Adaptive filter theory
On the self-similar nature of Ethernet traffic (extended version)
IEEE/ACM Transactions on Networking (TON)
Using adaptive linear prediction to support real-time VBR video under RCBR network service model
IEEE/ACM Transactions on Networking (TON)
Characterizing user behavior and network performance in a public wireless LAN
SIGMETRICS '02 Proceedings of the 2002 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
The Limitations of Artificial Neural Networks for Traffic Prediction
ISCC '98 Proceedings of the Third IEEE Symposium on Computers & Communications
The changing usage of a mature campus-wide wireless network
Proceedings of the 10th annual international conference on Mobile computing and networking
Characterizing mobility and network usage in a corporate wireless local-area network
Proceedings of the 1st international conference on Mobile systems, applications and services
FGCN '07 Proceedings of the Future Generation Communication and Networking - Volume 01
Traffic prediction from wireless environment sensing
WCNC'09 Proceedings of the 2009 IEEE conference on Wireless Communications & Networking Conference
Lightweight proactive queue management
IEEE Transactions on Network and Service Management
Multiresolution FIR neural-network-based learning algorithm applied to network traffic prediction
IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
| Hi-index | 0.00 |
Several studies have analyzed traffic traces collected from real world deployments of wireless networks. The vast majority of these studies have employed descriptive statistics with the aim of obtaining insights into the different aspects of wireless networks. While the contributions of all these studies are valuable, they mainly provide guidelines on design and deployment of wireless networks on a longer term perspective. The ability to predict at shorter timescales such as on the order of a few minutes empowers the network management entity with extra intelligence to optimize network performance taking into account anticipated traffic conditions. This paper proposes a simple traffic prediction mechanism using the Recursive Least Squares algorithm and highlights its applications in proactive network management. The performance of the proposed prediction mechanism is also evaluated using publicly available data set collected from a real world wireless network. Results from this study show that the RLS algorithm is capable of accurately predicting the traffic load and shows good adaptive behaviour. Moreover it is intuitively simple and results in a lightweight implementation.