This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
This chapter provides introduction about my project i.e. QOS in WIMAX, research challenges, aim and objectives, recommended approach and the methodologies used in the project.
In the recent years, many wireless technologies have been developed by several wireless communities to provide wireless communications with in different ranges, such as local Area Network (LAN), Metropolitan Area Network (MAN) and Wide Area Network (WAN). However, these wireless technologies have a limited transmission rate and range. In order to meet the future requirements, Worldwide Interoperability for Microwave Access (WIMAX) is an extended version of MAN due to its maximum transmission rate and its range and it does it with wide range of quality of service (QOS) capabilities. In a wireless communication quality of service plays a vital role to achieve the user requirements and providers.
WIMAX is related with the IEEE802.16 standard, which determines five types of user service. These entire standards define a wireless connection MAC protocol with a mechanism for QOS support. Depending on the vendor requirements, scheduling algorithms are designed.
Ad-hoc scheduling scheme, which supports five types of service flow which are specified in IEEE 802.16 and makes the bandwidth allocation fair and dynamic, and utilize the network resources in an efficient manner. Initially the scheduler uses an ad-hoc scheduling scheme, where all user request are separately scheduled using different scheduling algorithms. In second level, it uses Dynamically Allocating Priority Queue (DAPQ) to allocate bandwidth dynamically to each user class based on the packet history.
Mobility is one of the important features of WIMAX. Basically, handoff or handover process is needed to provide continues service to user. It is a process of changing the channel related with the current connection while a service in a progress. Initially, it is activated either by crossing a station boundary or by week signal strength in the current channel. Designing of poor handoff scheme goes in to heavy signalling traffic and decrease in quality of service (QOS).
"How to schedule the data packets efficiently by sharing the resources equally among them to achieve better Quality of Service (QOS)"
"How to minimize the handoff response time and waste of wireless network resources during the handoff process"
Scheduling the packets in an efficient manner by sharing the resources among them equally and provide better quality of service (QOS).
To evaluate the different scheduling schemes and the various QOS parameters associated with them
To utilize the network resources efficiently among all service classes on separate scheduling algorithms
To minimize the starvation problem by using Dynamic Allocation Priority Queue
Minimize the handoff latency during handoff process to enhance the mobility feature in WIMAX.
Analyze the behaviour of handoff process system in WIMAX
To investigate various handover processes and issues associated with them
To come out with a better process that can improve the Quality of service (QOS)
Approach and methodology:
Approach to achieve aim1:
Initially, I reviewed some background concepts related to scheduling process and analyzed the Quality of Service requirements such as, packet loss and delay, latency, jitter etc. I investigated various service classes, which are specified in IEEE802.16 standard and some scheduling algorithms such as, Round Robin (RR), Weighted fair Queue (WFQ) etc.
Ad-Hoc scheduling scheme, which carry-out the entire scheduling process is in two stages. In each stage scheduler perform its work by utilizing available network resources efficiently among all services.
Approach to achieve aim2:
I analyzed the behaviour of the general handoff process and studied how the process is initiated and executed. I learned certain things such as how the process is triggered, what kind of parameters are considered during the process, how the stations are communicate with each other, how the data transmission is will be done and etc.
Handoff process method using Mobility Patter is designed in such a way that method estimates the target BS scanning time and predicts it an efficient manner by minimizing the scanning time. Moreover, it avoids unnecessary scanning and reduces the waste of network resources and reduces the handoff latency.
The reminder of the entire project is organized as follow. Chapter 2 provides the background related to scheduling schemes and evolution of handoff process and problems identification. Chapter 3 describes the research done by others regarding scheduling and handover process. My own contribution such as methodologies used to solve the problems is explained in chapter 4. Chapter 5 provides future work, recommendations etc. Results and conclusion is given in chapter 6.
Chapter 2: Overview of related concepts and problem identification:
Scheduling algorithms in WIMAX:
Proportional fair scheduling scheme:
Initially, proportional fair scheduling was developed by QUALCOMM Company, which is recognized in the IS-856 standard. It s designed especially for downlink traffic scheduling to avoid conflicts between the users with high data rate and fairness. It does this by observing the trade-off between systems throughput and starvation of lowest priority users. Proportionate fair scheduling works based on following priority function.
