Change search
Link to record
Permanent link

Direct link
BETA
Stadler, Rolf, Prof.
Publications (10 of 14) Show all publications
Samani, F. S., Stadler, R., Johnsson, A. & Flinta, C. (2019). Demonstration: Predicting Distributions of Service Metrics. In: 2019 2019 IFIP/IEEE Symposium on Integrated Network and Service Management (IM): . Paper presented at 2019 IFIP/IEEE Symposium on Integrated Network and Service Management, IM 2019; Arlington; United States; 8 April 2019 through 12 April 2019 (pp. 745-746). Institute of Electrical and Electronics Engineers (IEEE), Article ID 8717915.
Open this publication in new window or tab >>Demonstration: Predicting Distributions of Service Metrics
2019 (English)In: 2019 2019 IFIP/IEEE Symposium on Integrated Network and Service Management (IM), Institute of Electrical and Electronics Engineers (IEEE), 2019, p. 745-746, article id 8717915Conference paper, Published paper (Refereed)
Abstract [en]

The ability to predict conditional distributions of service metrics is key to understanding end-to-end service behavior. From conditional distributions, other metrics can be derived, such as expected values and quantiles, which are essential for assessing SLA conformance. Our demonstrator predicts conditional distributions and derived metrics estimation in real-time, using infrastructure measurements. The distributions are modeled as Gaussian mixtures whose parameters are estimated using a mixture density network. The predictions are produced for a Video-on-Demand service that runs on a testbed at KTH.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2019
Keywords
Service Engineering, Service Management, Machine Learning
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-254135 (URN)000469937200144 ()978-3-903176-15-7 (ISBN)
Conference
2019 IFIP/IEEE Symposium on Integrated Network and Service Management, IM 2019; Arlington; United States; 8 April 2019 through 12 April 2019
Note

QC 20190625

Available from: 2019-06-25 Created: 2019-06-25 Last updated: 2019-06-25Bibliographically approved
Moradi, F., Stadler, R. & Johnsson, A. (2019). Performance Prediction in Dynamic Clouds using Transfer Learning. In: 2019 IFIP/IEEE Symposium on Integrated Network and Service Management, IM 2019: . Paper presented at 2019 IFIP/IEEE Symposium on Integrated Network and Service Management, IM 2019; Arlington; United States; 8 April 2019 through 12 April 2019 (pp. 242-250). IEEE, Article ID 8717847.
Open this publication in new window or tab >>Performance Prediction in Dynamic Clouds using Transfer Learning
2019 (English)In: 2019 IFIP/IEEE Symposium on Integrated Network and Service Management, IM 2019, IEEE, 2019, p. 242-250, article id 8717847Conference paper, Published paper (Refereed)
Abstract [en]

Learning a performance model for a cloud service is challenging since its operational environment changes during execution, which requires re-training of the model in order to maintain prediction accuracy. Training a new model from scratch generally involves extensive new measurements and often generates a data-collection overhead that negatively affects the service performance. In this paper, we investigate an approach for re-training neural-network models, which is based on transfer learning. Under this approach, a limited number of neural-network layers are re-trained while others remain unchanged. We study the accuracy of the re-trained model and the efficiency of the method with respect to the number of re-trained layers and the number of new measurements. The evaluation is performed using traces collected from a testbed that runs a Video-on-Demand service and a Key-Value Store under various load conditions. We study model re-training after changes in load pattern, infrastructure configuration, service configuration, and target metric. We find that our method significantly reduces the number of new measurements required to compute a new model after a change. The reduction exceeds an order of magnitude in most cases.

