The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
La virtualisation des fonctions réseau (NFV) permet aux opérateurs de réseaux de fournir de manière flexible diverses fonctions virtualisées pour des services tels que l'Internet des objets (IoT) et les applications mobiles. Pour répondre aux multiples exigences de qualité de service (QoS) dans des environnements réseau variables dans le temps, les fournisseurs d'infrastructure doivent ajuster dynamiquement la quantité de ressources de calcul, telles que le processeur, affectées aux fonctions de réseau virtuel (VNF). Pour assurer un contrôle agile des ressources et une adaptabilité, la prévision de la charge du serveur virtuel via des technologies d'apprentissage automatique constitue une approche efficace du contrôle proactif des systèmes réseau. Dans cet article, nous proposons un mécanisme d'ajustement pour les régresseurs basé sur l'oubli et l'ensemble dynamique exécuté dans un temps plus court que celui de nos travaux précédents. Le cadre comprend une méthode de réduction des données d'entraînement basée sur une régression de modèle clairsemé. En dressant une courte liste de données d'entraînement dérivées du modèle de régression clairsemée, le temps de réapprentissage peut être réduit à environ 57 % sans dégrader la précision de l'approvisionnement.
Takahiro HIRAYAMA
National Institute of Information and Communications Technology (NICT)
Takaya MIYAZAWA
National Institute of Information and Communications Technology (NICT)
Masahiro JIBIKI
National Institute of Information and Communications Technology (NICT)
Ved P. KAFLE
National Institute of Information and Communications Technology (NICT)
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Takahiro HIRAYAMA, Takaya MIYAZAWA, Masahiro JIBIKI, Ved P. KAFLE, "Sparse Regression Model-Based Relearning Architecture for Shortening Learning Time in Traffic Prediction" in IEICE TRANSACTIONS on Information,
vol. E104-D, no. 5, pp. 606-616, May 2021, doi: 10.1587/transinf.2020NTP0010.
Abstract: Network function virtualization (NFV) enables network operators to flexibly provide diverse virtualized functions for services such as Internet of things (IoT) and mobile applications. To meet multiple quality of service (QoS) requirements against time-varying network environments, infrastructure providers must dynamically adjust the amount of computational resources, such as CPU, assigned to virtual network functions (VNFs). To provide agile resource control and adaptiveness, predicting the virtual server load via machine learning technologies is an effective approach to the proactive control of network systems. In this paper, we propose an adjustment mechanism for regressors based on forgetting and dynamic ensemble executed in a shorter time than that of our previous work. The framework includes a reducing training data method based on sparse model regression. By making a short list of training data derived from the sparse regression model, the relearning time can be reduced to about 57% without degrading provisioning accuracy.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2020NTP0010/_p
Copier
@ARTICLE{e104-d_5_606,
author={Takahiro HIRAYAMA, Takaya MIYAZAWA, Masahiro JIBIKI, Ved P. KAFLE, },
journal={IEICE TRANSACTIONS on Information},
title={Sparse Regression Model-Based Relearning Architecture for Shortening Learning Time in Traffic Prediction},
year={2021},
volume={E104-D},
number={5},
pages={606-616},
abstract={Network function virtualization (NFV) enables network operators to flexibly provide diverse virtualized functions for services such as Internet of things (IoT) and mobile applications. To meet multiple quality of service (QoS) requirements against time-varying network environments, infrastructure providers must dynamically adjust the amount of computational resources, such as CPU, assigned to virtual network functions (VNFs). To provide agile resource control and adaptiveness, predicting the virtual server load via machine learning technologies is an effective approach to the proactive control of network systems. In this paper, we propose an adjustment mechanism for regressors based on forgetting and dynamic ensemble executed in a shorter time than that of our previous work. The framework includes a reducing training data method based on sparse model regression. By making a short list of training data derived from the sparse regression model, the relearning time can be reduced to about 57% without degrading provisioning accuracy.},
keywords={},
doi={10.1587/transinf.2020NTP0010},
ISSN={1745-1361},
month={May},}
Copier
TY - JOUR
TI - Sparse Regression Model-Based Relearning Architecture for Shortening Learning Time in Traffic Prediction
T2 - IEICE TRANSACTIONS on Information
SP - 606
EP - 616
AU - Takahiro HIRAYAMA
AU - Takaya MIYAZAWA
AU - Masahiro JIBIKI
AU - Ved P. KAFLE
PY - 2021
DO - 10.1587/transinf.2020NTP0010
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E104-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2021
AB - Network function virtualization (NFV) enables network operators to flexibly provide diverse virtualized functions for services such as Internet of things (IoT) and mobile applications. To meet multiple quality of service (QoS) requirements against time-varying network environments, infrastructure providers must dynamically adjust the amount of computational resources, such as CPU, assigned to virtual network functions (VNFs). To provide agile resource control and adaptiveness, predicting the virtual server load via machine learning technologies is an effective approach to the proactive control of network systems. In this paper, we propose an adjustment mechanism for regressors based on forgetting and dynamic ensemble executed in a shorter time than that of our previous work. The framework includes a reducing training data method based on sparse model regression. By making a short list of training data derived from the sparse regression model, the relearning time can be reduced to about 57% without degrading provisioning accuracy.
ER -