1. User-Centric Management of Wireless LANs Group 9 ZhengDe Li 3344514 XinTong Zhang 3379583 DanPing Tong 3401729 Nashid Farhad 3423251 LiKai Feng 3427408

2.  Summary  User–centric vs. AP-centric network management framework  Theoretical analysis & simulation  FCAPS relationships  Conclusion  References User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Outline 2

3.  ↑ Deployment density of WLANs  ↑ # of APs deployed around user  Challenge of high-density WLANs: Effective management of APs  Optimize users’ throughput  Throughput depends on: -Load of AP -Channel conditions to AP -Load of other devices within its sensing range  Challenge of effective network management: - AP association  balancing load of each AP -Channel selection  balancing load in each channel  throughput ↑ Summary User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Fig. 1. Illustration of different sensing ranges of users and APs 3

4. Traditional NM in WLANs User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian  Basic Management Entities SME MLME PLME SME MLME PLME  Related Management Activities Parameter Configuration Channel Selection Power Configuration User Information Management AP Association Priority Management Network Monitoring Downlink Monitoring Uplink Monitoring Fig. 2. Network management flow in traditional WLANs. 4

5.  AP-centric NM framework （ Traditional ） - Exclusively AP one-side control the transmission & None that relate to users - NM procedures: simple + workable - Helpless to enhance the throughput - Not suitable  Differentiated access priorities  Support different QoS  User-centric NM framework (Proposed) -The WLAN management is sharing pipeline btw users & APs - Maximize users’ throughput - Optimize AP aggregate throughput and dynamic APs’ channel selection -Network conditions & access priority  considered by two NM activities -User actively senses the network condition & delivers the sensed results with its access priority to the AP candidates User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian AP-centric V.S. User-centric NM framework 5

6. AP periodically delivers config info to neighboring users Configuration Info: channel, data rate, MAC etc. AP Sends Beacon User actively collects network data (user’s network snapshot) Sends to all candidate APs in its receiving range User Sends Channel Condition to AP AP calculates potential throughput and return to user Association Response Frame AP computes potential throughput User selects and sends a new request to the AP with the max potential throughput &AP reply and store user’s info A new association established User joins the AP with the most potential throughput AP Periodically updates snapshots and performance of associated users A new association triggered AP optimizes channel selection according to user data User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian How User-centric NM framework Works? 6

7. Simulations  Software used: Network Simulator version 2 (NS2)  Simulation Parameters  Simulation environment Indoor environment 7 User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian

8.  Throughput ↓ as # of users ↑.  Throughput of the user with high access priority is greater than that with low access priority. User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Effect of Access Priority 8

9.  AP-associated based on the user-centric NM framework always outperforms other frameworks  25%-90% throughput gain User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Throughput Comparison 9

10.  User-centric management provides highest throughput – max 69.3%  As the APs get saturated with load, the throughput becomes similar for all the system. User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Throughput Comparison – Channel Selection 10

11.  C : {0,1} = mapping factor for the lower bound  C R > 1 = constant for leveraging improvement degree  As the lower bound increases throughput decreases  As the improvement degree increases throughput increases  But, if both lower bound and improvement degree increases throughput decreases User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Impact of Dual-Threshold on Throughput 11

12.  Fault management------detect & correct associations  e.g. AP association triggered by user  Configuration management------sharing pipeline info  beacon frame by AP periodically  snapshot by user  Accounting management  Performance management—measure & record  AP calculates the potential throughput to user  user records collected info by snapshot  Increased NM traffic  Security management   Security Vulnerability – can be prone to attack via fake user snapshots. User- centric under FCAPS NM User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian 12

13. Traditional WLAN management system × can’t optimize users’ throughput ∵ ineffectively utilize the network conditions The user-centric network management framework  takes maximizing the throughput of users as the fundamental principle  provides an effective infor. sharing pipeline btw users & APs  assist the key network management activities Theoretical analysis  user-centric NM framework provide complete information  attaining the optimization of NM activities The simulation results throughputs can be significantly ↑ in NM activities over user-centric framework User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian Conclusion 13

14. Questions? User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian 14

15. [1] Y. Zhu, Q. Ma, C. Bisdikian, C. Ying, “A user-centric network management framework for high-density wireless LANs," in Proc. 2009 IFIP/IEEE International Symp. Integrated Netw. Management, pp. 218-225, 2008. [2] V. A. Siris and D. Evaggelatou, “Access point selection for improving throughput fairness in wireless LANs," in Proc. 10th IFIP/IEEE Int. Symp. Integrated Netw. Management, 2007, pp. 469-477. [3] Y. Zhu, Z. Niu, Q. Zhang, B. Tan, Z. Zhou, and J. Zhu, “A multi-AP architecture for high- density WLANs: protocol design and experimental evaluation," in Proc. IEEE SECON, 2008, pp [4] Y. Fukuda, T. Abe, and Y. Oie, “Decentralized access point selection architecture for wireless LANs," in Proc. 3rd Wireless Telecommun. Symp., 2004, pp. 137-145.28-36. [5] S. Vasudevan, K. Papagiannaki, C. Diot, J. Kurose, and D. Towsley, “Facilitating access point selection in IEEE 802.11 wireless networks,“ in Proc. 5th ACM SIGCOMM Conf. Internet Measurement, 2005, pp. 293-298. [6] J. Geier (2002, Feb. 11), “Assigning 802.11b access point channels," Wi- Fi Planet. Available: http://www.wi-fiplanet.com/tutorials/article.php 972261.http://www.wi-fiplanet.com/tutorials/article.php 972261 [7] Y. Hua, Y. Zhu, and Z. Niu, “A channel assignment scheme in high density WLANs to mitigate pesudo capture effect," in Proc. IEEE GLOBECOM, 2007, pp. 4796-4800. User-Centric Management of Wireless LANs - Y. Zhu, Q. Ma, C. Bisdikian References 15