Place, publisher, year, edition, pages
IEEE, 2019
Keywords
Service Management, Performance Prediction, Machine Learning, Neural Networks, Transfer Learning
National Category
Other Computer and Information Science
Identifiers
urn:nbn:se:kth:diva-254134 (URN)000469937200055 ()2-s2.0-85067071723 (Scopus ID)9783903176157 (ISBN)
Conference
2019 IFIP/IEEE Symposium on Integrated Network and Service Management, IM 2019; Arlington; United States; 8 April 2019 through 12 April 2019
Note

QC 20190626

Available from: 2019-06-26 Created: 2019-06-26 Last updated: 2019-06-26Bibliographically approved
Uddin, M., Stadler, R. & Clemm, A. (2018). A bottom-up design for spatial search in large networks and clouds. International Journal of Network Management, 28(6), Article ID e2041.
Open this publication in new window or tab >>A bottom-up design for spatial search in large networks and clouds
2018 (English)In: International Journal of Network Management, ISSN 1055-7148, E-ISSN 1099-1190, Vol. 28, no 6, article id e2041Article in journal (Refereed) Published
Abstract [en]

APPENDIX Information in networked systems often has spatial semantics: routers, sensors, or virtual machines have coordinates in a geographical or virtual space, for instance. In this paper, we propose a design for a spatial search system that processes queries against spatial information that is maintained in local databases inside a large networked system. In contrast to previous works in spatial databases and peer-to-peer designs, our design is bottom-up, which makes query routing network aware and thus efficient, and which facilitates system bootstrapping and adaptation. Key to our design is a protocol that creates and maintains a distributed index of object locations based on information from local databases and the underlying network topology. The index builds upon minimum bounding rectangles to efficiently encode locations. We present a generic search protocol that is based on an echo protocol and uses the index to prune the search space and perform query routing. The response times of search queries increase with the diameter of the network, which is asymptotically optimal. We study the performance of the protocol through simulation in static and dynamic network environments, for different network topologies, and for network sizes up to 100 000 nodes. In most experiments, the overhead incurred by our protocol lies well below 30% of a hypothetical optimal protocol. In addition, the protocol provides high accuracy under significant churn.

Place, publisher, year, edition, pages
WILEY, 2018
National Category
Communication Systems
Identifiers
urn:nbn:se:kth:diva-239491 (URN)10.1002/nem.2041 (DOI)000449676600008 ()2-s2.0-85052708680 (Scopus ID)
Note

QC 20181128

Available from: 2018-11-28 Created: 2018-11-28 Last updated: 2018-11-28Bibliographically approved
Yanggratoke, R., Ahmed, J., Ardelius, J., Flinta, C., Johnsson, A., Gillblad, D. & Stadler, R. (2018). A service-agnostic method for predicting service metrics in real time. International Journal of Network Management, 28(2), Article ID e1991.
Open this publication in new window or tab >>A service-agnostic method for predicting service metrics in real time
Show others...
2018 (English)In: International Journal of Network Management, ISSN 1055-7148, E-ISSN 1099-1190, Vol. 28, no 2, article id e1991Article in journal (Refereed) Published
Abstract [en]

We predict performance metrics of cloud services using statistical learning, whereby the behaviour of a system is learned from observations. Specifically, we collect device and network statistics from a cloud testbed and apply regression methods to predict, in real-time, client-side service metrics for video streaming and key-value store services. Results from intensive evaluation on our testbed indicate that our method accurately predicts service metrics in real time (mean absolute error below 16% for video frame rate and read latency, for instance). Further, our method is service agnostic in the sense that it takes as input operating systems and network statistics instead of service-specific metrics. We show that feature set reduction significantly improves the prediction accuracy in our case, while simultaneously reducing model computation time. We find that the prediction accuracy decreases when, instead of a single service, both services run on the same testbed simultaneously or when the network quality on the path between the server cluster and the client deteriorates. Finally, we discuss the design and implementation of a real-time analytics engine, which processes streams of device statistics and service metrics from testbed sensors and produces model predictions through online learning.

Place, publisher, year, edition, pages
Wiley, 2018
Keywords
quality of service, cloud computing, real-time network analytics, statistical learning, machine learning
National Category
Telecommunications Computer Sciences
Identifiers
urn:nbn:se:kth:diva-225070 (URN)10.1002/nem.1991 (DOI)000427120900006 ()2-s2.0-85029351383 (Scopus ID)
Funder
VINNOVA, 2013-03895
Note