Pi (t) =fi (t)/Fi (t)
fi (t) is a current data rate.
Fi (t) is a smoothing average of the service rate.
t is a time slot
Queues are served first, which have a highest value of Pi (t) at time slot t and the following function is used to update the average throughput of the queue.
Fi (t+1) = (1-1/Tc) Fi (t) + (1+Tc) fi (t)
Condition, if connection "i" is served at time slot t.
Fi (t+1) = (1-1/Tc) Fi (t)
Condition, if connection "i" is not served at time slot t
Tc is a time -constant, which is used to find out moving average time. Short-time starvation on the queue can be controlled by adjusting the Tc parameter. To improve its channel capacity, the scheduler waits for long-time for particular connection. When a large number of users exist in a queue, the scheduler schedules the connection having with high data rate at some particular time slot. Proportionate fair scheduling is easy to implement and efficient. However, it fails to deliver QOS requirements for a connection, which are mentioned for various classes such as delay, jitter, throughput etc.
Cross- layer scheduling scheme:
Many scheduling scheme are proposed but most of them depends on different algorithm to address various users' service classes to deliver QOS requirements. Some of them are used for resource allocation and grants a reserve quality of service per connection. To calculate priority of each connection, cross-layer scheduling considers various metrics of different classes.
UGS: Service rate and packet error rate, because it demands throughput guarantee, packet loss i.e. VoIP.
RTPS: Maximum delay and packet error rate, after which packet is waste i.e. video streaming.
NRTPS: Minimum reservation rate and packet error rate i.e. file transfer protocol (FTP).
BE: No need any guarantee. However, packet rate should maintain i.e. e-mail.
Channel condition priority of a particular connection is calculated based on the parameters. Initially some fixed number of timeslots is allocated separately for UGS service. After that according to the priority, remaining time RTPS, NRTPS and BE. Earliest deadline first (EDF) algorithm is used to manage queue for real-time polling service. PFS scheduling is used to manage queue for non-real time polling service. The traffic in queue for BE is handled based on a best rate discipline. The implementation of the algorithm is based on the following priority function.
Pi (t) =Vclass*fi (t) (Fi (t)*Fi (t)) if Fi (t)>1
Pi (t) =Vclass if Fi (t) <=1
According to the priority of a particular class, Vclass variable can be set. Normally coefficients are selected such as Vrtps>Vnrtps>Vbe. The parameter such as ri (t)/Ri (t) indicates normalized channel quality of a particular connection. Because Ri (t) and ri (t) indicates maximum data and current data rate.
Fi (t) =Ki (t)-Si (t) for real-time connection, where Ki is a delay requirement and Si (t) is maximum delay requirement.
Fi (t) =ni/N (t) for non-real time connection, where N (t) is a data rate, ni is a average data rate.
By using above equation priority is calculated only based on normalized channel quality and there are no specific QOS requirements.
TCP-Aware uplink scheduling scheme:
This is a simple algorithm, which can able to handle BE service class. This class don't have any QOS requirements and there is no guarantee to use bandwidth request mechanism for BE class. It cannot allocate the bandwidth equally among all connection in BE class because all connections can't utilize whole bandwidth allocated to them. Depending on the sending rate, algorithm calculates the bandwidth of that particular connection. Also sending rate can change dynamically. To allocate the bandwidth properly, TCP-aware uplink scheduling scheme works as following.
Comparison of QOS scheduling schemes in WIMAX
No. of classes
Ad-Hoc scheduling scheme
Proportionate fair scheduling scheme
Cross-layer scheduling scheme
TCP-Aware uplink scheduling scheme
WIMAX QOS service classes:
WIMAX is related with the IEEE802.16, which defines five QOS classes.
Unsolicited Grant Scheme (UGS).
Extended Real- time Polling Service (ERTPS).
Real-time Polling Service (RTPS).
Non-real time Polling Service (NRTPS).
Best Effort (BE).
Figure: Application classes of WIMAX
Each and every class has its own QOS requirements such as delay/jitter constraints and minimum throughput requirements.
UGS: This service class is designed to allow fixed periodic bandwidth allocation. There is no need to send other requests once the connection is established. For achieving constant bit rate (CBT) in real-time network traffic, UGS is used. Regular polling is must because of vary in bandwidth requirements. The following parameters are associated with it like
Maximum sustained rate (MST).
