Performance Analysis of Fast Handover Schemes in IEEE

Performance Analysis of Fast Handover Schemes in IEEE 802. 16 e Broadband Wireless Networks Liping WANG The Graduate University for Advanced Studies, Japan Fuqiang LIU School of Electronics and Information Engineering, Tongji University, P. R. China Yusheng JI National Institute of Informatics (NII) and The Graduate University for Advanced Studies, Japan The Graduate University for Advanced Studies

Outlines n n n Background Redundancies in IEEE 802. 16 e MAC layer handover process Fast Handover schemes q q n n Single neighbor BS scanning Fast ranging and pre-registration Performance analysis Conclusion 12/30/2021 The Graduate University for Advanced Studies 2

IEEE 802. 16 Broadband Wireless Networks 12/30/2021 The Graduate University for Advanced Studies 3

Handover in IEEE 802. 16 e n Handover types supported by IEEE 802. 16 e q q q n HHO/HO (Hard Handover) FBSS (Fast BS Switching) MDHO (Macro Diversity Handover) Up till now, only HHO and FBSS are adequately defined in the standard for practical use 12/30/2021 The Graduate University for Advanced Studies 4

The MAC Layer HO of IEEE 802. 16 e n Stages of HO process q q Network topology acquisition n Network topology advertisement n MS scanning or association of neighbor BSs HO execution n n 12/30/2021 Cell reselection HO Decision and Initiation Connection release Target network re-entry The Graduate University for Advanced Studies 5

Redundancies in 802. 16 e MAC Layer HO 1 n In network topology acquisition stage q q q There is only one BS can be selected as target BS for handover actually The results gained from scan may be invalid because of the changes of neighbor BSs’ channel quality The throughput will decrease if the scan/association process occupies too many resources 12/30/2021 The Graduate University for Advanced Studies 6

Redundancies in 802. 16 e MAC Layer HO 2 n In HO execution stage q q Interruption of data transmission only happens during the network re-entering process The time of interruption is called handover delay Service will be terminated if the handover delay is too long for applications Using fast ranging and pre-registration schemes can reduce the handover delay, hence increase the successful probability of handovers 12/30/2021 The Graduate University for Advanced Studies 7

Fast Handover Schemes -1 n Single neighbor BS scanning q q The main waste of resources is caused by redundant scan and association process of neighbor BSs Target BS estimation algorithm, in which MS only scans or associates the neighbor BS with the best CINR (Carrier to Interference-plus-Noise Ratio), is proposed 12/30/2021 The Graduate University for Advanced Studies 8

Fast Handover Schemes -2 n Fast Ranging q q Target BS uses a Fast_Ranging_IE to grant an dedicated uplink ranging opportunity to MS in its broadcasting UL-MAP message MS does not need to do contention-based ranging 12/30/2021 The Graduate University for Advanced Studies 9

Fast Handover Schemes -3 n Pre-registration q Target BS obtains the service flow and authentication information of this MS through backbone networks before handover q Target BS still has to send CID updating message to the MS by using REQ-RSP message q The average time of pre-registration process is half of the average time for the normal registration process 12/30/2021 The Graduate University for Advanced Studies 10

Performance Analysis – Assumptions (1) n Scan time and handover delay are studied in the IEEE 802. 16 e TDD model with frame length assumed to be 5 ms q q q Rload is defined as the cell load ratio TSync , the average time required for downlink synchronization, is assumed to be 10 ms TCont_rng, the average time required for contention-based ranging, is assumed to be directly proportional to Rload, and calculated as equation (1) n n 12/30/2021 “ ” mean the smallest integer not less than the number inside The Graduate University for Advanced Studies 11

Performance Analysis – Assumptions (2) q TRng, the average time required for fast ranging, is also assumed to be directly proportional to Rload and can be calculated as equation (2) n q q TAuth, the average time required for authorization, is assumed to be 150 ms TReg, the average time required for registration is assumed to be 35 ms 12/30/2021 The Graduate University for Advanced Studies 12

Performance Analysis – Scan Time (1) n n Scan time of different scanning types is studied Definations: q q q n: the number of neighbor BSs that chosen for scan or association TScn_0: the average time requires for ‘Scanning Type’ = 0, i. e. scan without association TScn_1: the average time requires for ‘Scanning Type’ = 1, i. e. scan/association without coordination TScn_2: the average time requires for ‘Scanning Type’ = 2, i. e. association with coordination TScn_3: the average time requires for ‘Scanning Type’ = 3, i. e. network assisted association reporting 12/30/2021 The Graduate University for Advanced Studies 13

Performance Analysis – Scan Time (2) n ‘Scanning Type’ = 0 (scan without association) q q n MS only needs to downlink synchronize with the target BS to get the quality information of its physical channel (3) ‘Scanning Type’ = 1 (scan/association without coordination) q q MS needs to downlink synchronize with target BS and execute a contention-based ranging (4) 12/30/2021 The Graduate University for Advanced Studies 14

Performance Analysis – Scan Time (3) n ‘Scanning Type’ = 2 (association with coordination) q MS needs to downlink synchronize with target BS and execute a fast ranging process (5) q n ‘Scanning Type’ = 3 (network assisted association reporting) q Serving BS packs all RNG-RSP messages from scanned neighbor BSs into a MOB_ASC-REPORT message, so the average time required for ranging is half of that for a normal fast ranging (6) q 12/30/2021 The Graduate University for Advanced Studies 15

Performance Analysis – Scan Time (4) n Figure 1. Scan time vs. cell load ratio (n=1). 12/30/2021 The Graduate University for Advanced Studies 16

Performance Analysis – Handover delay (1) n n Handover delay of different handover schemes is studied Definations q q DHO_P 1: the handover delay of handover policy 1, i. e. handover with contention-based ranging DHO_P 2: the handover delay of handover policy 2, i. e. handover with fast ranging DHO_P 3: the handover delay of handover policy 3, i. e. handover with contention-based ranging and preregistration DHO_P 4: the handover delay of handover policy 4, i. e. handover with fast ranging and pre-registration 12/30/2021 The Graduate University for Advanced Studies 17

Performance Analysis – Handover delay (2) n In handover policy 1, MS must obtain a ranging opportunity through contention, and execute a re-authorization and re-registration process (7) q n In handover policy 2, MS has to execute a reauthorization and re-registration process (8) q 12/30/2021 The Graduate University for Advanced Studies 18

Performance Analysis – Handover delay (3) n In handover policy 3, contention-based ranging and pre-registration are adopted. q q q Although target BS knows the information of MS, it still should send REQ-RSP message including CID updating information to MS. The average time for pre-registration is half of the average time of a complete registration process (9) 12/30/2021 The Graduate University for Advanced Studies 19

Performance Analysis – Handover delay (4) n In handover policy 4, fast ranging and preregistration schemes are adopted (10) q 12/30/2021 The Graduate University for Advanced Studies 20

Performance Analysis – Handover delay (4) n Figure 2. Handover delay vs. cell load ratio. 12/30/2021 The Graduate University for Advanced Studies 21

Conclusion n Several fast handover schemes for IEEE 802. 16 e broadband wireless networks are proposed analyzed: q q In the first scheme, single target BS with best CINR is chosen for scanning, which reduces the scan time in network topology acquisition process especially in association modes Secondly, fast ranging and pre-registration schemes are used to optimize network re-entering process, which reduces handover delay. 12/30/2021 The Graduate University for Advanced Studies 22

Thank you! Q&A 12/30/2021 The Graduate University for Advanced Studies 23