The document discusses code assignment algorithms for OVSF codes used in WCDMA networks. It proposes a new code assignment algorithm for NOVSF (Nonblocking OVSF) codes. The key points are: 1) OVSF codes have a blocking property that limits bandwidth utilization. Code reassignments can improve utilization but introduce delays and overhead. 2) NOVSF codes share OVSF codes with spreading factor 8 across time slots, removing the blocking property. 3) The proposed NOVSF code assignment algorithm assigns time slots without requiring any code reassignments, improving throughput over algorithms for OVSF codes that use reassignments.

1. Telecommunication Systems 25:3,4, 417–431, 2004
 2004 Kluwer Academic Publishers. Manufactured in The Netherlands.
A Code Assignment Algorithm for Nonblocking OVSF
Codes in WCDMA
KIRAN VADDE and HASAN ÇAM e-mail: hasan.cam@asu.edu
Department of Computer Science and Engineering, Arizona State University, Tempe, AZ 85287, USA
Received 11 October 2003
Abstract. OVSF codes are used as channelization codes in WCDMA. Due to code blocking property of
OVSF codes, the bandwidth available in the system is severely limited. Code reassignments mitigate the
impact of the blocking property at the expense of causing delays and decreasing the throughput of the
system. Nonblocking OVSF (NOVSF) codes have been proposed to alleviate the adverse effect of code
reassignments. This paper presents a code assignment algorithm for NOVSF codes, which does not require
any code reassignments. Simulation results show that NOVSF codes achieve better throughput than OVSF
codes, even though code reassignments are allowed in the assignments of OVSF codes.
Keywords: OVSF codes, NOVSF codes, WDCMA, code reassignments, code blocking, code–slot assign-
ment
Introduction
Third generation (3G) wireless networks such as WCDMA aim to provide multimedia
services for mobile users anywhere and at any time. These networks need to transmit not
only voice but also images, videos and multimedia data. They are required to support
bursty traffic, which is significantly different from voice traffic carried in the existing
second-generation wireless systems. To support a variety of multimedia applications,
the system must support variable transmission rates for different users [Adachi et al., 2].
Orthogonal Variable Spreading Factor (OVSF) codes are used as channelization
codes in WCDMA for data spreading on both downlink and uplink [Adachi et al., 1].
OVSF codes determine the data rates allocated to users because data rates are a func-
tion of spreading factors of OVSF codes. By varying the spreading factor, i.e., OVSF
code assigned, different data rates can be supported. WCDMA supports data rates up to
2.048 Mbps in 5 MHz bandwidth using variable spreading factors. Because OVSF codes
require a single RAKE combiner at the receiver, they are preferable to multiples of or-
thogonal constant spreading factor codes which need multiple RAKE combiners at the
receiver. When a particular OVSF code is allocated to a requesting user, its descendant
and ancestor codes cannot be used simultaneously because their encoded sequences be-
come indistinguishable. Consequently, the code blocking property of OVSF codes leads
to poor utilization of network capacity.

2. 418 VADDE AND ÇAM
Code reassignments are proposed in [Minn and Siu, 13] to improve the utilization
of OVSF codes. As the code assignments and releases are made, the OVSF code tree
gets fragmented. Due to this, even if the system capacity is available, an arriving request
may not be granted a code due to already made code assignments. In such cases, a new
request can be assigned a code by reassigning some of the assigned codes. Code reas-
signments improve the system throughput by reducing the code wastage. However, code
reassignments cause significant delays and overhead in code assignment, as the number
of requests, which are to be reassigned, may be large. All the receivers involved in code
reassignments should be informed about the new code assignment, which consumes both
the system time and bandwidth.
Nonblocking OVSF (NOVSF) codes [Çam, 4, 5] are introduced to remove code-
blocking property. The basic idea behind NOVSF codes is to share those OVSF codes
with spreading factor 8 in time. Each of these codes has 64 time slots, one or more
of which can be assigned to a channel. These codes achieve better throughput and re-
sult in good spectral efficiency. Many improvements can be made to the current code
assignment algorithms as the blocking property is removed.
In this paper, we propose a new code assignment scheme for NOVSF codes, which
does not use any code reassignments. The algorithm assigns the required number of
slots to the arriving request based on the data rate requested. Performance of the pro-
posed algorithm is compared with the dynamic code reassignment algorithm of OVSF
codes. Simulation results indicate that the throughput obtained when NOVSF codes are
employed is greater than that when OVSF codes are used.
The remainder of this paper is organized as follows. Section 1 discusses some of
the code assignment schemes proposed in the literature. Section 2 describes drawbacks
of OVSF codes. Section 3 discusses the construction of NOVSF codes and code assign-
ment algorithms for dedicated channels (DCH) and Downlink Shared Channel (DSCH).
Performance of NOVSF code assignment algorithm for DCH and OVSF code assign-
ment algorithm is compared in section 4. Concluding remarks are made in section 5.
1. Related work
Dynamic code assignment schemes have the ability to enhance statistical multiplexing
and spectral efficiency of WCDMA systems. Code assignment schemes determine the
best way to allocate codes to channels. Because the OVSF codes are valuable resources
in WCDMA-based system, they should be properly managed to support as many users
with different requirements as possible using cost-efficient algorithms.
Various code assignment schemes are introduced in [Minn and Siu, 13; Kam
et al., 11; Cheng and Lin, 8; Fanatacci and Nannicini, 9; Tsaur and Lee, 16; Assarut
et al., 3], which require some degree of code reassignments to allocate code to a new
arriving request. The algorithm in [Cheng and Lin, 8] assigns code to low data rate
users in a manner that maximizes the available number of low SF codes corresponding
to high data rate codes. They assume that a user can be assigned multiple OVSF codes,
increasing the complexity of the receiver since multiple RAKE receivers are required.

3. OVSF CODES IN WCDMA 419
In [Minn and Siu, 13], code-blocking problem is mitigated by reassigning existing
users to new codes in a manner to maximize the available number of low SF codes,
without addressing the support for different types of traffic. This algorithm may lead
to a chain of reassignments, resulting in a lot of overhead for informing receivers about
the change of code assignments. On the other hand, the code assignment algorithm
in [Fanatacci and Nannicini, 9] supports both real-time traffic and non-real time traffic,
without addressing the problem of efficiently sharing OVSF codes between a number of
bursty traffic users.
Another algorithm [Kam et al., 11] shares bandwidth between bursty traffic sources
with different QoS requirements by dynamically changing the spreading code and band-
width at the cost of increased complexity. Tsuar and Lee proposed Forest for OVSF-
Sequence-Set-Inducing Lineages (FOSSIL) codes [Tsaur and Lee, 16]. FOSSIL codes
are constructed from OVSF code tree, and are mainly used to achieve blind rate detec-
tion. Each user is assigned a set of codes rather than a single code and a code is selected
from the set assigned, based on the data required at that instant. The advantages of this
approach are that stations can adjust their symbol rate without the need for a separate
control channel and reliable detection of the symbol rate without exchanging any rate
messages. The main goal of this approach is to achieve blind rate detection using OVSF
codes, so the algorithm does not address the issue of maximum throughput.
Assarut et al. [3] proposed Region Division Assignment (RDA) algorithm. The
main objective of this algorithm is to provide services for all supported rates. In order to
achieve this, RDA divides the OVSF code tree into regions for each supported rate. RDA
does not consider code reassignments. When a new request arrives, even if the capacity
is available in the region corresponding to the rate requested, if the code corresponding
to the request is not available, the request is rejected. This leads to lower utilization of
the network capacity.
In [Çam and Vadde, 7], the performance of OVSF code reassignment technique
is compared with a preliminary code assignment technique for NOVSF codes, and the
simulation results are very encouraging for NOVSF codes.
2. OVSF codes
Due to the code blocking property of OVSF codes, system capacity is reduced up to
25–30% [Minn and Siu, 13]. Hence, code reassignments are needed to maximize the
system throughput. However, code reassignments do not come for free, there are some
overheads involved in the code reassignments process. The delays caused by code reas-
signments are discussed in detail in [Goria et al., 10].
According to [Goria et al., 10], signaling radio bearers are used to inform the all
receivers, which are tuned to the codes of same layer. Signaling Radio Bearers (SRB)
are used for other purposes too. Other functions of SRBs include handover, negotiation
and renegotiation of QoS parameters. The mapping of these SRBs onto the physical
channels is implementation dependent [15].

4. 420 VADDE AND ÇAM
In order to reduce the signaling overhead ARROWS [Goria et al., 10] uses four
radio signaling bearers with combined aggregate data rate of 3.4 Kb/s. If these bearers
are actually mapped on to the data channels, the overhead will be much more. So, to
minimize the signaling overhead, separate channel is used for signaling radio bearers.
This also permits the sending of reassignments information layer by layer.
Since these signaling bearers are separate and serve different purposes, they are
very busy and may not be available always, in such case, the code reassignment proce-
dure has to wait until the bearers are available. This delay can be in the order of seconds,
after which the actual code reassignments take some more time, totaling some time,
which is long enough to interrupt the existing call/service. In order to avoid this inter-
ruption, it is better not to perform the code reassignments, in which case, the arriving
request is dropped.
The actual probability of signaling radio bearers to be busy depends on the load
on the network. If the network is heavily loaded, then more number of handovers and
service negotiations occur, making the probability of SRBs being free less. On the other
hand, if the network is lightly loaded, then the probability of SRBs being idle is relatively
high. Determining the exact probability of SRBs being idle requires a lot of statistical
data and analysis. Current literature on the subject does not provide this information,
due to the obvious variations in the actual traffic and network implementations. In this
paper, the performance of OVSF code assignment algorithm is analyzed by running the
simulations for different values of SRB free probability, which gives an idea of the out
performance of the NOVSF codes without loss of generality. Simulations are run by
fixing the probability of SRBs being idle as 0.90, 0.75, 0.50 and 0.25.
Drawbacks of OVSF codes are:
1. Code reassignments. Dynamic OVSF code reassignment algorithm performs as
many reassignments as needed to accommodate the new arriving request. Code
reassignments do not come for free, as they have the overhead of informing all the
receivers involved in code reassignments about the new code assignment.
2. Request processing delay. As the code reassignment procedure takes some time to
complete, some of the requests which come during the reassignment procedure of
the previous request have to be queued. Due to this queuing, processing of some
requests is delayed.
3. Blocking of services. Code reassignment information is sent on Signaling Radio
Bearers (SRBs) that are used for other functions like negotiation of QoS parame-
ters, call handover, so they may not be available when code reassignments are
needed. This may lead some calls to be dropped, which increases the number of
call rejects i.e. blocking probability and decreases the system throughput. In ad-
dition, while SRBs serve code reassignments, other services that need SRBs are
blocked.

5. OVSF CODES IN WCDMA 421
3. Code and slot assignment algorithm for NOVSF codes
NOVSF codes are discussed in [Çam, 4, 5]. The main idea behind NOVSF codes is as
follows. As shown in figure 1, there are initially eight orthogonal codes with SF 8 each,
namely, A, B, C, D, E, F, G and H. Each of the 64 time slots indeed corresponds to
one 8-bit chip sequence. Hence there are 64 eight-bit chip sequences in 64 time slots,
resulting in a total of 64 · 8 = 512 chips. The data rate supported by each time slot is
equivalent to the data rate that an OVSF code with SF of 512 supports. If the data rate
supported by a time slot is denoted by R, then the data rate supported by X time slots
that are allocated to a user equals X · R. Note that the time slots that are allocated to
a user need not be consecutive. The arriving requests are assigned a number of time
slots of the eight codes corresponding to the data rate requested by the arriving request.
Depending on the data rate requested, which corresponds to SF 512, 256, 128 or 64, 1 or
2 or 4 or 8 time slots are assigned. Time multiplexing provides flexibility in supporting
many different data rates supported by the codes. We can achieve intermediate data rates
by varying the number of slots assigned to a request. In standard OVSF code tree, there
are eight branches of code trees when layer 2 is considered. The reassignments within
these branches are not necessary in NOVSF codes case, as the branches are flattened into
time slots. Hence, the number of reassignments is minimized by using NOVSF codes.
Downlink Shared Channel (DSCH) is introduced in WCDMA to remove the
wastage of codes for bursty data sources. DSCH is assigned a branch of the OVSF
code tree, codes of which are shared by number of bursty users in time and code. DSCH
improves the system throughput by reducing the number of idle codes. As per the current
proposal, the branch of OVSF code tree assigned to DSCH is fixed. On the other hand,
if a data source is not bursty and has certain quality of service (QoS) requirements, it is
Figure 1. NOVSF code tree.

6. 422 VADDE AND ÇAM
assigned a dedicated channel (DCH). In DCH, a OVSF code is reserved for the entire
duration of the data source transmission. When NOVSF codes are employed, branch
translates to bunch of time slots. A bunch of time slots of a NOVSF code are assigned
to DSCH and these time slots are shared in time by certain number of bursty sources.
First, we present the NOVSF code assignment algorithm for dedicated channels (DCH)
and then we present the DSCH slot allocation algorithm.
3.1. Code/slot assignment algorithm for dedicated channels (DCH)
Code/slot assignment algorithm assigns time slots of one or more NOVSF codes to arriv-
ing requests. The code assignment algorithm calculates the required number of free slots
based on the spreading factor requested. This algorithm does not do code reassignments
in the process of code and slot assignment. The flowchart of the algorithm is illustrated
in figures 2 and 3 describes the assignment algorithm. The algorithm initially checks
if the arriving request has traffic class belonging to conversational traffic (class 1) and
streaming traffic (class 2), which require dedicated channels. If the traffic of the request
belongs to traffic classes 3 and 4, the request is assigned to DSCH, as the nature of the
traffic is bursty and DSCH schedules this traffic according to their deadlines. The al-
gorithm assumes that DSCH is assigned a particular NOVSF code time slots, which are
shared among number of bursty traffic users. The algorithm only assigns NOVSF code
slots for requests, which require dedicated channels.
The call is granted the free slots if the total number of free slots in any code is
greater than the required number of free slots. If the search for the required number
of idle slots in one code fails, the algorithms searches if there are required number of
non-overlapping idle slots in multiple codes. In case, the search succeeds, these non-
overlapping time slots are assigned to the request.
The time slots have to be non-overlapping with respect to the index of the time
slot. This is required as we assume that the transmitter does not have the capability of
transmitting with multiple codes simultaneously. If the above searches fail, the call is
rejected.
The code assignment algorithm never reassigns a code in the dedicated channels,
saving a lot of bandwidth and time.
In the flow chart of the algorithm shown in figure 2, S is the number of slots re-
quired to satisfy the request and SF is Spreading Factor. System capacity is checked to
see if the requested rate can be granted. Depending on the available capacity, the request
is either rejected or further processed. If sufficient capacity is available, the algorithm
computes the number of time slots needed to satisfy the rate request using the following
formula.
Number of slots required to satisfy the request(S) = 512/(requested SF).
As each NOVSF code has 64 time slots, the above calculation uses 512 in the
numerator.
After this calculation step, all the NOVSF codes are checked for vacant time slots.
If the required number of slots is free in more than one code, then, from these codes,

7. OVSF CODES IN WCDMA 423
Figure 2. Flow chart of the assignment algorithm.

8. 424 VADDE AND ÇAM
Algorithm CSA
Begin
Calculate the number of slots required; S = 512/(SF corresponding to data rate requested)
If System capacity available
If any NOVSF code has required number of free slots
Select best fit NOVSF code and assign the slots to the request
Else If there are non-overlapping time slots from more than one NOVSF code
Allocate non-overlapping time slots to the request
Else
Reject the request
Else Reject the request
End
Figure 3. NOVSF code/slot assignment algorithm.
Algorithm DSCHDSA
Begin
Form ready flow set from the flow queues using scheduling policy
Calculate the number of NOVSF slots needed for each flow in the ready flow set
Do
Select the next ready flow from the ready flow set
If the number of free NOVSF slots in the current schedule time T >= required number
of time slots of the selected flow
Allocate required NOVSF free slots to the selected flow
Delete the flow from ready flow set
Else
Allocate the remaining free NOVSF time slots in the current DSCH frame to the selected
flow
Make the current flow ready for next schedule time T
Adjust the supported data rate
While there are free NOVSF time slots and ready flow set not empty
End
Figure 4. DSCH assignment algorithm.
the code with the minimum number of free slots available is chosen and allocated to the
request. This step is done in order to get more space at some codes, so that high rate
requests can be satisfied with the more spaced codes. In other words, the best fit for the
request is selected and assigned rather than assigning the first fit. The slots allocated
need not be consecutive. The receiver is informed about the slot assignment during the
call setup procedure. It then decodes the data from only the assigned time slots.
3.2. Slot assignment algorithm for DSCH
Figures 4 and 5 show the assignment algorithm and flowchart for allocation of DSCH
NOVSF slots to traffic flows. The assignment algorithm is executed for each DSCH
schedule time T of base station. There are 320 NOVSF time slots in each DSCH

9. OVSF CODES IN WCDMA 425
Figure 5. Flow chart of the DSCH assignment algorithm.

10. 426 VADDE AND ÇAM
frame. 2560 chips can be transmitted in each DSCH frame [14]. NOVSF code has
64 time slots and each time slot corresponds to 8 chips, therefore each NOVSF code
can transmit 64 · 8 = 512 chips; 5 · 512 = 2560 chips. Therefore, each DSCH cor-
responds to duration of 5 NOVSF codes. 5 · 64 = 320 time slots. Therefore, we
have 320 NOVSF time slots in each DSCH frame. The assignment algorithm assigns
these 320 time slots to flows in DSCH. We assume that a scheduling policy, which
determines the next flow to be served and the number of packets of that flow to be
transmitted, is already in place. There are many scheduling algorithms in the litera-
ture [Malik and Zeghlache, 12; Çam and Challa, 6], a designer can choose one of these
policies based on the specific implementation. At the start of the current schedule time T ,
the scheduling policy selects the ready flow set and the amount of data of each flow in
the ready flow set that needs to be transmitted to meet the respective flow service re-
quirements. The code/slot assignment algorithm, then selects each flow in the ready
flow set and allocates the time slots of DSCH NOVSF code. The algorithm calculates
the number of time slots needed to transmit the data of the selected flow and checks if
there are required number of free time slots in the NOVSF code reserved for DSCH. If
yes, the algorithm assigns the required number of free time slots to the flow and deletes
the flow from ready flow set. NOVSF codes provide the flexibility of assigning variable
number of NOVSF time slots based on the data rate required. If there are any unassigned
or free NOVSF slots after the current flow has been assigned its share of time slots, the
algorithm checks if there is any flow in the ready flow set and assigns the free slots to
the flow accordingly. If a flow in the ready flow set requires more time slots than that are
available, the algorithm assigns the available free time slots and makes this unsatisfied
flow ready in the next schedule time so that it gets a chance to transmit rest of its packets
in the next DSCH transmission frame.
4. Performance analysis of OVSF and NOVSF codes
In this section, we present the comparison analysis results obtained by simulation of
the NOVSF code assignment algorithm for DCH and OVSF dynamic code reassignment
algorithm proposed in [Minn and Siu, 13]. Simulations have been performed in OPNET
to compare the throughput difference and the number of code rejects in the system when
OVSF codes are employed and NOVSF codes are deployed. We consider a cellular
network with hexagonal cells. Each cell is serviced by a node-B which generates code
requests. The simulation uses the following parameters. Call arrival process is Poisson
with mean arrival rate λ = 1–100 calls/s. Call duration is exponentially distributed
with a mean value of 1/µ = 0.25 s. Maximum spreading factor – 512 and minimum
spreading factor 8. Probability of SRBs being idle values 0.90, 0.75, 0.50 and 0.25.
Parameters evaluated are throughput, number of code reassignments and number of code
rejects.
Figure 6 shows throughputs of NOVSF and OVSF codes when the proposed algo-
rithm and code reassignments are used in NOVSF and OVSF codes, respectively. Note
that the throughputs of OVSF codes are obtained for different bearer free probabilities.

11. OVSF CODES IN WCDMA 427
Figure 6. Comparison of throughputs between OVSF with various SRB free probabilities and NOVSF.
As seen from the above figure, the throughput obtained is higher in NOVSF codes. The
difference between the throughput obtained in OVSF codes, when the bearer free proba-
bility is 0.90 and that obtained in NOVSF codes assignment algorithm is less. In OVSF
codes case, when the bearer free probability is 0.90, the number of cases where SRBs
are not free is very less, so it performs almost all of the code reassignments, thereby ob-
taining throughput nearly equal to that of NOVSF codes. Since SRBs are used heavily
for other functions as mentioned earlier, the assumption of bearer free probability to be
0.90 is not very realistic. The cellular cell must be very lightly loaded in order for bearer
free probability value to be this high, which will not be the case in real world scenarios.
In this particular case, even though the throughput gain in NOVSF codes is less, since it
does not use SRBs, other services will not be blocked, which is not the case with OVSF
codes.
When the bearer free probability is 0.75, for OVSF codes, the throughput obtained
in NOVSF codes is greater than that of OVSF codes at all loads. As seen from the above
figure, as the bearer free probability decreased to 0.50, the throughput obtained using
NOVSF codes is higher than that in OVSF codes at all loads. In OVSF codes case,
since, half of the time, signaling radio bearers will not be available, requests will be
rejected more often, and hence the throughput of the system drops.

12. 428 VADDE AND ÇAM
Figure 7. Comparison of number of call rejects between OVSF with various SRB free probabilities and
NOVSF.
The throughput obtained in NOVSF codes is very much higher than that obtained
using OVSF codes, when the bearer free probability is 0.25. When the cell becomes
busy, i.e., as more number of users enter the cell, signaling radio bearers get busier
and busier. When code reassignments are needed in OVSF codes, the SRBs will not
available, hence more calls are dropped, thereby decreasing the system throughput.
Figure 7 shows the comparison graph of the number of call rejects between OVSF
dynamic code reassignment algorithm at various bearer free probabilities and NOVSF
code assignment algorithm.
As seen from the figure, the number of call rejects for OVSF code assignment
algorithm and NOVSF code assignment algorithm are not the same. At all loads, the
number of call rejects in OVSF code assignment algorithm is greater than that in NOVSF
code assignment algorithm.
As the SRB free probabilities decrease, for OVSF codes, the number of code rejects
increase, since SRBs are not available to perform code reassignments, many requests are
rejected.
When SRB free probability is 0.90, even though the number of code rejects in
OVSF codes is higher than NOVSF codes, the throughput obtained is nearly equal as
seen in figure 6. The reason for this is that OVSF code assignment algorithm performs as
many code reassignments as needed, hence high data requests starve low data requests.

13. OVSF CODES IN WCDMA 429
Figure 8. Comparison of number of code reassignments in OVSF with various SRB free probabilities.
Due to this, many small data rate requests are rejected, whereas NOVSF codes reject
some high data requests. Hence, even if the number of call rejects is different, the
throughput obtained in both the algorithms is nearly equal. NOVSF code assignment
algorithm, allows a proper mix of the data rates.
For SRB free probabilities 0.75, 0.50 and 0.25, OVSF code assignment algorithm
rejects many requests due to unavailability of signaling radio bearers, hence the through-
put drops. On the other hand, since NOVSF code assignment algorithm does not use sig-
naling radio bearers, the SRB free probability does not affect the number of code rejects
and the overall throughput.
Figure 8 shows the graph for number of code reassignments for OVSF code as-
signment algorithm. NOVSF code assignment algorithm does not reassign any codes of
already existing requests to satisfy new arriving requests. Hence, there will not be any
code reassignments when NOVSF codes are deployed in the system. As seen from the
figure, for SRB free probabilities 0.75 and 0.90, the number of code reassignments for
OVSF code reassignment algorithm increase with the load. As the load in the system
increases, the base station has to perform more number of code reassignments to accom-
modate the arriving requests. Since SRB free probabilities are a bit high, the base station
finds SRBs free more often, so performs code reassignments more frequently. Hence,
the number of code reassignments increase with the load.

14. 430 VADDE AND ÇAM
For bearer free probabilities 0.5 and 0.25, since the base station does not find the
SRBs free most of the time, the number of successful requests will be less, hence, the
increase in number of code reassignments does not change much with the load.
As seen in the simulation results, the throughput obtained in NOVSF code as-
signment algorithm is more than that obtained using OVSF code assignment algorithm.
Simulations results are shown for four different SRB free probabilities. As the SRB free
probability decreases, NOVSF codes outperform OVSF codes. The number of call re-
jects in OVSF code assignment algorithm is greater than that in NOVSF code assignment
algorithms, which decreases the throughput obtained. The number of reassignments in
OVSF code assignment algorithm depends on the SRB free probability and load in the
system. As OVSF dynamic code reassignment algorithm performs as many code reas-
signments as needed, it leads to an overhead in time and bandwidth consumption. The
incoming requests may have to be queued, as the RNC is busy performing code reas-
signments for the previous request.
5. Conclusions
Spreading codes are a scarce resource in WCDMA systems and must be properly man-
aged. OVSF code assignment scheme with code reassignments reduces the system re-
sources wastage and utilizes the system capacity in an effective manner. However, code
reassignments cause a lot of overhead, as all the receivers of the previously assigned
codes need to be informed about the code reassignment, which takes considerable sys-
tem time and bandwidth. Code assignment algorithm using NOVSF codes is presented
to improve the system throughput by not allowing any code reassignments in the code
assignment scheme. Simulation results show that NOVSF codes achieve higher through-
put than that achieved by OVSF codes employing code reassignments.
References
[1] F. Adachi, M. Sawahashi and K. Okawa, Tree-structured generation of orthogonal spreading codes
with different lengths for forward link of DS-CDMA mobile radio, IEE Electronics Letters 33(1)
(1997) 27–28.
[2] F. Adachi, M. Sawahashi and H. Suda, Wideband DS-CDMA for next generation mobile communi-
cations systems, IEEE Communications Magazine 36(9) (1998) 56–69.
[3] R. Assarut, U. Yamamoto and Y. Onozato, Region division assignment of orthogonal variable-
spreading-factor codes in W-CDMA, in: IEEE Vehicular Technology Conf., Vol. 3, October 2001,
pp. 1884–1888.
[4] H. Çam, Nonblocking OVSF codes for 3G wireless and beyond systems, in: Proc. of the Internat.
Conf. on 3rd Generation Wireless and Beyond’2002, June 2002, San Francisco, USA, 2002.
[5] H. Çam, Nonblocking OVSF codes and enhancing network capacity for 3G wireless and beyond
systems, Computer Communications, to appear.
[6] H. Çam and N. Challa, Supporting rate and soft delay guarantees for data traffic in the downlink
shared channel of WCDMA, in: Proc. of the Internat. Conf. on Wireless Networks (ICWN)’2002, Las
Vegas, USA, June 2002.

15. OVSF CODES IN WCDMA 431
[7] H. Çam and K. Vadde, Performance analysis of nonblocking OVSF codes in WCDMA, in: Proc. of
the Internat. Conf. on Wireless Networks (ICWN)’2002, Las Vegas, USA, June 2002.
[8] R.G. Cheng and P. Lin, OVSF code channel assignment for IMT-2000, in: IEEE Vehicular Technology
Conf., Vol. 3, May 2000, pp. 2188–2192.
[9] R. Fanatacci and S. Nannicini, Multiple access protocol for integration of variable bit rate multimedia
traffic in UMTS/IMT-2000 based on wideband CDMA, IEEE Journal on Selected Areas in Commu-
nications 18(8) (2000) 1441–1454.
[10] P. Goria, C. Guerrini and A. Vaillant, Signaling delay of code allocation strategies, in: 1st Mobile and
Wireless Telecommunication Summit, 2002.
[11] A.C. Kam, T. Minn and K.Y. Siu, Supporting rate guarantee and fair access for bursty data traffic in
W-CDMA, IEEE Journal on Selected Areas in Communications 19(11) (2001) 2121–2130.
[12] S. Malik and D. Zeghlache, Improving throughput and fairness on the downlink shared channel in
UMTS WCDMA networks, in: European Wireless’2002, Florence, Italy, February 2002.
[13] T. Minn and K.Y. Siu, Dynamic assignment of orthogonal variable-spreading factor codes in
WCDMA, IEEE Journal on Selected Areas in Communications 18(8) (2000) 1441–1454.
[14] Physical channels and mapping of transport channels onto physical channels (FDD), Third Genera-
tion Partnership Project Technical Specification Group Radio Access Network, TS 25.211 V3.12.0
(September 2002).
[15] Radio link control/medium access control (RLC/MAC), Third Generation Partnership Project Tech-
nical Specification Group Radio Access Network, TS 144.160 V5.0.0 (July 2002).
[16] L. Tsaur and C. Lee, Symbol rate adaptation and blind rate detection using FOSSIL (forest for
OVSF-sequence-set-inducing lineages), in: IEEE Internat. Conf. on Communications, Vol. 6, 2001,
pp. 1754–1759.