Network-Initiated Handoff | WiMAX HANDOFF CONTROL

The network can initiate handover depending on its current status. Such a decision can be made after the evaluation of the payload of different BSs or the data throughput at the reference points. In profile A, ASN GWs may initiate handover of MSs under their control. The network-initiated handoff procedure depicted by Figure 1 begins by a prehandover operation, during which the serving ASN GW collects status information from the BSs and the MSs to decide whether a network-initiated handover is required. If it is the case, the ASN GW sends a HO_Directive message to the serving BS to order it to handoff some MSs to other BSs while providing it with a list of recommended BSs and starting a timer. 

The ASN GW may also specify how many payload should be migrated to other BSs to achieve load balancing as it may indicate the list of the recommended MSs that need to be handed over. The serving BS should respond by a HO_Directive_Rsp message to make the ASN GW stop the timer. The serving BS selects some candidate MSs based on the information maintained by it and the list given by the HO_Directive message and then it may order some candidates to achieve scanning to get their neighbors’ information. The serving BS will then select some suitable MSs for handover and send separately a HO_Req message relative to each MS to the serving ASN-GW. The following procedure is the same as the process of MS-initiated handoff described earlier. When the process of handover preparation is finalized by the network, the serving BS will send a MOB-BSHO_REQ message to each MS to order it to hand over to the target BS.

Figure 1: The preparation phase of a network-initiated handoff.

Base Station Initiated Handoff | WiMAX HANDOFF CONTROL

The serving BS may decide to no longer manage an MS and initiate handoff for it. This occurs generally when the serving BS can no longer provide the required QoS or when it detects that the MS is moving out of its coverage area. Although the causes of a BS-initiated handoff are similar to the causes of an MS-initiated handoff, it is useful to let the BS decide to centralize the handoff procedure. In fact, the MSs are generally tiny equipments with limited power and computing resources; therefore, it is important to implement the handoff process at the BS level. 

The serving BS continues broadcasting the MOB_NBR-ADV message for the served MSs, but it orders the MS that needs to perform handoff via a MOB_BSHO-REQ message to start scanning the neighboring BSs. The MOB_BSHO-REQ message transmitted on the basic connection defines a list of recommended target BSs along with service level predictions and channel details. Upon receiving that message, the BS starts the scanning procedure and sends back a MOB_BSHO-RSP message to the serving BS indicating a list of recommended BSs. 

The rest of the handoff process is similar to the MS-initiated handoff case. In fact, the MS waits for the list of the target BSs and then sends a HO-IND message to its serving BS. Upon receiving the fast ranging IE, the MS sends the RNG-REQ ranging request message to the target BS to register with it. The BS-initiated handoff process, described earlier, is depicted by the flow chart in Figure 1.

Figure 1: The BS initiated handoff at the MS level.

Mobile-Initiated Handoff | WiMAX HANDOFF CONTROL

An MS may decide to change its serving BS after losing signal quality or after detecting that a higher QoS can be disserved by another BS. In such situations, the MS will initiate the handoff. Nevertheless, IEEE 802.16e specifications did not specify the methodology of deciding whether or not to perform the handoff; they only focused on the mechanisms that should be implemented to collect information about the neighboring BSs for taking the handoff decision. In fact, each BS should transmit on the broadcast connection a mobile neighbor advertisement message MOB_NBR-ADV informing the listening MSs of the characteristics of any neighboring BS. Such a message includes the identifier of each neighboring BS, its frequency, the supported services, and its available radio resources such as its available channels. Upon getting such information, each MS should be able to take the handoff decision in the light of scanning of possible target BSs. The scanning procedure begins when the MS sends its serving BS a MOB_SCN-REQ message to inform the serving BS that it wishes to scan the neighboring BSs. The message indicates a length of time in frames for this interval and the type of association that will be used for scanning. The aim of association is to enable the MS acquiring and recording ranging parameters and service availability information to select the proper BS target. 

The specifications define three levels of possible associations, which are the association without coordination, the association with coordination, and the network-assisted association reporting. With the association without coordination, the target BS has no knowledge of the MS. Association with coordination means that the serving BS will coordinate association with the requested target BS and then respond to the requesting MS. With the network-assisted association reporting, the serving BS coordinates association with the requested target BS, a RNG-RSP message is sent back over the backbone to the serving BS. The latter collects all received RNG-RSP messages from all scanned BSs and then sends them to the MS in the form of a MOB_ASC_REPORT message. 