QC 20180328

Available from: 2018-03-28 Created: 2018-03-28 Last updated: 2018-03-28Bibliographically approved
Ahmed, J., Josefsson, T., Johnsson, A., Flinta, C., Moradi, F., Pasquini, R. & Stadler, R. (2018). Automated diagnostic of virtualized service performance degradation. In: IEEE/IFIP Network Operations and Management Symposium: Cognitive Management in a Cyber World, NOMS 2018. Paper presented at 2018 IEEE/IFIP Network Operations and Management Symposium, NOMS 2018, 23 April 2018 through 27 April 2018 (pp. 1-9). Institute of Electrical and Electronics Engineers Inc.
Open this publication in new window or tab >>Automated diagnostic of virtualized service performance degradation
Show others...
2018 (English)In: IEEE/IFIP Network Operations and Management Symposium: Cognitive Management in a Cyber World, NOMS 2018, Institute of Electrical and Electronics Engineers Inc. , 2018, p. 1-9Conference paper, Published paper (Refereed)
Abstract [en]

Service assurance for cloud applications is a challenging task and is an active area of research for academia and industry. One promising approach is to utilize machine learning for service quality prediction and fault detection so that suitable mitigation actions can be executed. In our previous work, we have shown how to predict service-level metrics in real-time just from operational data gathered at the server side. This gives the service provider early indications on whether the platform can support the current load demand. This paper provides the logical next step where we extend our work by proposing an automated detection and diagnostic capability for the performance faults manifesting themselves in cloud and datacenter environments. This is a crucial task to maintain the smooth operation of running services and minimizing downtime. We demonstrate the effectiveness of our approach which exploits the interpretative capabilities of Self- Organizing Maps (SOMs) to automatically detect and localize different performance faults for cloud services. © 2018 IEEE.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers Inc., 2018
Keywords
Fault detection, Fault localization, Machine learning, Service quality, System statistics, Video streaming, Conformal mapping, Learning systems, Quality of service, Self organizing maps, Automated detection, Automated diagnostics, Cloud applications, Self organizing maps(soms), Virtualized services
National Category
Telecommunications
Identifiers
urn:nbn:se:kth:diva-238086 (URN)10.1109/NOMS.2018.8406234 (DOI)2-s2.0-85050672220 (Scopus ID)9781538634165 (ISBN)
Conference
2018 IEEE/IFIP Network Operations and Management Symposium, NOMS 2018, 23 April 2018 through 27 April 2018
Note

Conference code: 137784; Export Date: 30 October 2018; Conference Paper; Funding details: VINNOVA; Funding details: 2013-03895, VINNOVA; Funding text: ACKNOWLEDGMENT This research has been supported by the Swedish Governmental Agency for Innovation Systems, VINNOVA, under grant 2013-03895.

QC 20190111

Available from: 2019-01-11 Created: 2019-01-11 Last updated: 2019-01-11Bibliographically approved
Samani, F. S. & Stadler, R. (2018). Predicting Distributions of Service Metrics using Neural Networks. In: Salsano, S Riggio, R Ahmed, T Samak, T DosSantos, CRP (Ed.), 2018 14TH INTERNATIONAL CONFERENCE ON NETWORK AND SERVICE MANAGEMENT (CNSM): . Paper presented at 14th International Conference on Network and Service Management (CNSM), NOV 05-09, 2018, Rome, ITALY (pp. 45-53). IEEE
Open this publication in new window or tab >>Predicting Distributions of Service Metrics using Neural Networks
2018 (English)In: 2018 14TH INTERNATIONAL CONFERENCE ON NETWORK AND SERVICE MANAGEMENT (CNSM) / [ed] Salsano, S Riggio, R Ahmed, T Samak, T DosSantos, CRP, IEEE , 2018, p. 45-53Conference paper, Published paper (Refereed)
Abstract [en]

We predict the conditional distributions of service metrics, such as response time or frame rate, from infrastructure measurements in a cloud environment. From such distributions, key statistics of the service metrics, including mean, variance, or percentiles can be computed, which are essential for predicting SLA conformance or enabling service assurance. We model the distributions as Gaussian mixtures, whose parameters we predict using mixture density networks, a class of neural networks. We apply the method to a Voll service and a KY store running on our lab testbed. The results validate the effectiveness of the method when applied to operational data. In the case of predicting the mean of the frame rate or response time, the accuracy matches that of random forest, a baseline model.

