Dr. Alexandru Murgu

InBCT 3.3.

SERAMON - Adaptive Service Aggregation in IP Mobile Networks

Team Leader:

Dr. Alexandru Murgu, Department of Mathematical Information Technology,
murgu@mit.jyu.fi, 014-2602757

Qualitative Aspects

• Heterogeneous user population characterization of the present available services in mobile IP networks is considered in terms of QoS / CoS differentiation.
• Adaptive multi-service management based on the dynamic estimation of built-in features of user populations on a long-term basis (trends on traffic samples and similarities) leads to IP service reformulation for detecting the statistical traffic abrupt changes of QoS and the corresponding class of service parameters.
• Dynamic tradeoff between the IP resources and service flexibility is handled via entropy aggregation as the implementation of IP session control mechanism.

Quality of Service Mechanism for IP Services

Multimedia service aggregation is designed to intelligently balance the quality of service (QoS) descriptors in edge/core media servers for optimum performance and reliability. QoS mechanisms are based on multiprotocol label switching (MPLS) streamlining and provisioning of differentiated services granular traffic, subject to service level agreements (SLAs) according to the following key operational aspects:

• QoS for streaming media network (streaming latency/throughput).
• IP streaming of differentiated service into representative aggregated classes.
• Service encapsulation for QoS class translation.
• Long-term service pattern/class estimation.
• Service aggregation of mobile user populations.
• Dynamic competition under resource constraints.

Adaptive Aggregation Research Issues

• QoS adaptivity implementation of network traffic streaming in order to respond to different requirements of multimedia communications (voice, data, video, web).
• aggregating addresses at the core servers (used in unicasting); multicast addresses are aggregated to encapsulate the multicast addresses containing the location information on group members; aggregation schemes for aggregate addresses either through a group identifier approach or an interface approach.
• interfacing: addresses are aggregated at the interface to allow group members and address allocation to be totally random; traffic mobility patterns are assessed and translated based on entropy type of statistical processing of group membership classifier scheme.

InBCT 3.3.

SERAMON - Adaptive Service Aggregation In IP Mobile Networks

• Mobiilissa IP verkossa meneillään olevien palvelujen karakterisointi heterogeenisena käyttäjäpopulaationa QoS/CoS eriyttämisen näkökulmasta.
• Palveluiden mukautuva hallinta, joka perustuu pitkäaikaisten käyttäjäpopulaatiossa esiintyvien piirteiden dynaamiseen arviointiin. Tämä johtaa tarjottavien IP palveluiden muokkaamiseen siten, että ne yrittävät statistisesti huomioida mahdollisia QoS ja CoS parametrien äkillisiä muutoksia.

Verkon multimediapalveluiden yhdistämisprosessi (aggregaatio) suunnitellaan älykkäästi mukauttamaan QoS määrityksiä reuna- ja keskuspalvelimilla optimaalisen suorituksen ja luotettavuuden takaamiseksi. Lisäksi pyritään tehostamaan QoS toteutusta mm. seuraavanlaisilla menetelmillä:

• QoS toteutus streaming tyyppisille medioille.
• Eriytettyjen IP palveluiden aggregointi tietyntyyppisiin luokkiin.
• Pitkäaikaisten palvelu- ja luokkatrendien ennustaminen.
• Liikkuvan käyttäjäpopulaation palveluiden aggregointi.
• Rajallisten resursseiden dynaaminen kilpailuttaminen.

Tarkoituksena mm. saada aikaan QoS parametrien mukautuminen vastaamaan multimediapalveluiden (ääni, data, video, www) erilaisia tarpeita, sekä ratkaista liikkuvaan käyttäjään liittyviä osoiteongelmia.

Lisätietoja yms.:
More information etc.:

A. Murgu, InBCT 3.3. SERAMON -- Adaptive Service Aggregation in IP Mobile Networks (DOC, 40 KB), Summary 2003.

A. Murgu, SERAMON - Adaptive Service Aggregation in IP Mobile Networks (PDF, 8.1 KB).

J. Takkinen, InBCT33_2002_JT.pdf (PDF, 32 KB). Lyhyt suomenkielinen esitys Seramon-hankkeesta.

A. Murgu, "Fuzzy Aggregation of Input-Output Service Level Dynamics in Multimedia Networks" (PDF, 104.8 KB), in Advances in Computer Cybernetics, Vol. XI, G.E. Lasker (ed.), pp. 54-62, 2002.

A. Murgu, "Information Conditioning Adaptation for Maximum Likelihood Mean Learning" (PDF, 129.4 KB), in "Learning and Adaptation in Stochastic and Statistical Systems", A. Murgu, G.E. Lasker (eds.), IIAS, 2002 (to appear).

A. Murgu, "Multi-class Service Dynamics Replication Using Evolutionary Selection Games" (PDF, 102.5 KB), in Proc. of 10th Intern. Symp. on Dynamic Games, ISDG-2002, St. Petersburg, Russia (to appear).

A. Murgu, "Kalman Filtering of Mixture Systems with Interval Dynamics" (PDF, 162.4 KB), in "Adaptive, Cooperative and Competitive Processes in System Modeling, Design and Analysis", A. Murgu, G.E. Lasker (eds.), IIAS, 2001, pp. 28-36.

A. Murgu, "Fuzzy Successive Averaging method in Multistage Optimization Systems" (PDF, 93.7 KB), in Proceedings of International Joint Conference on Neural Networks, IJCNN-2001, July 15-19, 2001, Washington, D.C., USA, pp. 2206-2211.

A. Murgu, "Minimum Cross Entropy Control in Stochastic Switching Systems" (PDF, 104.7 KB), in Proceedings of IEEE International Symposium on Information Theory, ISIT-2001, June 24-29, 2001, Washington, D.C., USA, pp. 205.

A. Murgu, "Switching Envelope for Statistical Aggregation of Traffic Flow Batches" (PDF, 108.5 KB), in Proceedings of Ninth International Symposium on Dynamics Games and Applications, ISDG-2000, Adelaide, Australia, December 18-21, 2000, pp. 375-383.

A. Murgu, O.Yu. Gorokhov, I.V. Semoushin, "Input-Output Statistical Inference for Switching Processes" (PDF, 180.4 KB), in Scientific Advances, Vol. 2, No. 9, Ulyanovsk State University, Russia, 2000, pp. 101-113.