ERTPS: Extended real-time polling service is designed to support voice over internet protocol (VOIP) with silence suppression. This service is similar to UGS, in that maximum sustained rate is allocated in active mode by base station (BS). However, during silent period no bandwidth is allocated by ertps. Because of the similarity between UGS and ERTPS, QOS parameters are same as these in UGS.
RTPS: This service is designed to support variable bit rate (VBR) real time traffic. To determine what allocation need to made, BS need a regular polling each MS. Bandwidth requirements are vary in RTPS. The QOS parameters of this service are maximum sustained traffic rate and minimum reserved traffic rate.
NRTPS: This service provides variable bit rate (VBR) traffic with no delay guarantee. This service supports file transfer protocol (FTP) traffic.
BE: most of the data traffic falls in to this class. This service follows the above rule, if it wants to grant a bandwidth to the MS.
Comparison of WIMAX QOS services classes:
QOS service classes
Meet real time service guarantee, there is no overhead.
Bandwidth is not utilize efficiently because of allocations are granted regardless of present need.
Data overhead efficiency and optimal latency.
Polling mechanism is not required.
Data transport efficiency.
Overhead is required and delay guarantee.
Able to provide good service for non-real traffic because of its minimum reserved rate.
It provide full service for BE traffic
No service guarantee.
Quality of Service challenges:
Scheduling process is a crucial factor in communication networks. To satisfy the basic Quality of Service requirements in the wireless environment, designing is a challenging component. Let's consider criteria while designing an algorithm such as,
Energy consumption and power control
The resources are limited in wireless networks. So by using the available resources, system should transmit the data without any packet lost and overhead. The overhead includes MAC overhead, burst overhead and fragmentation. Bandwidth request is represent in number of bytes. There arise a question in optimizing the frame and how to share SDUs in to MPDUs.
The request for bandwidth is calculated in bytes that mean that would not interpret straight ahead to number of slots. This is because a slot can contain any number of bytes reckoning on the modulation technique used for that. Additionally, packet percentage is also a considerable issue. The channel state condition data and the leading with error rate are useful in making decisions for the modulation and coding schemes for each and every user.
Apart from checking QOS requirements, the remaining resources should be apportioned fairly. Time to meet fairness is significant because fairness can be determined as short or long term.
Power control and energy consumption
Maximum power permissible should be consider by the scheduler. Base station should accept signal to noise ratio (SNR) and bit error rate (BER) to transmit data. Depending on the subscriber station location, scheduler calculates the desirable power. MS scheduler should use maximum transmission power since there is a limited power for mobile users.
Base station need to handle many connections at a time. The scheduling algorithm should be efficient, simple and it should use minimum resources.
In order to meet the future requirements an algorithm has to operate efficiently as the connections increases.
Introduction to mobility management:
To realize that the important of wireless connectivity in future wireless broadband network such as WIMAX, a good mobility management is crucial. It consist handoff management and location management. With the location management, location of the mobile can be changed when it is in ideal mode and also it allows the wireless network to locate the roaming MS and establishes the connection. Handoff management is a process of switching from one MS to another in order to maintain the service continuity. In WIMAX handoff process the following components are exists.
Mobile station (MS)
Service Access Point (SAP)
Target Access Point (TAP)
Serving Gateway (SGW)
Target Gateway (TGW)
Home Agent (HA)
Access point plays an important role as attachment point for mobile station and it can able to terminate link layer and air link MAC protocols. The GW defined as a point for IP forwarding and IP layer attachment point for the MS. The layer-2 and layer-3 handoff can be defined as.
Layer-2 handoff is also called as micro mobility. It cab able to modify the attachment point in air link by keeping the internet protocol attachment point unmodified. The entire process can detect the modified signal strength and pass the connection to the SAP. In addition, it establishes a new connection to the target access point. L2HO has low packet loss and minimum handoff latency because of its transparent nature to the upper layer protocols.
Layer-3 is also called as macro mobility and it can modify the internet protocol attachment point of a users. During this process it can update the home agent by using MIP4/MIP6 protocol with care-of-address of the mobile station. Moreover, when it movies from serving gateway to target gateway the mobile station should registered and authenticated with the home agent.
Handoff Schemes in WIMAX:
Layer-2 handoff process consist some similar attributes in all modern communication systems. In 3GPP, 3GPP2 and WIMAX, it uses common methods to achieve L2Ho. Basically handoff process is divided in to two categories. They are
MAC layer handoff
Network layer handoff
In MAC layer handoff, the subscriber station selects the new station based on the on characteristics such as
Quality of service
Some other matrices
In the network layer handoff, the subscriber station finds and allow new routing path in order to maintain the wireless connection with the corresponding nodes. There are three kings of MAC layer handoff schemes are specified in IEEE802.16 standard. They are
Hard Handoff (HHO)
Macro Diversity Handover (MDHO)
Fast Base Station Switching (FBSS)
Figure: Handoff Schemes
In HHO process, if current SS want to establish a connection with a new base station first the SS need to disconnect the old connection with base station. During this process, the wireless connection is interrupted temporarily because the processing time of HHO is crucial for real-time applications, such as voice over internet protocol (VOIP), electronic conference (E-conference. The support of HHO is compulsory in WIMAX networks.
In MDHO process both subscriber station and base station maintain some set of base stations. The subscriber station have ability to communicate with all existing base station in diversity set at the same time maintain a wireless communication during the handoff process. MDHO has ability to reduce handoff latency because of its optimized handoff scheme nature. Both MDHO and FBSS need extra physical layer support.
Handoff process in a Traditional Architecture:
Traditional network architectures are designed in such a way that it can easily accommodate large number user in outdoor cellular environment. This architecture can also easily deploy for 3G operators. Architecture accommodates large number of gateway and high capacity of carrier grade control nodes in the central office. Large number of wireless radios is interconnected via point-to-point T1/E1 links. There is no direct connection exists between cell-sites and radio. The entire traffic and all signalling must go through central office. In the access network, the radio network controller (RNC) controls node B. GPRS core network is formed by gateway GPRS support node (GGSN) and Serving GPRS support node (SGSN), providing the connection to internet protocol network.
Figure: Reference model of UTMS network
In this chapter I explained the concepts related to Quality of Service, scheduling and handoff process. Such as Quality of Service classes in WIMAX and their requirements, QOS challenges, classification of various scheduling algorithms and issues, evolution of handoff mechanism and its importance, various handoff models and its types.
Chapter3: Relevant work done by others :
Many people proposed various scheduling algorithms in the wireless community among them Maheshweta serkar and Harpreet sechdeva designed a downlink packet scheduling (DPS). They explained how it is evolved, a brief explanation about DPS. Moreover, they test the performance of the DPS by conducting simulation test. Let see how they explained about DPS.
Downlink packet scheduling (DPS):
IEEE802.16 standard does not determined any sort of scheduling algorithm for transmission of packets in both directions i.e. downlink and uplink. There major objective is to design an efficient scheduling algorithm that can provide a quality of service among various types of traffic flow depending on their QOS requirements. The main aim of the proposed algorithm is to minimize the packet delay and loss for all type of traffic. This can be gained by implementing the algorithm at base station (BS).
The scheduler divides the entire scheduling process in to two stages. In the first stage, the scheduler splits the entire resources in to four unequal parts and allocates that to a particular traffic class. After the allocation process, scheduler schedules the packets which are waiting in a queue at base station (BS). Thus scheduler scheme supports four types of service class each of them associated with its own queue. In second stage, scheduler marks the decision by consider the service policy for each particular queue.
QOS requirement policies:
Based on the QOS requirements it follows following four different policies.
Fixed bandwidth service policy for UGS traffic.
Earliest deadline first-round robin service policy for rtps traffic.
Weighted Round Robin service policy for BE traffic.
The authors use C++ programming language to test the performance of the scheduler.
They suggest that the scheduler process the transmitting packets to the subscriber station not by one subscriber station at a time. However, one type of traffic at a time depending on its priority, this is because of various service classes that are associated with specific requirements. For example, NRTPS and UGS follow strict packet deadlines and packets with delay, but these two services can able to allow packet loss. For NRTPS service would not allow any sort of packet loss at any cost instead it adapts larger delays.
Figure: Under various conditions the average packet percentage that cannot meet delay deadline for various service classes.
The proposed scheduling algorithm scheme designed to overcome problems. DPS follows the following path to transmit the packets, first UGS next RTPS packets, followed by NRTPS and finally BE packets. The scheduler contains a traffic classifier which maps the entire traffic in to four categories as mentioned previously.
Scheduler working principle:
Bandwidth allocator follows a strict priority rules from highest to lowest i.e. UGS, RTPS, NRTPS and BE. One of the main disadvantage of the scheduling algorithm is it follows strict priority discipline due to that policy, the packets with lower priority will be starved. In order to overcome this problem, author introduces a concept called traffic policy module in each subscriber station which forces the connection to stay in its limit. Based on UGS requirements 802.16 allocate fixed bandwidth.
The packets with earliest deadline will be scheduled first in a RR manner because the connections in RTPS are applied earliest deadline first-round robin (EDF-RR). Depending on the packet deadline all of them are served accordingly because of its delay sensitive. Weighted RR (WRR) is applied for NRTPS based on its weights the scheduler schedules the packets in RR prioritizing manner. The rest of the bandwidth is distributed equally among all BE connection.
Authors tested the performance of DPS algorithm via simulation by using C++ programming language in UNIX environment. They mainly focused on two performance metrics. They are
Average packet loss.
Average packet delay.
Figure: Simulation results for traffic generated and served under various network conditions.
Designer described that the primary goal of the designed scheduling algorithm to share resources equally among each traffic so that it can meet performance standards. DPS scheduler allows the packet transmission per traffic type but not per subscriber station.
They verified the performance of the algorithm via simulation test using C++ language in UNIX environment. Results clearly show to minimize the packet delay and packet loss. In addition, the proposed algorithm provides fairness in resource allocation among various traffic types.
Mac layer handoff eee802.16:
MAC layer handoff method is proposed by Doo Hwan Lee, he explained the entire process in two stages and he clearly described how the process is initiated, executed and terminated after successful completion. He said the process goes under several levels and conducted the simulations test in order to test the performance and handoff latency. Let's see the entire process,
Basically, the MAC layer handoff method is divided in two stages. They are,
Network Topology Acquisition
Actual Handoff process
Initially, NTA is conducted it include Network topology advertisement, Mobile station scanning of neighbour base stations and association procedures. Then followed by actual handoff process, it includes cell reselection, handoff decision and initiation, synchronization with downlink and uplink parameter adjustment, ranging and uplink parameter adjustment, mobile station re-authorization, re-register, terminating with the serving base station.
Figure: Procedure of MAC layer handoff method
Network Topology Acquisition:
In the first stage, before handoff process initiation a network topology acquisition is conducted by both mobile station and serving base station using backbone network before the handoff process initiation. Base station sends the information regarding network topology periodically using MOB_NBR_ADV message, which includes the information regarding neighbouring base stations. Ignoring the broadcast message, the base station synchronizes with neighbouring base stations. Next followed by neighbouring base station scanning, mobile station select some neighbour base station as host base station in order to initiate handover process.
Mobile station carry out downlink synchronization process with each and every neighbouring base station by using MOB_SYN_REQ or MOB_SYN_RSP message. Content or non-content based ranging is initiated to measure the characteristics of the physical channel. By serving BS the entire incoming data is buffered to mobile station during mobile station scanning process. Connection establishment process between MS and BSs is a last step in NTA. Initial ranging is an optional process, which is triggered during mobile stations scanning interval time.
There are some issues exists in this stage. Primarily, mobile station starts the scanning and synchronization process with one or more neighbouring base station as a target BS by conducting neighbouring base station scanning process. The overall system throughput is decreased by conducting the neighbouring BS scanning process when data transmission is paused and supplied.
Figure: Network Topology Acquisition procedure.
Actual Handoff Process:
Basically, handoff process is triggered when the mobile station moves from serving base station to target base station. Once it is migrated, it goes in to several levels in order complete handoff process.
The mobile station initiates this level with the help of information provided by the network topology acquisition stage because the operations, which are carried out in this stage are same as operations in stage one.
Handoff Decision and Initiation:
This level is processed when the target BS receives MOB_MSHO_REQ message from MS. Then the ranging process is triggered after handoff process initiation.
Figure: Decision, Initiation and ranging procedure of handoff process,
Synchronize with new downlink and uplink obtains parameters:
Mobile station obtains new downlink and uplink parameters with DCD or UCD message by synchronizing the target BS. In order to complete initial ranging process both station exchange RNG_REG or RNG_RSP message. This process may be done by conducting contention or non-contention based process.
After all physical parameters are adjusted successfully; the network re-entry process is executed. The re-entry process includes authorization of mobile station and registration of new base station. The target BS request authorization information in order to authorize the Server through network backbone and using REG-REQ or REG-RSP message a new BS registration is performed. The processes include,
Once the target BS registration is done successfully, the serving BS receives the message from MS to notify handoff process is completed.
Here the designer conducted a simulation test to minimize the handoff latency and its performance. He analysed the handoff operation time experimentally. Here he showed some results of performance versus time.
Figure: operational time of handoff process.
The designer concludes that the MAC layer handoff method can able to establish the scanning process before the actual handoff process and with this method response time of the handoff can be reduced.
In this chapter I described the research work done by other people. I explained briefly about the downlink scheduling scheme such as working principals, considerable parameters, issues and results. Moreover, I analyzed MAC layer handoff process and its issues. Clear explained about the process of solving the problems.
Ad-Hoc scheduling scheme:
Packet scheduling for WIMAX has been investigated by many groups of people in wireless community. Many packet scheduling schemes are introduced and classified in various forms like integrated schemes versus uplink and downlink scheduler and others based on centralized scheduling schemes versus decentralized scheduling schemes. Most of the scheduling tasks in centralized schemes are performed in the base station (BS) but in decentralised scheduling the tasks are performed in both stations i.e. base station and the subscriber. Commonly, packet scheduling in WIMAX are separated based on real-time versus non-real-time schedulers.
In the basic schemes, just normal traditional queuing field is considered for all user classes. In such schemes, it uses only one or two basic queue fields in order to separate the traffic and Applying different weights to several classes and priorities is also a common strategy. The scheduling scheme based on the complex scheme, in which some parameters related to medium characteristics should considered and traffic by combine several queuing disciplines.
In hierarchical based schemes, the scheduling is planned in multiple levels. In first level, the traffic from dissimilar service classes is separated first and then separated traffic is scheduled in second level. The different levels of scheduling process can be performed in base station (BS) or overall task could be separated between base station (BS) and subscribed station (SS). An ad-hoc scheduling scheme, where different queuing discipline is used for each and every service flow. Due to its flexibility and unique quality of service requirements, it is used in many schedulers for IEEE 802.16 traffic. In hybrid schemes, multiple scheduling schemes are combined together in order to built more complex schemes. Another complex scheme, which is based on sophisticated token bucket strategy and it allows a tight control over the flow of traffic.
Ad-hoc scheduling is divided in to two stages.
Dynamically allocating priority queue (DAPQ).
In the first stage, four different algorithms are used for five user classes of service in order to meet their Quality of Service (QOS) requirements. To hold the Unsolicited Grant Service (UGS) traffic, the scheduler uses Earliest Deadline First (EDF). All packets are scheduled first based on its maximum latency and its arrival time. In order to allow delay guaranteed service, dynamic priority based scheduling algorithm is used by the EDF. Also it consider some traffic characteristics of the queue, such as
Minimum inter-arrival time.
Maximum service time.
Packet length and delay bounds.
Weight Fair Queuing (WFQ) is employed for Extended Real Time Polling Service (ERTPS) and Real Time Polling Service (RTPS) connections, which is almost similar to Generic Fair Queuing Algorithm. This algorithm main intention is to provide all service with equal allocation of resources' among all users. The basic idea behind the fair queuing is to transmit the data from each and every service. The advantage of WFQ is to allocate the network resources to all the users equally. This algorithm works based on resource reservation technique. However, the disadvantage is of this algorithm, it does not follow strict time limit on each packet. All packets in ertps and rtps are scheduled based on their weights.
In order to maintain nrtps connection, the scheduler users Round Robin (RR) queuing. The strategy is to allocate the resources among the users in RR fashion. It uses time slots to handle all requests in an efficient manner without priority. RR is easy to implement and maintain. Moreover, it does not cause starvation problem for low priority services.
Best Effort service (BE) connection will be handled in FIFO manner and it is suitable for best effort traffic.
Figure: Ad-Hoc scheduling scheme.
Depending on their priority the scheduler goes through all levels in a queue i.e. from highest priority followed by lowest priority. If any packet is waiting in a queue immediately that packet is moved to Active Queue List (AQL). Once the packet is added to AQL, the scheduler moves that packet to final stage of scheduling i.e. to DAPQ component. Then the DAPQ adds some portion of data to each and every packet and assigns the required bandwidth to it. If there are no packets waiting in a queue then the scheduler removes from AQL and bandwidth will not be allocated to it.
Dynamically Allocating Priority Queue (DAPQ) is an advanced version on normal priority queue. The disadvantage of normal priority queue, it follows strict priority policy i.e. from highest to lowest and allocates the bandwidth per flow. In IEEE802.16, the service classes are specified in following order.
Unsolicited Grant Service (UGS)
Extended Real Time Polling Service (ERTPS)
Real Time Polling Service (RTPS)
Non-Real Time Polling service (NRTPS)
Best Effort Service (BE)
In normal priority queue, the highest priority packets will be served first if the flow continues for a long time the packets with lowest priority could go in to starvation problem. In order to overcome this problem, an advanced priority queue uses some reservation techniques. Moreover, it improves the bandwidth allocation by using bandwidth reallocation method.
DAPQ maintain some database (i.e. a list of all active queues) in the AQL until all packets are served in the queue. Once all the packets are served, then that queue is removed from the AQL. The scheduler strictly follows the above relationship among user classes.
UGS Ëƒ (ERTPS, RTPS) Ëƒ (NRTPS, BE)
In order to determine the number of bits at the head of the queue and to keep database of the packets, which are waiting in active queue the scheduler goes through all active queues. This process is repeated until the queue is become empty then scheduler moves to the next active queue. The above process goes through several loops. In each loop, scheduler updates the AQL, such as available frame portion and number of packets in a queue. For each service cycle it goes through several steps, by serving the highest priority first then followed by lowest once. In each service highest priority queue will be processed until available bandwidth is lower than the requested bandwidth by the higher priority queue. This is a best solution for the lower priority queue.
Handoff using mobility pattern
Hand-off method using mobility pattern, the basic idea behind this approach is to minimize the handoff response time i.e. avoiding the unnecessary searching time during handoff process. Mobility pattern table is setup at the base station that helps the MS to predict the target BS. Mobility pattern table is initiated when the BS station establish the connections for MS. For every successful process, the information entered in the table is updated based on the decision received from the BS. Moreover, to minimize the packet loss ratio the BS transmits the data packet to target BS.
Figure: Handover on WIMAX
The subscriber station predicts the target BS by using mobility pattern table. The information regarding previous base station and the target base station is maintained in the mobility pattern table by serving current base station. Mobility pattern table includes three elements,
Previous base stations (BSs) ID
Target base stations (BSs) ID
Previous base station is nothing but it is the station at which mobile station formerly connected to a right association. On other hand, target BS is defined as a MS that is associated after moving from the previous BS. Handoff time is a certain period, with in that number of mobility entries appears.
Initially, mobility patter is empty when it is created and it is updated when entries are added during handoff process. During the process, the target base station is select by the MS based on signal strength and other parameters. The MS sends a MOB_HO_IND message to the current serving base station. The previous base station and the aim BS IDs are piggybacked in that message. Once the serving base station receives that message, the IDs of both stations are updated automatically in the mobility patter table. Then the serving base station checks the updates in the table. If the IDs of both station exists in the table, then the BS increase the value of handoff time by 1. Otherwise, the handoff value is set to 1.
The signal strength of the serving base station gradually degrades, when MS migrates from serving BS. The mobile station initiated the scanning process in order to find another base station when the predefined signal strength of base station drops. Mobile station retrieves the required information from neighbour base stations, when it receives the MOB_NBR_ADV message from the serving base station during that movement. In the scanning process, the PHY channel information is used for future selection from the base stations list.
Before the scanning process is initiated by the MS, it sends scanning request to current serving base station. The message includes the base stations list and then the selection of base station is done according to PHY channel information of neighbouring base stations. The neighbouring base stations are included in scanning request message if it meets the requirements of MS. Moreover, IDs of previous BSs are also included in that msg. The current BS finds the candidate BSs up on the received scanning request message based on the previous BSs IDs list, which are saved in the mobility pattern table.