Place, publisher, year, edition, pages
IEEE, 2018
Series
International Conference on Network and Service Management, ISSN 2165-9605
Keywords
Service Engineering, Machine Learning, Generative Models, Network Management
National Category
Probability Theory and Statistics
Identifiers
urn:nbn:se:kth:diva-245985 (URN)000458916300006 ()2-s2.0-85060906697 (Scopus ID)978-3-9031-7614-0 (ISBN)
Conference
14th International Conference on Network and Service Management (CNSM), NOV 05-09, 2018, Rome, ITALY
Note

QC 20190311

Available from: 2019-03-11 Created: 2019-03-11 Last updated: 2019-08-20Bibliographically approved
Pasquini, R. & Stadler, R. (2017). Learning End-to-end Application QoS from OpenFlow Switch Statistics. In: 2017 IEEE CONFERENCE ON NETWORK SOFTWARIZATION (IEEE NETSOFT): . Paper presented at 3rd IEEE Conference on Network Softwarization (IEEE NetSoft) - Softwarization Sustaining a Hyper-Connected World - En Route to 5G, JUL 03-07, 2017, Bologna, ITALY. IEEE
Open this publication in new window or tab >>Learning End-to-end Application QoS from OpenFlow Switch Statistics
2017 (English)In: 2017 IEEE CONFERENCE ON NETWORK SOFTWARIZATION (IEEE NETSOFT), IEEE , 2017Conference paper (Refereed)
Abstract [en]

We use statistical learning to estimate end-to-end QoS metrics from device statistics, collected from a server cluster and an OpenFlow network. The results from our testbed, which runs a video-on-demand service and a key-value store, demonstrate that the learned models can estimate QoS metrics like frame rate or response time with errors bellow 10% for a given client. Interestingly, we find that service-level QoS metrics seem "encoded" in network statistics and it suffices to collect OpenFlow per port statistics to achieve accurate estimation at small overhead for data collection and model computation.

Place, publisher, year, edition, pages
IEEE, 2017
Keywords
Machine Learning, Network Analytics, OpenFlow, Quality of Service, Software-Defined Networking
National Category
Computer and Information Sciences
Identifiers
urn:nbn:se:kth:diva-215847 (URN)000410943600016 ()2-s2.0-85029372779 (Scopus ID)978-1-5090-6008-5 (ISBN)
Conference
3rd IEEE Conference on Network Softwarization (IEEE NetSoft) - Softwarization Sustaining a Hyper-Connected World - En Route to 5G, JUL 03-07, 2017, Bologna, ITALY
Note

QC 20171017. Forskningsfinansiärer: Brazilian National Council for Scientific and Technological Development, CNPq, through the Science Without Borders Program, 232586/2014-4  ; Swedish Governmental Agency for Innovation Systems, VINNOVA, through project SENDATE-EXTEND ; Swedish Research Council through the ACCESS Linnaeus Centre 

Available from: 2017-10-17 Created: 2017-10-17 Last updated: 2018-01-13Bibliographically approved
Stadler, R., Pasquini, R. & Fodor, V. (2017). Learning from Network Device Statistics. Journal of Network and Systems Management, 25(4), 672-698
Open this publication in new window or tab >>Learning from Network Device Statistics
2017 (English)In: Journal of Network and Systems Management, ISSN 1064-7570, E-ISSN 1573-7705, Vol. 25, no 4, p. 672-698Article in journal (Refereed) Published
Abstract [en]

We estimate end-to-end service metrics from network device statistics. Our approach is based upon statistical, supervised learning, whereby the mapping from device-level to service-level metrics is learned from observations, i.e., through monitoring the system. The approach enables end-to-end performance prediction without requiring an explicit model of the system, which is different from traditional engineering techniques that use stochastic modeling and simulation. The fact that end-to-end service metrics can be estimated from local network statistics with good accuracy in the scenarios we consider suggests that service-level properties are "encoded" in network-level statistics. We show that the set of network statistics needed for estimation can be reduced to a set of measurements along the network path between client and service backend, with little loss in estimation accuracy. The reported work is largely experimental and its results have been obtained through testbed measurements from a video streaming service and a KV store over an OpenFlow network.

