quality of Experience (QoE) from a user perspective in the context of personal cloud storage services. In order to gain insights in existing use practices and to detect possible QoE influencing factors and relevant features, we conducted an online survey amongst users of personal cloud storage services and applications in general (N=349) and one specifically targeted at Dropbox users (N=49). Both studies point to accessibility, data sharing and data back-up as key features of personal cloud storage services. In addition, performance, i.e. upload and download speed, goes hand in hand with synchro-nization speed and should be taken into account when measur-ing QoE. However, cost, privacy and security related issues are as important factors as well.

Student works and theses

Theses: