Co-existence and interworking of 802 by pengtt


									Co-existence and Interworking of 802.11 and 802.16 networks

                                                 Prepared by

                                                 Nirmal Andrews

                                          Table of Contents

CONTENTS                                                                       PAGE

List of tables                                                                  3

List of figures                                                                 4

1. Differences and characteristic of 802.11 and 802.16                          5

        1.1. 802.11e Quality of Service mapping                                 5

2. Advantages and feasibility of 802.11/802.16 co-existence and Interworking    7

3. Architecture and deployment models of 802.11 and 802.16 coexistence          9

4. Research issues in an integrated 802.11/802.16 Network                       15

        4.1. Co-existence of 802.11 and 802.16: Issues and previous work        16

        4.2. Interworking of 802.11 and 802.16: Issues and previous work        19

References                                                                      21

                                            List of Tables

TABLES                                                                    PAGE

1. Difference between operation of 802.11 and 802.16 except for 802.11e    4

                                             List of Figures

FIGURES                                                                                   PAGE

1. Mapping by EDCF in 802.11e                                                              6

2. 802.11e Access Categories (AC). Four service flows                                      7

3. Advantages of the synergy of 802.11 and 802.16                                          8

4. A deployment scenario where 802.16 (WiMax)/ 802.11 (WiFi) is used „on the go‟           9

5. Dense Urban area coverage using 802.11 – 802.16                                         10

6. 802.11 and 802.16 interworking network serving less populated areas                     11

7. City wide model for deployment of integrated 802.11/802.16. Larger cells are            12
  WiMax cells and the cells inside them are WiFi cells.

8. The detailed structure of 802.11 and 802.16 interworking network                        13

9. Flowchart representation of the various related work in interworking and coexistence    15
  based on protocol layers

1. Differences and characteristic of 802.11 and 802.16

         The study of coexistence and interworking of 802.11 and 802.16 networks are almost impossible
without the knowledge of the differences in operation and access schemes. Therefore this table below
gives a brief overview of each technology and the differences between them. Based on this table it could
easily seen that interworking between both the networks need a lot of coordination at various levels of the
network namely physical, MAC , routing and transportation layers.

  Networks         Wi-Fi (802.11 - based on              WiMax (802.16 - based on commonly used
                commonly used technical name)                         technical name)
   Mode of         1) Infrastructure                     1) Point to multipoint (PMP)
  Operation        2) Ad-Hoc                             2) Mesh
    PHY            1) DHSS                               1) Wireless MAN SCa : 2-11 GHz (PtP)
                   2) FHSS                               2) Wireless MAN OFDM: 2-11 GHz(PtMPt)
                   3) Infrared                           3) Wireless OFDMA: 2-11 GHz (PtMPt), more
                                                         4) Wireless HUMAN (High-speed unlicensed):
                                                            2-11 GHz (PtMPt)
    MAC                     CSMA/CA (most                           TDMA/TDD (most common)
    access                    common)
   Systems      802.11 a     802.11 b    802.11 g     802.16      802.16 a       802.16 d        802.16 e
  Spectrum       5 GHz       2.4 GHz     2.4 GHz       10-66      <11 GHz     10-66 GHz and      <6 GHz
                                                       GHz                       <11 GHz
  Data rate     54 Mbps      11 Mbps     54 Mbps        120       75 Mbps       120 Mbps         15 Mbps
 Modulation      OFDM         DSSS        OFDM        QPSK,       OFDM, OFDM,OFDMA unknown
                                                     16-QAM,      QPSK,     ,HUMAN (High
                                                     64-QAM     16-QAM,           Speed
                                                                 64-QAM       Utilization)
              Table 1: Difference between operation of 802.11 and 802.16 except for 802.11e

         This table specifically does not mention 802.11e since it is different from the rest of the 802.11
systems and its features and structure make it highly adaptable to co-existence and interworking. The
above mentioned 802.11 systems do not focus a lot on Quality of Service architecture and so they usually
have only one type of service flow which is Best Effort. Considering the fact that Quality of Service plays
an important role while integrating 802.11 and 802.16, we should understand that 802.16 have four
different service flows for guaranteeing Quality of Service. The details of the service flows will be
discussed in a while. The mapping of Quality of Service between 802.11 systems with Best effort service
flow to 802.16 with four different service flows creates quite a challenge. Therefore, 802.11e which was
developed for providing better Quality of Service standards for the 802.11 systems have categorized
service flow which makes it more adaptable to co-existence and integration than its peers.

1.1. 802.11e Quality of Service mapping

       The importance of matching Quality of Service services between networks for internetworking
makes it important to know the details of the Quality of Service structure of 802.11e. In this section we
expound upon the details of 802.11e features which enabled in Quality of Service provisions.

        802.11e as mentioned above proposes enhanced mechanisms which provide good Quality of
Service to an application based on traffic type. There are a few improvements proposed by 802.11e for
providing the good Quality of Service. Some of them are:

1) Introduction of enhanced distributed channel access (EDCA), which is a contention based channel
access, which improved the DCF (Distributed Coordination function).

2) Introduction of Hybrid Coordination Function controlled channel access, a controlled channel access,
which improves the PCF (Point Coordination Function).

3) These two entities are controlled by a Hybrid Controller which is module found in the Quality of
Service Access point.

         The other major concept introduced was traffic differentiation using traffic specification
(TSPEC). The TSPEC describes the Quality of Service requirement of a traffic stream (TS) from a STA
or mobile node using 802.11e. The frame format of TSPEC gives us details of traffic and required Quality
of Service. In 802.11e the traffic is divided into 4 information categories which are background, best
effort, video and voice. These 4 information categories are further assigned user priorities. The user
priorities are numbered 0 to 7. The details of the user priority are given in 802.1D standards specification.
Now, each node or STA which has Quality of Service support in them has four queues (AC) for the 8 user
priorities. A packet with a certain user priority is said to belong to a traffic class with that user priority.
Transmission is provided to the queue with the highest priority.

                                    Fig 1: Mapping by EDCF in 802.11e

                      Fig 2: 802.11e Access Categories (AC). Four service flows.

2. Advantages and feasibility of 802.11/802.16 co-existence and Interworking

    Integration of 802.11 and 802.16 is an interesting topic because of the benefits of the synergy both
economically and technically from the users‟ as well as the vendors‟ perspective. 802.11 and 802.16 are
complementary technologies in the fact that 802.11 is used for very high speed WLAN ( Wireless LAN)
connectivity and 802.16 (Wireless MAN) is used for high speed WMAN connectivity. Both these
technologies use IP-based technologies to provide connectivity to Internet and when this is combined
with the equipments being standardized by WiFi Alliance and WiMax Forum makes it feasible to deploy
and use easily. According to Motorola and Intel (2007) the impact of the synergy between 802.11 and

    1) Enables high connectivity, where users can connect to either of the networks based on their
       location, Quality of Service requirements and coverage.
    2) Enables efficient global roaming.
    3) Enables provision of broader range of services both at home as well as on the road.
    4) Enables usage of both licensed and license exempt spectrum based on service requirements.
    5) Enables economical coverage of large areas.

The above mentioned reasons make the integration highly desired and attractive. Integrating 802.11 and
802.16 provides an efficient broadband connectivity, which is both convenient and affordable. For service
providers it brings new deployment models and for users it brings new usage models.

     Necessities for the interworking between 802.11 and 802.16 as suggested by Motorola and Intel
(2007) are: multi-mode subscriber devices that can communicate with both 802.11 and 802.16 networks,
session continuity. Specifications allowing transition of users from 802.16 networks to other networks
have been developed by the Network Working Group (NWG). An ASN (Access Service Network)
gateway has been designed by the NWG for managing access to services such as AAA (Authentication,
Authorization and Accounting), DHCP, session and mobility management. These developments will
facilitate in making interworking of 802.11 and 802.16 feasible.

                                      Why integrate
                                      802.11 and

                        Quality of Service            Reduces service         Global
                        provision                     cost                    Roaming

                         Fig 3: Advantages of the synergy of 802.11 and 802.16

    Apart from Motorola and Intel (2007), there are few other authors who provide arguments for the
integration of 802.11 and 802.16 based on the impact of the synergy that the integration can produce.
Gakhar et al. (2005) states that efficient interworking between 802.11 and 802.16 is needed to support
Quality of Service for delay sensitive bandwidth intensive applications such as VoIP, video transmissions,
large volume FTP, etc. 802.11 provides high data rate but not the Quality of service but, in an
Interworking network between 802.11 and 802.16, the architectural design has 802.16 as the serving
network and so Quality of Service can be maintained since 802.16 has an efficient Quality of Service
policy. Fantacci and Tarchi (2006), finds that the one of the main advantage of the interworking is the
possibility of seamless integration between heterogeneous networks within an urban area. Another
advantage mentioned is the provision of Quality of Service which was not a priority in 802.11 networks.

    Ali-Yahiya et al. (2006), mentions that there are several advantages of both 802.11 and 802.16
working Interworking. The authors explain that integration would benefit the operator by providing larger
network access coverage with cheap investment. From the users perspective the benefit is in the provision
of a ubiquitous network access with a guaranteed service. Gunasekaran and Harmantzis (2006) stated that
wireless ISPs‟ (Internet Service Providers) can use the most mature technology by integrating 802.11 and
802.16. 802.11 can be used the reach end user and 802.16 can be used as backhaul which will minimize
the backhaul cost and efficiently reduce the time of service provisioning. The authors make a very
interesting statement that “If proper planning and deployment is done the integrated network can turn a
whole region within geographic boundaries to a „hot zone‟ ”.

    Berlemann et al. (2006), points out that the impact of the 802.11 and 802.16 synergy is in expanding
coverage over rural areas where the deployment of wired networks is too expensive due to the marginal
density of population. Additionally the authors also indicate that 802.16 can provide multi-hop, relay-
based wireless backhaul serving 802.11 WLAN hotspots. The multi-mode relays can benefit from an
Interworking capability between both standards. Thomas, Nicholas J. et al (2006), discusses the
possibility of co-existence of 802.11 and 802.16 and the authors‟ state that the possibility has increased
dramatically based on the increased usage expected in the future years. Especially since 802.11 is widely
established and new technologies are still coming in the possibility of co-existence is becoming a reality.

3. Architecture and deployment models of 802.11 and 802.16 coexistence

    Motorola and Intel (2007) recommended four different deployment models for the integrated 802.11
and 802.16 architecture. The different deployment models were:

1) Broadband „on the go‟
   Deployment of 802.11 and 802.16 in the user devices, allows service providers to offer transparent
   service between 802.11 in hotspots and 802.16 beyond the hotspots. Deployment of 802.16 in areas
   with high density of Internet users extends broadband services beyond hotspots providing both users
   and service providers the utility and value of service.

        Fig 4: A deployment scenario where 802.16 (WiMax)/ 802.11 (WiFi) is used „on the go‟.

2) Last mile broadband
   In rural areas where the population density is scarce the expansion of DSL will be very expensive and
   in urban areas where it could be difficult to add wired connections to existing multiple dwelling units
   (MDU) the integrated network can be used to extend broadband connectivity in the last mile. 802.16
   will be used as backhaul which will make the extension less costly.

3) Broadband campus Coverage
   802.16 (WiMax) can cover larger area. It can provide connectivity beyond the individual buildings to
   an entire campus as blanket coverage. This allows service providers to offer choice of connectivity to
   users enabling them to connect to the 802.11 (WiFi) network in the building or the 802.16 network
   which covers the campus. This usage of 802.16 can reduce the number of 802.11 access points
   needed to cover the entire campus, thereby reducing maintenance cost.

4) Citywide broadband
   Cities are now trying to implement city wide wireless broadband coverage using 802.11 mesh
   networking. 802.11 mesh networking although is efficient proves to be highly expensive. The cost
   rises from power supply and wired connections to each of the access points in the mesh network. The

    integration of 802.11 and 802.16 provides a cost effective alternative for the 802.11 mesh network by
    using 802.16 as the backhaul to connect the access points from the mesh portals to the Internet

    Gunasekaran V. et al., (2006) in their study of financial assessment of citywide Wi-Fi (802.11)/
WiMax (802.16) deployment classified the possible scenarios of deployment of an integrated 802.11 and
802.16. The classifications are: dense urban area and low density area. The deployment structure
described by the authors is essentially an extension of the underlying 802.11 network deployment model
which exists. Camponovo et al., (2003) describes the two 802.11 deployment structures: selected
locations or hotspots covering specific locations where internet usage could be high such as airports,
hotels, cafes‟, etc. and extensive or outdoor which could be an example of areas with low population
density where 802.11 is used to provide outdoor coverage

   In the dense urban areas majority of the SOHO (small offices, home offices) and households are in
multi dwelling units or apartment complexes. 802.16 backhaul can deliver megabits of data to the multi
dwelling units from where 802.11 can be used to distribute services to individual users or rooms.

                         Fig 5: Dense Urban area coverage using 802.11 – 802.16

    The second case is where the population density is low. This scenario differs in the way that it
requires a larger number of 802.11 access points to cover the area. Gunasekaran et al. explores the option
of a single private owner model for operating the services and selling it directly to the customers. In this
model explored by the authors either the local government makes an agreement with the company to build
and own the network. The other possibility is that the local government can make deals with multiple
owners to grant license to operate, providing the service providers the city‟s existing infrastructure like

street poles and lamp post to build the network reducing the capital cost and operating cost, which enables
the operators to offer lower cost of service to customers.

              Fig 6: 802.11 and 802.16 interworking network serving less populated areas.

    Gunasekaran et al., also provides a detailed layout of how the infrastructure should look like in an
integrated 802.11/802.16 network. According to the authors the backhaul should be designed using
802.16 mesh network. This reduces the backhaul cost and makes efficient use of the wired backhaul. In a
36 square mile radius assuming that 9 802.16 cells of diameter coverage 2 miles can be fitted, it would
require 9 wired connections for each of the BS in the cell, but in case of a 802.16 mesh 1 wired
connection could substitute the 9 wired connections instead. Once the 802.16 cells are in place, the next
step is to pack these 802.16 coverage cells with 802.11 access points (AP) and using 802.11 as the last
mile connection. The usage of 802.11 as last mile makes use of the fact that 802.11 is ubiquitously
available and already established widely. The authors suggest the usage of 802.11 a/b and g to provide an
extensive option for users as well as operators to provide best service. The mesh node communication can
use one technology and client to node can use another technology. 802.11 a can be used for voice uses
and 802.11b can be used for data users.

    Fig 7: City wide model for deployment of integrated 802.11/802.16. Larger cells are WiMax cells and
                                  the cells inside them are WiFi cells.

    The authors study the most optimal deployment of 802.11/802.16 network based on financial aspects.
The assumptions made by the authors for the modeling are that if multiple operators exist then 802.16 BS
uses sectorization instead of cell splitting. Six sectors are used with 18Mbps capacity in each sector. 15-
25 802.11 cells can cover the area of one 802.16 cell. Deployment of more 802.11 AP is better since
demand per unit area would never exceed the capacity of a single AP. The cost model also assumes that
75% of the area covered will be low population density area and the rest 25% will be urban densely
populated areas. The average bandwidth for light users is 250kbps and for heavy users is about 500 kbps.
Another assumption made is that 25% of the residential users are heavy users. The various service options
provided in the model are on demand connection service as well as monthly subscription. The model
designed based on all the above mentioned assumptions provide breakeven for the operators within 2
years of operation. The benefit for the end user is lower cost of service and multiple service provisions.
The monthly subscription cost of USD 15 for light users and USD 30 for heavy users. For business users
the service is provided for USD 60 per month. On demand users will be charged USD 5 per day of usage.
USD 20 per month will provide voice over WiFi (VoWiFi) which allows users to make unlimited local
and long distance calls for an entire month for just USD 20. It can be observed that the model proposed
by the authors is very attractive in terms of financial benefit for both users and operators, but this paper
does not discuss the feasibility of such a project with 802.16 mesh and 802.11 last mile. There are no

details to establish the operational viability of this model technically, which makes it weak despite of the
attractive financial model. However, the authors give detailed insight into various deployment

    Jing, Xiangpeng and Raychaudhuri, Dipankar (2005) in their study of co-existence of WiFi and
WiMax simulated the multiple 802.16 BSs‟ and 802.11 hotspots‟ scenario along with different 802.16 SS
geographic distributions. When the SS are distributed uniformly it is observed that there is a throughput
increase of 15% in 802.11 network and when the SS are clustered around the hotspot the throughput
increases by 160% for any larger clusters the throughput degrades. Figure 5 shows the model being
discussed here. This study identified various 802.16 SS node distribution and spatial clustering and their
effects on the throughput. This study provides another insight into a possible deployment structure of
802.11 and 802.16 integrated networks to achieve better performance.

                 Fig 8: The detailed structure of 802.11 and 802.16 interworking network

     Yahiya Ali et al., (2006) also describe a common scenario of 802.11 and 802.l6 interworking which
will be a closer look into the model. 802.16 networks consist of several base station (BS) and subscriber

station (SS). Each BS covers an area called a cell where multiple SS can communicate with the BS. The
SS aggregates traffic coming from different terminals or end users such as 802.11, Ethernet users, 802.16
users, etc. The border of 802.16 cells is overlapped with 802.11 cells. The 802.11 users will send their
traffic to the 802.16 BS via special access points names ESS (Extended Subscriber Station) in the 802.11
cells, which can transmit the 802.11 traffic in a form compatible to 802.16. ESS essentially acts as a
translator and carries different traffic and sends them to 802.16 BS. Figure 5 expounds on the model
discussed here. This could also be described as a tight coupling scenario in interworking networks.

4. Research issues in an integrated 802.11/802.16 Network

                        Coexistence and Interworking of (Wi-Fi) 802.11
                        and (WiMax) 802.16

    Physical           Data link            Network               Transport            Application

                            1) Bandwidth based vertical
                            handoff (Jing et al., 2005)
                            2) Hybrid 802.11/802.16
                            routing protocol. (Ibanez et al,

            MAC                                                1) IROISE QoS Interworking
            1) BSHC for MAC                                    architecture. (Gakhar et al., 2005)
            frame control.                                     2) 802.1D/Q bridge for end to
            (Berlemann, 2006)             Service              end QoS. (Fantacci and Tarchi,
            2) Software upgrade                                2006)
            to MAC of 802.16
            BS. (Berlemann,                                    1) Bandwidth Threshold based
            2006)                                              (Yahiya et al, 2006)
                                       sharing                 2) Policy based resource
            3) MAC bridging (
            Fantacci, 2006)                                    management (Yahiya et al, 2007)

                            1) Co-existence analysis between 802.11 and 802.16
                            systems. (Thomas et al., 2006)
 Resource                   2) Reactive cognitive algorithm (Jing, 2005)
 sharing                    3) CSCC etiquette protocol (Jing and Raychaudhri, 2005)
                            4) Citywide WiFi/WiMax deployment topologies.
                            (Gunasekaran, 2006)

  Fig 9: Flowchart representation of the various related work in interworking and coexistence based on
                                              protocol layers

4.1. Co-existence of 802.11 and 802.16: Issues and previous work

    As the name suggest co-existence involves just the study of various topologies in which 802.11 and
802.16 co-exist within the same coverage region and how the co-existence of the networks affect various
operating parameters of each network since they share the same resources such a frequency of operation,
overlapping channels, etc. 802.11 had been an established network around the world in the unlicensed
bands of 2.4 and 5 GHz. The 802.11 protocol allows a fair co channel sharing in ad-hoc manner with the
use of CSMA/CA access mechanism. The introduction of IEEE 802.16 which operates in the frequency
range of 2-11 GHz increases the chances of the co-existence of 802.11 and 802.16. The operation in a
similar frequency range would mean sharing the same resources such as channels and operation bands.
Now, since 802.16 was developed for operation in licensed band and it uses PCF, channel sharing with
other networks is not a feature that 802.16 facilitates. This makes the understanding of co-existence of
802.11 and 802.16 of utmost interest and importance.

     Thomas, Nicholas J. et al (2006) proposed a study to observe the effects and issues of coexistence of
802.11 and 802.16 on each systems performance. The mutual impact of the co-existence of both the
networks was studied in detail through a MAC layer event based simulation. In this study the authors
study three possibilities of coexistence which are: co-existence effects between 802.11 among each other,
co-existence of 802.16 among each other and co-existence of 802.11 with 802.16. The parameters varied
in the study were RTS/CTS, connectivity, packet length, the distance of coverage region and interference
region. Two measures of connectivity were used which were minimum connectivity i.e. the nodes on the
edge of coverage could just get to the AP and full connectivity where the users on the edge of the
coverage region could get to node on the opposite end of the coverage region. The results show that
coexistence on the same channel of combinations of 802.16 and 802.11 systems can degrade the
performance of both systems significantly. Study indicates that co-existence of 802.11 with itself can
approach a single systems limit, but the protocol in itself has poor efficiency especially in cases with
RTS-CTS, large number of nodes and short packets. The throughput tends to half the single system limit
on an average of the offered load in co-existence of 802.11 with itself without RTS-CTS and short packet
transmissions. This study indicates that co-existence of 802.16 with itself is almost impossible. The
systems must be separated by at least one coverage radius. In the coexistence of 802.11 with 802.16 it can
be observed that performance of both the systems will suffer. The total throughput drops to about 10% of
the combined offered load. Therefore it can be seen that coexistence of these networks reduce system
performance because of the interference and also because there exists no coordination or call admission
protocols to perform the coexistence in a controlled fashion.

    Since it has been proved that the coexistence of 802.11 and 802.16 without any control scheme would
lead to severe interference and performance degradation, Jeng Xiangpeng et al (2005) designed a
cognitive radio based reactive algorithm for sharing radio resources in frequency, space and time. Three
reactive schemes designed are dynamic frequency selection (DFS) scheme which chooses the band with
least interference to minimize interference effects, power control (PC) was used to minimize transmit
power as required based on interference and time agility (TA) was exploited to adjust traffic pattern to
avoid interference. The co-existence scenario considered is Wi-Fi and WiMax in which Wi-Fi hotspots
are inside a WiMax cell and share the 2.45GHz frequency band. Bandwidth allocation starts with center
frequency 2412MHz in both the system. 802.16 use OFDM with 20MHz bandwidth providing three non-
overlapping channels for 802.16 in the 2.4 GHz band. 802.11 use DSSS with a bandwidth of 22MHz per

channel providing 11 overlapping channels itself. Simulation is based on one 802.16a cell and one
802.11b hotspot and then a more realistic study with multiple hotspots and 802.16 BS is also considered.
The results show that the reactive cognitive radio algorithms can significantly improve the spectrum
efficiency and throughput and can reduce interference. Using time agility throughput of the hotspot could
be increased by 30% and using power control scheme the throughput at the 802.16 SS can be increased by
4 times while reducing the throughput of 802.11 hotspots by a very small amount.

         Despite of the improvement in system performances based on reactive schemes proposed above,
it appeared that reactive schemes adjust PHY parameters based on local observations alone, which are
insufficient in scenarios which involve hidden stations or receivers. It is generally accepted that
unlicensed band control schemes such as listen before talk are not applicable in hybrid scenarios due to
„hidden-station‟ problems. To counter this problem Jing, Xiangpeng and Raychaudhuri, Dipankar
(2005) designed a proactive scheme named as common spectrum coordination channel (CSCC) etiquette
protocol. The common channel shares radio parameter information of nodes based on which the nodes
initiate appropriate spectrum sharing policies. Each node is considered to have a narrow band radio with
low bit rate through which it listens to and broadcasts announcements of its parameters. This scheme
solves the hidden station problem since the range of the CSCC broadcast is designed to exceed that of the
regular service data range, and receivers can announce their presence in a larger range to optimize
spectrum usage. Spectrum coordination is done using both adaptations in frequency and power based on
the CSCC broadcast received. Simulation scenarios include single 802.16 BS with single 802.11 hotspot
and more realistic situations of multiple BS and hotspots. Using CSCC it can be observed that in the case
of a single BS with single hotspot the DL throughput of 802.16 is improved by 35% with the change in
distance between the SS and hotspot. In the multiple BS and hotspot scenario different 802.16 SS
geographic distributions are studied. When the SS are distributed uniformly there is an 802.11 throughput
increase of 15% and when the SS are clustered around the hotspot the throughput increases by 160% for
any larger clusters the throughput degrades. This study identified various 802.16 SS node distribution and
spatial clustering along with CSCC and observed that it can improve the system throughput significantly
by solving the hidden station problem. Both the above mentioned solutions are based on the PHY layer
and they do not consider MAC layer solutions and any layer above MAC. Handoff is a key process to
successful communication in a network.

     Berlemann, Lars et al (2006) studies the coexistence of 802.11(a) and 802.16 within the license
exempt frequency band and proposes a solution for interference avoidance in the coexistence scenario.
The frame based medium access of 802.16 makes it very essential that rigorous protection should be
provided for the 802.16 frames from WLAN before and during the 802.16 transmissions. The authors
focus on coexistence in the 5GHz UNII band. Since this paper deals only with coexistence and not with
interworking the solution proposed is of the form of software implantations rather than any hardware
models. Existing solutions such as dynamic frequency selection and power control do not take into
account Quality of Service support. When Quality of service is demanded along with mitigation of
interference in coexistence a deterministic control of radio resources is required for all coexisting wireless
systems. The proposed solution as mentioned above is a software upgrade of the MAC of 802.16 BS. In
this solution full control of the radio resource is provided to 802.16 and 802.11 gets access only with the
permission of 802.16. The solution targets the avoidance of an idle medium for duration of DIFS or
longer before and during the 802.16 MAC frame transmission. The 802.16 MAC can be divided as
preamble indicating the start, DL bursts, contention intervals for initial ranging and bandwidth request

followed by UL bursts. Whereas, the 802.11 transmission is contention based and if the node finds the
channel free for more than or equal to DIFS time period it transmits its packets. 802.11 is allowed to
transmit between two 802.16 MAC transmissions, but if the time for the arrival of next 802.16 MAC
frame is lesser than the maximum time taken for 802.11a transmission then the 802.16 BS blocks the
medium to 802.11. Any ongoing 802.11 transmission is allowed to be completed but no new
transmissions are allowed. DL bursts are continuous without any intervals, but UL transmissions could
have intervals equal to or more than DIFS if an SS does not allocate a UL burst. In this case also the
802.16 BS blocks the medium to 802.11a access. The contention slots for initial ranging and bandwidth
allocation if left unallocated could also lead to idle medium for time greater than DIFS. So the 802.16 BS
has to block the medium for initial ranging slots since they are longer than DIFS by default. So slots for
initial ranging are allocated after the contention slots. Even though coexistence with Quality of Service is
achieved by 802.16, this solution could prove to be highly unfair to 802.11a users.

    Every network has to deal with handoffs because of the cell based structure of network deployment
and operation. When a user moves from one cell to another he loses signal and the call gets cancelled in
case the BSs‟ in the neighboring cells do not pick up the call. When the handoff takes place between
similar type of networks it is called horizontal handoff and when the handoff is in a hybrid network it is
called vertical handoff. Vertical handoffs are much more complicated since methods used to measure
signal strength in one type of network might not be the same as in other networks. This could lead to
unnecessary handoffs due to miscalculations. The integrated 802.11 and 802.16 networks face the same
issue of vertical handoff. The unnecessary handoffs could lead to drastic reduction in throughput and call
handling efficiency.

    Nie, Jing et al., (2005) discusses the issue of vertical handoff within 802.16/ 802.11 interworking
networks and provides a solution based on bandwidth as a metric for handoff decision. The authors
mention that the previous work done in vertical handoff has been between a high bandwidth network and
low bandwidth network, but in the case of 802.11/ 802.16 both the networks are high bandwidth
networks. Since both the networks in consideration here provide high bandwidth, the handoff should be
made carefully. If the user is in the 802.16 network, the authors suggest calculation of bandwidth in the
MAC layer and make initial registration before handoff by network layer. The authors derive a formula to
obtain bandwidth from NAV. From the 802.11 NAV the 802.16 network calculates the available
bandwidth in 802.11. If the bandwidth provided by 802.11 is higher than that offered to 802.16 currently
the handoff is made. Now if the user is in 802.11 network, then the handoff decision is made based on
RSS (Received signal strength) and bandwidth. Since the RSS can fluctuate frequently the handoff is
made based on aggregate RSS and compared to a threshold RSS, which is minimum required RSS for
good communication. The authors also propose two hop relay architecture with mobile nodes to be relay
gateways. The nodes with only one interface can use nodes with dual interfaces as relays to the 802.16
network. This allows 802.11 nodes to discover and utilize better quality radio channels in 802.16 network
to achieve higher bandwidth and wider coverage without introducing protocol complexity. However, the
users do not take into account interference and cooperation issue in their solution.

4.2. Interworking of 802.11 and 802.16: Issues and previous work

    Coexistence just studies the effect that the coexisting networks have on each other. However,
interworking goes deeper since it has to consider MAC as well as Quality of Service mapping issues for
the combined working of both the interworking networks. Different uses of Interworking of 802.11e and
802.16 are using 802.16 as a backhaul in rural areas where fixed network deployment is expensive. It
could also be useful in scenarios where 802.16 will be deployed as multi-hop, relay-based wireless
backhaul serving 802.11 WLAN hotspots.

    Berlemann, Lars et al (2006) indicates in their work that information exchange between spectrum
sharing networks enables interworking but it is not required for co-existence. Interworking allows
networks to coordinate the spectrum sharing among each other. However the information exchange
requires a common frame structure and full control of MAC. The author proposes a single common
communication device, capable of operating in both 802.11e and 802.16, centrally controls coordination
and Interworking between 802.16 and 802.11e when operating in same frequency channel. The single
common communication device is named as Base Station Hybrid Coordinator (BSHC). The BSHC is a
combination of 802.11e Hybrid Coordinator and 802.16 BS. To 802.16 it appears as BS and to 802.11e it
appears as Hybrid Coordinator. The Interworking is made possible in BSHC by integration of 802.11e
transmissions into the MAC frame structure of 802.16. An optional period for contention based access
may be placed between two consecutive 802.16 MAC frames. The allocated time intervals for
transmissions are considered as DL/UL burst by 802.16 and Transmission opportunity by 802.11e. In case
of a communication between 802.11e and 802.16 the BSHC receives the data to be transmitted in the
transmission opportunity time from 802.11e and transfers it 802.16 MAC frame in between the DL burst
if 802.16 MAC. The similar OFDM based transmission allows facilitates the Interworking between
802.11e and 802.16. The results based on Interworking using a common control device BSHC shows that
maximum available throughput and optimal partitioning of MAC frame depend on the number of 802.11e
nodes (QSTAs) served by the BSHC. Increasing number of QSTAs decrease the overall system capacity
because of the contention in contention period of transmission. Authors work show that Interworking
influences the MAC of all participating networks and the restrictions and requirements of each protocol
have to be combined to obtain Quality of Service under coexistence scenarios.

     The previous models designed for integrated operation of 802.11 and 802.16 have considered both the
networks to be operating in the same frequency and hence have focused specifically on interference
avoidance more than operational issues. Tang-Lin et al (2009) however have designed a model focusing
on the operation under the assumption that 802.11 and 802.16 networks will be operating using different
frequencies which avoids interference. The authors propose a connection oriented unified architecture for
the integration of 802.11 and 802.16. The model introduces an 802.11/802.16 access point and the authors
name it W2- AP. The W2- AP acts as either a bridge in translating frames between 802.11 and an Ethernet
interface or as a relay node transferring frames between 802.11 and 802.16. This is made feasible by
introducing a modified convergence MAC layer of 802.11 interfaces which is designed by embedding
802.16 subscriber MAC function within the original 802.11 MAC. The W2- AP and the devices work
together to enable an 802.11 hotspot which facilitates connection oriented transmissions and
differentiated services. The authors also introduce a bandwidth allocation scheme named THBA, this
scheme is similar to the one used by Niyato and Hossoin (2007) to study the resource allocation in the
802.11/802.16 integrated network. However, Niyato and Hossoin (2007) provided only a skeletal

structure of the bandwidth allocation in two levels. Tang-Lin et al. (2009) gives the intricate details of
how this is to be performed. The authors use the model in order to provide continuity in Quality of
Service. The individual nodes request for bandwidth to the W2- AP and the W2- AP in turn requests
bandwidth to the 802.16 BS. There involves a delay in this process between the BS providing the
bandwidth to W2- AP and W2- AP in turn distributing the bandwidth among the individual nodes. The
authors simulate the scenario and study its impact on three service flows which are video, VoIP and web.
The end to end delay was studied to understand the impact on Quality of Service. It is observed that the
proposed THBA scheme with changes in W2- AP provide significant improvement in end to end delay
due to the connection oriented unified approach. VoIP calls achieve committed Quality of Service and are
not affected as the network load or connection increase progressively. It enables differentiated services
between 802.11 and 802.16.

     As mentioned earlier 802.16 has predefined Quality of Service framework, whereas legacy 802.11
and the 802.11a/b/g do not have good Quality of Service frameworks. However, 802.11e as described
earlier has an efficient Quality of Service framework with different service flow and traffic
categorization. The approaches to the Quality of Service provisioning are however different in both
802.11e and 802.16. The MAC protocols of both the networks vary and so direct mapping of Quality of
Service frameworks of 802.11e and 802.16 are not possible. Therefore designing an integrated Quality of
Service is a challenging issue that is faced in interworking. Gakhar, Kamal et al (2005) proposed a
Quality of Service framework for 802.11e/802.16 interworking based on the mapping of Quality of
Service requirements of an application and also defined the necessary messaging procedures required for
the mapping. The authors propose an architecture which is a Radio Gateway (RG) which works as a SS
for 802.16 and Quality of Service AP for 802.11. A common service flow structure is designed which
combines the service flows of both 802.16 and 802.11e which makes mapping possible. The common
service flows are names C1-C4 which varies from CBR (Constant bit rate) with Real-Time Traffic (C1) to
VBR (Variable bit rate) with real time as well as precious (minimum packet loss data) data. Based on the
data it receives from specific network it maps the service flow from the networks to the common service
flow which is understood by both the networks and then transmits the data based on Quality of Service
requirements and policies. However a drawback of this solution is that initial setup time is required since
it involves buffering, mapping, setting up new connection, etc. Synchronization of data arrival should be
ensured in this proposed solution of using RG. Timely service of Quality of Service sensitive application
despite if traffic handling in RG should be ensured too.

    Fantacci and Tarchi (2006) proposed a bridging device capable of transparently interconnecting
both 802.11a and 802.16. The authors propose the solution based on two major goals which are traffic
priority and implementation issues. Based on the two goals there are two bridging solutions discussed.
The first solution is based on providing end to end Quality of Service independent of the wireless
technologies in the interworking network. The second solution reduces the implementation complexity
while sacrificing the Quality of Service. The main difference between the 802.11 and 802.16 protocol
stack is the absence of convergence sublayer (CS) or similar layer within 802.11 which handles traffic
class management. For this reason the CS cannot be connected directly to 802.11 MAC. Therefore the
authors propose using a MAC bridge that follows IEEE 802.1D/Q standards. The 802.1D/Q supports 8
priority levels as in 802.11e. The second solution proposed by the authors neglects the Quality of Service
requirements and in this solution the 802.16 network acts as a level 2 tunnel allowing the 802.11 areas to
be operating as a unique (single) network entity. In order to optimize bandwidth allocation, the payload

area of the PDU generated at the 802.11 network should be tightly packed. The payload length is
specified by 802.16 SS. The solution aims at fragmenting longer SDU and packs together the shorter
SDUs to optimize the PDU length. The impact of the solution is to optimize the performance by choosing
the optimal size of PDU at the MAC layer.

     Existing study of Quality of Service in integrated networks focused on designing Quality of Service
models. However, Ali Yahiya et al (2007) concentrate on the quality of service maintenance during
mobility or vertical handoff between the networks. The authors study the seamless support of ongoing
sessions based on a Quality of Service mapping model they design. They examine the performance of
802.16 in accepting the ongoing sessions and providing them the same quality of service that they had
when they were in their home network which is 802.11. They introduce the architecture for Interworking
and signaling protocols to perform the handoff process efficiently. The architecture defined is a loosely
coupling between 802.11 and 802.16 which means that the networks will be integrated at IP layer and
relies on IP protocols. In other words the integration takes place at the network layer. The authors use
bandwidth reservation policy in 802.16 which sets aside a dedicated amount of resource for the handoff
users in such a manner that the existing 802.16 SS users do not have to suffer loss of quality of service.
The authors state that vertical handoff from 802.11 to 802.16 requires to involve Quality of Service
mapping in both SS and BS. They define which from Quality of Service class each 802.11 is assigned to
which service flow in 802.16 and also provide mapping from traffic specification negotiated in 802.11 in
Action.ADDT request to DSA-REQ negotiated in 802.16. the Different Quality of Service parameters
such as packet loss rate, maximum throughput are studied for service flow sessions such as video, voice
and FTP. 802.16 can support 24 voice sessions and 21 video sessions seamless transferred from 802.11
without any packet drop. In the case of FTP it is observed that the throughput increases after the handoff
to 802.16 which is due to the contention free scheduling in 802.16. The simulation results indicate that,
with the mapping assigned by the author‟s, limited number of 802.11 video and voice sessions are
seamless handed off and supported by 802.16.

     Yahiya Ali et al., (2006) studies resource allocation in an integrated 802.11/802.16 network. The
authors also propose a resource allocation scheme for the efficient performance of the interworking
network. The policy based on threshold is used for managing resources between traffic of both types. Two
thresholds are used corresponding to direct and indirect traffic to 802.16 BS. The direct traffic is the
traffic from 802.16 users and indirect traffic is 802.11 users traffic. The architecture in this interworking
scenario is tight coupling which makes 802.11 traffic dependent on 802.11 BS for reaching the internet.
Therefore in this proposed solution for efficient management of resources, priority is given to 802.11
users over 802.16 users. There are several methods of bandwidth allocation which are complete sharing,
complete partitioning and a method which integrates both. The authors use the integrated method where
the resource is partitioned partially and the rest of the available resource is left for sharing. The resource
in this discussion is bandwidth in terms of channel. The users are blocked only when the respective
partitioned channels as well as the shared channels are over. The model is analyzed using Markov chains.
The results indicate that with variation in partitioning and load the blocking probability also varies. The
solutions do not clearly indicate that 802.11 is getting higher priority neither does it indicate the benefit of
using the two threshold scheme. However, it proposes a possible method of resource sharing.

   From various papers in the area of resource allocation in integrated 802.11/802.16, it can be seen that
mainly there are two methods used which are Markov chain models and game theory. Before we discuss

the papers that deal with game theory, it would be appropriate to give a brief overview of game theory, its
working and advantages.

    Game theory is a mathematical model used to study the interaction among different entities in a
wireless network. The major components in game theory modeling are player, strategy and payoff.
Different players will choose different strategies and the payoff shows the outcome of the game. The
players choose the strategies such that all players are satisfied. The game could be modeled as cooperative
or non-cooperative. The state where all the players are satisfied is referred to as equilibrium and the most
common equilibrium is called Nash equilibrium. Other than Nash equilibrium there are other solutions
such as min-max solution, Stackelberg equilibrium, etc. The advantage of using game theory formulation
over other optimization theory is that game theory aims at providing individual optimal solutions which is
a better model in situations where multiple entities interact with each other to achieve their interests.

     Niyato and Hossain (2007) in their study observed that the existing solutions for Quality of Service
provision based on resource allocation never satisfied the requirement of all individual users. Therefore,
the authors aimed at providing Quality of Service provision taking into consideration the satisfaction of
all individual users and because of the nature of the solution which the authors aimed to achieve, they use
game theoretic approach for modeling the solution. Quality of Service in this solution is based on
bandwidth management and admission control algorithms. The authors use a hierarchical model for the
bandwidth allocation. However the admission control uses utility as the basis for the control. The
hierarchical bandwidth allocation is done in two levels; the first level uses a bargaining game to allocate
resources to the SS and the WLAN APs‟ so that both of them are satisfied. Once the bandwidth between
WLAN AP and SS have been allocated satisfactorily then in the second level based on the bandwidth
allocated to the SS, a cooperative game through coalition is used to achieve fair sharing of the allocated
bandwidth between the different connections with SS such as UGS, PS, BE, etc. The admission control
for SS is however based on a utility function and calls are admitted only if the total utility increases. In
the case of WLAN the admission control is based estimated traffic load. If the traffic load is high and the
AP does not have the required bandwidth it places a request to the BS and the BS allocates a new
bandwidth for AP and if the resource requires still does not match the call is rejected. The solutions
obtained by the authors indicate that individual user Quality of Service is considered in this solution and
fair bandwidth allocation is provided among the connections in the integrated network.

     Niyato and Hossoin (2007) extended the idea discussed above to a different scenario of 802.11 in a
multihop 802.16 network where 802.16 mesh is used as a backhaul for 802.11 hotspots. 802.16 supports
both point to multipoint single hop as well as multihop mesh networking. The system model used was
802.16 as mesh routers would carry the traffic of the SS and WLAN APs‟ to the gateway router to
connect to the internet. The WLAN APs are called as edge routers and they have dual radios to
communicate between WLAN and 802.16. The resource allocation in this case was not just between
802.11 APs and 802.16 SS, in this scenario it also includes resource sharing with traffic from relay
connections. The authors use the bandwidth allocation and admission control schemes in this scenario to
get a fair allocation to the different traffic model. The traffic from the mesh connection is determined by
transmission rate which is obtained from the burst size and signal to noise ratio of channel. Whereas, the
traffic load from WLAN is measured using probability of success and probability of collision. The
bandwidth allocation again uses bargaining game strategy. The bargaining game provides a fair and
efficient solution due to Pareto optimality and fairness is achieved by obtaining the equilibrium. The

admission control as mentioned earlier is based on utility function such that the packets are admitted only
if the total utility increases. The results show that for fairness the bargaining game has to adjust the burst
size accordingly and WLAN is given priority over 802.16 SSs‟.

    Both the works mentioned above only concentrate on bandwidth allocation and admission control.
They do not deal with pricing issues in an integrated 802.11/802.16 environment. Here the 802.16 BS
charges the 802.11 APs for sharing the licensed band of 802.16. However later, Niyato and Hossoin
(2007) model a pricing strategy for the integrated networks. To design a pricing model in an
802.16/802.11 integrated environment the resource demand should be available because the resource
requirement may change based on price changes. Also the Quality of Service requirements for the 802.16
SS should also be considered. Since the 802.16 SS has a fixed bandwidth demand, they subscribe to
802.16 BS with a flat rate. However, 802.11 network has an varying demand. The solution is obtained
using the Stackelberg equilibrium. This equilibrium is a leader follower algorithm where the leader uses
the followers best response information to provide optimal supply quantity so as to gain highest profit.
Profit of 802.16 is maximized and 802.11 routers are satisfied with the bandwidth sharing and pricing.
The solution is easy if the bandwidth requirement and 802.11 response are all available before hand.
Since this is not the case in practical situations, the authors use genetic algorithm to study a pattern of
bandwidth demand at 802.16 BS and 802.11 APs. Revenue from the SS is based on Quality of Service
and revenue from 802.11 APs is based on bandwidth allocation. Numerical results indicate that at
equilibrium 802.16 BS charges the same price for every 802.11 router even though the bandwidth
demands differ and when traffic arrival increases the 802.16 BS increases the price on 802.11 APs to
compensate for the loss in revenue by degradation of quality of service to SSs. The solution obtains a
pricing model with 802.16 BS gaining maximum profit while the 802.11 APs are also kept satisfied.

    Ibanez et al (2008), states in their work that the growth of WLAN promotes the development of
multihop routing protocols. AODV is one such protocol that is mentioned by the authors. AODV adapts
quickly to the dynamic conditions, processes data rapidly, has low network utilization and uses unicast
route mechanism that makes data transmission easy between source and destination. However, in an
802.11/802.16 integrated network the existing AODV mechanism does not work since it involves
changing networks along with the multihop route discovery and delivery. The authors modify the existing
AODV to accommodate the integration of 802.11 and 802.16. The algorithm sends out a request packet
until it can reach a destination node or the 802.11/802.16 gateway. If the destination is outside the
coverage of 802.11 network, the AP retransmits the request packet. Once the 802.11/802.16 gateway or
the destination node receives the packet it sends a reply packet to the source which helps in route

    It has been observed from the study of research issues in the area of Interworking of 802.11 and
802.16 that the main problems in an integrated 802.11/802.16 network are:

1) Protocol adaptation

2) Quality of Service provision

3) Resource allocation

4) Pricing

            [This reference is just for the report shown above arranged in alphabetical order]

1.    Ali-Yahiya, T. and Sethom, K. and Pujolle, G.," Seamless Continuity of Service across WLAN and
      WMAN Networks: Challenges and Performance Evaluation," 2nd IEEE/IFIP International
      Workshop on Broadband Convergence Networks, 2007.
2.    Dusit Niyato and Ekram Hossain, "A Hierarchical Model for Bandwidth Management and
      Admission Control in Integrated IEEE 802.16/802.11 Wireless Networks," IEEE WCNC, 2007.
3.    D Niyato, E Hossain, "Integration of WiMAX and WiFi: Optimal Pricing for Bandwidth Sharing,"
      IEEE communications Magazine, 2007.
4.    Dusit Niyato and Ekram Hossain, "Integration of IEEE 802.11 WLANs with IEEE 802.16-Based
      Multihop Infrastructure Mesh/Relay Networks: A Game-theoretic Approach to Radio Resource
      Management," IEEE NETWORK, 2007.
5.    HT Lin, YY Lin, WR Chang, RS Cheng, "An Integrated WiMAX/WiFi Architecture with QoS
      Consistency over Broadband Wireless Networks," 6th IEEE Consumer Communications and
      Networking Conference, CCNC, 2009.
6.    Ibanez, S.R.; Santos, R.A.; Licea, V.R.; Block, A.E.; Ruiz, M.A.G, "Hybrid WiFi-WiMAX
      Network Routing Protocol," Electronics, Robotics and Automotive Mechanics Conference, 2008.
      IEEE CERMA 2008, pp. 87-92.
7.    J. Nie, X. He, Z. Zhou, and C. Zhao, “Communication with bandwidth optimization in IEEE 802.16
      and IEEE 802.11 hybrid networks,” in Proc. IEEE ISCIT 2005, vol. 1, Oct. 2005, pp. 26–29.
8.    K. Gakhar, A. Gravey, A. Leroy, “IROISE: A New QoS Architecture for IEEE 802.16 and IEEE
      802.11e Interworking”, 2nd IEEE/Create-Net International Workshop on Deployment Models and
      First/Last Mile Networking Technologies for Braodband Community Networks, IEEE
      BROADNET, Oct. 2005.
9.    L Berlemann, C Hoymann, GR Hiertz, S Mangold, "Coexistence and interworking of IEEE 802.16
      and IEEE 802.11 (e)," IEEE 63rd Vehicular Technology Conference, IEEE VTC, 2006.
10.   L. Berlemann, C. Hoymann, G. Hiertz, and B. Walke, “Unlicensed Operation of IEEE 802.16:
      Coexistence with 802.11(A) in Shared Frequency Bands,” in Proc. IEEE International Symposium
      on Personal, Indoor and Mobile Radio Communications, IEEE PIMRC, Sep. 2006.
11.   Motorola and Intel, “ WiMax and WiFi Together: Deployment models and user scenario,” White
      paper, 2007.
12.   NJ Thomas, MJ Willis and KH Craig, "Analysis of co-existence between IEEE 802.11 and IEEE
      802.16 systems," IEEE SECON, 2006.
13.   Romano Fantacci and Daniele Tarchi, “Bridging Solutions for a heterogeneous WiMAX-WiFi
      Scenario,” Journal of Communications and Networks, Vol. 8, No. 4, December 2006.
14.   T. Ali Yahiya, H. Chaouchi, A. L. Beylot, G. Pujolle, Threshold Based WiMax Resource
      Reservation, IEEE mobility, 2006.
15.   V. Gunasekaran and F. Harmantzis, "Financial Assessment of Citywide Wi-Fi/WiMAX
      Deployment," Communications & stratégies(Montpellier), pp. 131-153, 2006.
16.   Xiangpeng Jing, Siun-Chuon Mau,D. Raychaudhuri and Robert Matyas, “Reactive Cognitive Radio
      Algorithms for coexistence between IEEE 802.11b and 802.16a Networks," IEEE GLOBECOMM,
17.   X Jing, D Raychaudhuri, "Spectrum co-existence of IEEE 802.11 b and 802.16 a networks using
      the CSCC etiquette protocol," IEEE DYSPAN, 2005.


To top