Next, the serving BS responds with a MOB_SCN-RSP message specifying the length of the approved scan and the association type that will be used. Upon receiving that message, the MS may scan its neighbors by synchronizing with a given BSs DL transmissions and estimating the quality of the physical channel. After performing scanning, the MS sends a MOB_MSHO-REQ message to its serving BS on the basic connection. That message includes a list of BSs recommended by the MS as targets. Upon receiving such message, the serving BS sends HO-prenotification messages to all BSs specified in the MOB_MSHO-REQ message and waits for the corresponding HO-prenotification-response messages to analyze them. Next, the serving BS generates the MOB_BSHO-RSP message indicating a list of target stations and sends it back to the MS on the basic connection. The MS may now perform or cancel the handoff; it informs its serving BS via a MOB_HO-IND message sent on the basic connection. If it performs handoff, the MS will inform its serving BS that it is leaving it while providing the parameters of the target BS and then it registers with the target BS. The target BS may, upon receiving the HO-prenotification message, include a fast ranging information element (IE) that provides the MS with a noncontention-based initial ranging opportunity to minimize the handoff process latency.

HANDOFF SCHEMES

Mobile WiMAX specifications support five types of accesses that reflect five mobility-related usage scenarios. First, fixed broadband access assumes a subscriber being in the same geographic location during the whole duration of access to the network services. Second, nomadic access supports the movement between different cells without managing handoff. Third, portable access provides nomadic access to a portable device with the expectation of a BE handoff. Fourth, simple mobility access supports subscribers moving at speeds up to 60km/hour; it provides service continuity despite mobility and fulfils brief interruptions during handoff. Both portable and simple mobility accesses implement the hard handoff concept. Finally, full mobility access guarantees service continuity at high speeds up to 160km/hour and achieves seamless handoff with less than 50 ms latency and less than 1 percent packets loss ratio. To support full mobility access, IEEE 802.16e specifications define three handoff methods, which are the hard handover (HHO), the fast base station switching (FBSS) handover, and the macrodiversity handover (MDHO). HHO is mandatory while both FBSS and MDHO are optional. Initially, HHO is the only type required to be implemented by certified mobile WiMAX equipments.

1: Hard Handoff

Hard handoff results in a sudden connection transfer from one managing BS to a second one as the MS can communicate with only one BS each time. Therefore, all connections with the serving BSs are broken before a new connection with the target BS is established. In Figure 1, the red thick line at the border of the cells indicates the place where HHO is executed. The threshold level hysteresis is used in practice to avoid the repeated switching of neighbors BS during a movement lengthwise to the cell boundaries. HHO is a less complex handoff type but it induces high latency; therefore, it is used for data as it is not suitable for real-time latency-sensitive applications such as VoIP.

Figure 1: The HHO process.

The handoff decision may be taken by either the BS, the MS, or a third entity in light of the periodic measurements done by the MS. In fact, each MS periodically processes a radio frequency scan during scanning intervals allocated by the serving BS and measures the signal quality of the neighboring BSs. Scanning consists in monitoring each possible frequency until a DL signal is received. The number of scanned frequencies depends on the regulatory-provisioned bandwidth, the physical specification, and the chosen bandwidth per channel, which depends on the physical specification. The MS is allowed to perform the initial ranging process to associate with one or more neighboring BSs. Once the handoff is decided, the MS starts the synchronization with the DL transmission of the target BS; it then performs ranging if this was not done during scanning, and it finally ends the connection with the previous BS. The undelivered MAC protocol data unit (MPDU) is stored at the BS until the timer expires.

2: Macrodiversity Handoff

MDHO is a form of soft handoff as the MS is allowed to maintain a valid connection simultaneously with more than one BS. When the MDHO is supported by both the MS and the BS, a diversity set, also referred to as active set, is maintained. The active set is a list of BSs that may be involved in the handoff procedure. Such BSs list is updated through the exchange of MAC management messages. These messages are sent based on the long-term CINR of BSs, which depends on two threshold values broadcasted in the DCD: the Add Threshold H_Add_Threshold and the Delete Threshold H_Delete_Threshold. A serving BS is dropped from the diversity set when the long-term CINR is less then H_Delete_Threshold. A neighbor BS is added to the diversity set when its long-term CINR is higher than H_Add_Threshold. The MS continuously monitors the BSs in the diversity set and defines an anchor BS among them. The MS synchronizes and registers to the anchor BS and performs ranging while monitoring the DL channel for control information. The MS communicates with anchor BS and active BSs in the diversity set as depicted in Figure 2. Two or more BSs transmit data on the DL so that multiple copies are received by the MS which needs to combine them using any of the well-known diversity-combining techniques.

Figure 2: The MDHO process.

To evaluate the performances, suggest the implementing of a handover delay timer that adds a short delay between the time when the handover conditions are met and the handover initialization is started. They assumed a MDHO handover and a periodic reporting of 1 s and then evaluated the number of handovers between 4 BSs and 42 MSs. The principle was to select the BS with the best signal strength for each MS and then create the diversity set for each MS based on defined thresholds and signal strengths. Three different cases were simulated. In the first case (variety I), the probability of affected CINR value is 0.5 percent (it means that the 99.5 percent of values are unaffected). In the second case (variety II), the probability of affected value is 0.05 percent (99.95 percent of values are unaffected). In the third case (variety III), no values were affected (100 percent of values are unaffected). There has been simulated 30 min. time interval and the values were evaluated with 1 s step for all varieties. Results are depicted in Figure 3. 

Figure 3: Number of initialized handovers in function of HDT duration.

3: Fast Base Station Switching

The fast base station handoff is also a form of soft handoff which is supported by both the BS and the MS. The MS maintains a list of BSs referred to as the active set and then continuously monitors it. As in the case of MDHO, the MS may perform ranging and maintain a valid CID with the BSs of the active set. Nevertheless, the MS is allowed to communicate with only one BS called the anchor BS as depicted in Figure 4. In fact, the MS is registered and synchronized with the anchor BS; both entities exchange UL and DL traffic including management messages. The anchor BS may be changed from frame to frame with respect to the BS selection scheme. In that case, the connection is switched to the new anchor BS without performing explicit handoff signaling as the MS simply reports the ID of the newly selected BS on the CQICH. Note that every frame can be sent via a different BS belonging to the active set.

Figure 4: Fast base station switching.

The anchor BS may be updated by implementing two mechanisms: the handover MAC management method and the fast anchor BS selection mechanism. The first updating mechanism is based on the exchange of five types of MAC management messages while the second updating mechanism transmit anchor BS selection information on the fast feedback channel. The new anchor BS should belong to the current diversity set; its selection is based on the signal strength reported by the MS. Adding BSs to the diversity set and removing others is done on the basis of the comparison of their long-term CINR to H_Add_Threshold and H_Delete_Threshold. Note that FBSS and MDHO have many similarities and achieve better performance compared to HHO. Nevertheless, they are more complex as they require the BSs of the active set or the diversity set to be synchronized, use the same carrier frequency, and share the network entry information.

4: Seamless Handoff

Seamless mobility intends to achieve a seamless handoff between different data networks and access technologies. Examples of such handoff occur between WiMAX and WiFi networks, WiMAX and UMTS networks, WiFi and 3G mobile networks including CDMA 2000 and UMTS, etc. Subscribers, who are becoming more and more demanding, wish to have voice, video, and data connections anywhere and anytime and expect to roam within and between networks at low cost while preserving the initial security level and QoS guarantees. To take up such challenges, the used handsets should integrate multiple radio interfaces and support a wide range of applications while implementing power management. Such mixed-network devices must be able to automatically detect and select the best wireless network. On the other hand, networks should provide high bandwidth while implementing intelligent networking mechanisms including seamless roaming and cross-network identity and authentication. For instance, an efficient and usage-model appropriate means of establishing identity should be provided. When considering seamless mobility between WiFi and WiMAX, it is evident not to initiate handoff when the WiFi coverage is available as WiFi provides high bandwidth and achieves good performance. The received signal strength level, provided by the physical layer and the bandwidth of network layer, may be the guiding parameters that determine when to initialize handover and how to choose the best WiFi access point. Link-layer triggers should be defined to help IP handoff preparation and execution while cross-layer information exchange should fasten the handoff process. To guarantee an interrupted user connection during handoff between different networks, IEEE proposes a new standard referred to as IEEE 802.21.

WiMAX HANDOFF

IEEE 802.16e standard defined the required procedures and functions that should be implemented at the physical and MAC layers to perform handoff. The mobile WiMAX version inherits from the IEEE 802.16e standard, but it also defines the protocols that should be implemented at the higher layers to support intra-/inter-ASN handoff, roaming, seamless handoff, and micro-/macro-mobility.

1: SUPPORTED ARCHITECTURE

As described earlier, an ASN includes at least one ASN GW responsible for communicating with the CSN and a BS managing the connections to the MSs in its coverage. An ASN GW may be associated with one or more BSs while a BS can be managed by one or more ASN GWs so that multiple vendors can simultaneously interoperate within the same ASN. The BS may be a serving BS or a target BS depending on its task during the handoff process. In fact, the serving BS is the BS related to the MS before handoff while the target BS is the BS associated with the MS after handoff. On the other hand, we distinguish the serving ASN GW, the target ASN GW, and the anchor ASN GW. The serving ASN GW is the ASN GW corresponding to the serving BS; the target ASN GW is the ASN GW connected to the target BS while the anchor ASN GW is the ASN GW receiving the CSN data addressed to the MS and relaying them to the serving ASN GW. Thanks to the anchoring ASN GW, the MSs mobility is transparent to the CSN that does not need to know which ASN GW is managing the BS that is serving the MS. Therefore, the anchoring function prevents the CSN from frequently changing IP addresses. If the serving ASN GW is directly receiving data from the CSN, it is also considered as the anchor ASN GW. Nevertheless, the anchor ASN GW does not need to be a serving ASN GW or a target ASN GW. The intra-ASN handoff is processed between BSs within the same ASN; it does not induce important delay and minimizes data loss. Besides, intra-ASN handoff does not result in a change of the MSs IP address because the mobility is transparent to the outside of the ASN. Contrarily, an inter-ASN handoff is processed between BSs belonging to different ASNs and involves ASN GWs associated with separate ASNs. These ASN GWs need to coordinate their actions by adopting either anchoring or reanchoring to make the handoff smooth to the MS.

2: FUNCTIONAL DECOMPOSITION

The ASN-anchored mobility management is defined as mobility of an MS not involving a change in the CoA; it applies to mobility in networks not based on MIP. The specifications identify three functions responsible for the handoff, the MS context, and the data delivery control. More specifically, the Handoff function, which is implemented on the serving, the relaying, and the target peers, manages the signaling messages exchange and takes decisions associated with the handover. Figure 1 illustrates a possible handoff scenario. First, the serving-handoff function sends a handoff request (HO_Req) and waits for the corresponding reply. That HO_Req should include at least the MS_ID identifying the MS that requests the handoff, the list of the candidate target BS identifiers (IDs), possibly the MS/Session information content, and the first requested bicast SDU sequence number. The relaying handoff function relays the HO_Req to multiple target handoff functions, which are in charge of analyzing the request, formulating, and sending the correspondent handoff responses (HO_Rsp). The HO_Rsp primitive includes at least the MS_ID and the list of the recommended target BS IDs; it may also carry other optional information. The received responses are forwarded by the handoff-relaying function to the serving-handoff function. The latter should send back a handoff confirmation (HO_Cnf) to the chosen target stating the final handoff action that may either be an initiation, a cancellation, or a handoff rejection. The HO_Cnf should at least indicate the MS_ID, the DL ARQ synchronization information per service flow describing the context necessary to restore communication from the point it has been interrupted and the UL ARQ synchronization information per service flow describing the context necessary to restore communication from the point it has been interrupted.

Figure 1: Handoff function network transaction.

On the other hand, the context function manages the MS context and related information while handling their exchange in the backbone to set up any state or retrieve any state in network elements. For instance, the MSs context in the context function associated with the serving/anchor handoff function needs to be updated. More specifically, the MSs context in the context function associated with the serving handoff-function will be transferred to the context function associated with the target handoff function. Context information transfer may be triggered to populate a new MSs security context at a target BS, inform the network of an MSs initial network entry, or inform the network of the MSs idle mode behavior. The specifications identify relaying context functions, context functions acting as context servers, and context functions acting as context clients. The relaying context function mediates information delivery between context-client and context-server functions. The context-server function stores the most updated session context information for the MS while the context-client function, which is associated with the functional entity having the 802.16 physical link, retrieves session context information stored at the context-server level during handoff processes.

Finally, the data path (DP) function, also referred to as the bearer function, establishes the routes and manages the current data packets transmission between two functional entities. More specifically, the DP function controls the setup of the bearer plane between two BSs, two gateways, or a gateway and a BS; it may implement the setup of tunnels and support multicast and broadcast. The specifications distinguish four DP functions with respect to their roles in the handoff process. First, the anchor DP function anchors the data path associated with the MS across handoffs by forwarding the received data packets toward the serving DP function; it may buffer some of the packets and maintain some state information regarding bearer for the MS during handoffs. Second, the serving DP function is implemented at the end of the DP and associated with the serving PHY(physical)/MAC function (e.g., the serving BS) to handle the transmission of all data packets destined to the MS. Third, the target DP function is associated with a target BS that has been selected as the target of the handoff; it communicates with the anchor DP function to establish the DP that will replace the current path after the termination of the handoff. If the handoff succeeds, the target DP function becomes the serving DP function. Fourth, the relaying DP function mediates message exchange between serving, target, and anchor DP functions.