For each and every entry in the table, the serving base station compares the value of previous base station and BSs IDs. After comparison if the value and IDs are same as the IDs in scanning request message, then the target base station which is entered in mobility pattern table is considered as a candidate BS. The ID of the base station is added to list of base stations. The group of candidate base stations, which are in mobility pattern table and the group of candidate BSs, which are specified in MOB_SCN_REQ are represent as available scanning base stations list. Apparently, scanning of all base stations in list takes long time which is not suitable for real time applications.
Accordingly, in this method the classification of available BSs is done based on their priority.
If handoff time of the previous base station and available base station is high according to the
MP table, during the handoff process the available base station have more opportunity to be associated. Therefore, base on the handoff time the base station directs the available base stations in the list. The serving base station allocates the scanning time to manage the candidate base stations and MOB_SCN_RSP message is returns to subscriber station.
According to the MP table, the entire list of BSs is retrieved by the base station in decreasing order of handoff time in this response msg. Once the scanning response message was sent, the serving base station initiates the buffering process, which forwards all packets to the MS.
Next, the MS successfully disconnects the connection with the serving base station and starts the scanning process again according to the list of available base stations.
In order to retrieve the detailed information regarding PHY channel from the candidate BS, the MS synchronize with the BS. Scanning process is safely terminated if the channel conditions meet the required requirements of the MS.
Figure: Flow chart of Handoff process using mobility pattern.
Chapter5: Critical Appraisal, Recommendations and Suggestions for Future work
The report describes clearly what I have done. I reviewed articles, journal and the work done by the other groups in wireless community using various sources. I analyzes evolution of the topics which I under taken and the issues which are associated with them. Moreover, I investigated various approaches and methods in order to solve the research questions, which I mentioned in the chapter1. I came up with efficient approaches that can able to give a solution to the questions. In addition, I compared methods with others approaches and I found some interesting conclusion. Let's see some parameters which I compared,
Comparison of QOS scheduling schemes in WIMAX
No. of classes
Ad-Hoc scheduling scheme
Downlink scheduling scheme
Proportionate fair scheduling scheme
Cross-layer scheduling scheme
TCP-Aware scheduling scheme
After comparing all the methods of my research, I believed that Ad-hoc scheduling schemes work efficiently and the way it gave the solution was explained in the chapter4.
Moreover, Handoff Management is a crucial feature in WIMAX to provide better Quality of Service (QOS). I investigated its evolution and methods used to minimize the issues in HM. Compare to all other method, I believe this approach is very efficient, simple and easy to implement. Moreover, it is compatible as well.
Critical Appraisal, Recommendation and Future work
When I done my graduation in Computer Science and Engineering, I came across many concepts related to networking and security background and I was fascinated towards them. That made me to choose Computer Networking course in London Metropolitan University. The dissertation which I took was based on Quality of Service in WIMAX. I took some part of the WIMAX and done research on it such as scheduling and handoff management.
WIMAX is a crucial technology that can able to provide enhanced wireless communication for mobile terminals with great range in MAN. Due to enormous complexity in the network and to meet the future requirements we have to propose sophisticated approaches and methods. The complexity of network is increasing day by day due to increase in users. Quality of service (QOS) plays a vital role in wireless communication community. So we need design and implement more and more highly developed approaches, schemes and methods in order to fulfil the future requirements and it should meet the Quality of Service (QOS) requirements as well.
My future work is to develop a new sophisticated scheduling algorithm by introducing QOS parameters (such as fairness, scalability etc.) and distribution of bandwidth utilization amount various services. The scheduling algorithm can deliver fairness among different service traffic simultaneously.
On the other hand, Mobility management is an important feature on WIMAX. I continue my research to implement better and fast handoff methods so that it can predict the target stations more in a less amount of time and minimize the whole scanning time without wasting the channel resources. Moreover, I'm will concentrate on security issues as well.
The Ad-Hoc scheduling scheme is an efficient method that can able to handle the service traffic flow in an efficient way by minimizing the starvation problem among packets with different priorities. It supports all Quality of service classes, which are mentioned in IEEE802.16 standard related to WIMAX. In addition, it shares the available network resources in better manner among all services equally. On other side, the handoff process using mobility pattern can reduce the valuable time to predict the target base station without wasting any resources by avoiding unnecessary scanning.