Place, publisher, year, edition, pages
SPRINGER, 2017
Keywords
Network management, Machine learning, Statistical learning, Feature selection, End-to-end performance Prediction, Network analytics, OpenFlow
National Category
Computer and Information Sciences
Identifiers
urn:nbn:se:kth:diva-217759 (URN)10.1007/s10922-017-9426-z (DOI)000414211400002 ()2-s2.0-85029795404 (Scopus ID)
Note

QC 20171121

Available from: 2017-11-21 Created: 2017-11-21 Last updated: 2018-01-13Bibliographically approved
Ahmed, J., Johnsson, A., Moradi, F., Pasquini, R., Flinta, C. & Stadler, R. (2017). Online approach to performance fault localization for cloud and datacenter services. In: Proceedings of the IM 2017 - 2017 IFIP/IEEE International Symposium on Integrated Network and Service Management: . Paper presented at 15th IFIP/IEEE International Symposium on Integrated Network and Service Management, IM 2017, 8 May 2017 through 12 May 2017 (pp. 873-874). Institute of Electrical and Electronics Engineers Inc.
Open this publication in new window or tab >>Online approach to performance fault localization for cloud and datacenter services
Show others...
2017 (English)In: Proceedings of the IM 2017 - 2017 IFIP/IEEE International Symposium on Integrated Network and Service Management, Institute of Electrical and Electronics Engineers Inc. , 2017, p. 873-874Conference paper (Refereed)
Abstract [en]

Automated detection and diagnosis of the performance faults in cloud and datacenter environments is a crucial task to maintain smooth operation of different services and minimize downtime. We demonstrate an effective machine learning approach based on detecting metric correlation stability violations (CSV) for automated localization of performance faults for datacenter services running under dynamic load conditions. © 2017 IFIP.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers Inc., 2017
Keywords
Dynamic loads, Learning systems, Automated detection, Datacenter, Different services, Fault localization, Load condition, Machine learning approaches, Fault detection
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:kth:diva-216283 (URN)10.23919/INM.2017.7987390 (DOI)2-s2.0-85029446145 (Scopus ID)9783901882890 (ISBN)
Conference
15th IFIP/IEEE International Symposium on Integrated Network and Service Management, IM 2017, 8 May 2017 through 12 May 2017
Note

QC 20171213

Available from: 2017-12-13 Created: 2017-12-13 Last updated: 2017-12-13Bibliographically approved
Pasquini, R., Moradi, F., Ahmed, J., Johnsson, A., Flinta, C. & Stadler, R. (2017). Predicting SLA Conformance for Cluster-based Services. In: 2017 IFIP Networking Conference (IFIP NETWORKING) and Workshops: . Paper presented at IFIP Networking Conference (IFIP Networking),JUN 12-16, 2017, Stockholm, Sweden. IEEE
Open this publication in new window or tab >>Predicting SLA Conformance for Cluster-based Services
Show others...
2017 (English)In: 2017 IFIP Networking Conference (IFIP NETWORKING) and Workshops, IEEE, 2017Conference paper, Published paper (Refereed)
Abstract [en]

The ability to predict conformance or violation for given Service-level Agreements (SLAs) is critical for service assurance. We demonstrate a prototype for real-time conformance prediction based on the concept of the capacity region, which abstracts the underlying ICT infrastructure with respect to the load it can carry for a given SLA. The capacity region is estimated through measurements and statistical learning. We demonstrate prediction for a key-value store (Voldemort) that runs on a server cluster located at KTH.

Place, publisher, year, edition, pages
IEEE, 2017
Keywords
Capacity Region, Feasible Region, Real-time Prediction, Statistical Learning, Service-level Agreement (SLA)
National Category
Social and Economic Geography
Identifiers
urn:nbn:se:kth:diva-224115 (URN)000425452000050 ()978-3-9018-8294-4 (ISBN)
Conference
IFIP Networking Conference (IFIP Networking),JUN 12-16, 2017, Stockholm, Sweden
Funder
VINNOVASwedish Research Council
Note

QC 20180312

Available from: 2018-03-12 Created: 2018-03-12 Last updated: 2018-03-12Bibliographically approved
Organisations

Search in DiVA

Show all publications