A REAL TIME TRAFFIC ENGINEERING SCHEME FOR

BROADBAND CONVERGENCE NETWORK(BCN)

Hwa-Jong Kim, Myoung-Soon Jeong and Jong-Won Kim

Department of electronics & Computer Engineering, Post-BK21 U-Home CI Team

Kangwon National University, 192-1 hyoja 2-Dong Chuncheon-Si Kangwon-Do, Korea

Keywords: Broadband Convergence Network (BCN), Next Generation Network (NGN), Traffic Engineering, real-time

monitoring.

Abstract: Recently, Broadband Convergence Network (BcN), a Korean version of Next Generation Network (NGN),

was introduced to guarantee pre-defined QoS for high speed multimedia service. The BcN is considering

charging users for premium services. For the BcN to be successfully diffused, however, a practical traffic

engineering (TE) tool is required because the BcN is composed of many kinds of subnetworks, and real time

feedback (traffic control) would be mandatory for the premium service. In the paper, a new TE scheme for

BcN, Rule Based Capture(RBC)/User Satisfaction Parameter(USP), is proposed to resolve the latent

problems of the BcN TE. The USP is designed to be admitted by many BcN subnetworks as a common

intermediate description of service quality instead of conventional QoS parameters. The RBC is introduced

to settle the real time TE issue against the vast accumulation of traffic monitoring data. The pilot RBC/USP

is implemented on the Linux platform and its performance is investigated. We found that the average traffic

log size is reduced to 0.058% for FTP, and 2.39% for streaming service by using the RBC/USP.

1 INTRODUCTION

Be advised that papers in a technically unsuitable

form will be returned for retyping. After returned the

manuscript must be appropriately modified. The

Broadband Convergence Network (BcN) was

introduced to provide high speed multimedia service

in Korea [1]-[6]. The BcN’s aim is to provide a

premium service by charging users according to

agreed QoS. Most of the BcN users are expected to

use multimedia traffic, so it is presumed that the

traffic analysis (or traffic monitoring) data is also

growing dramatically. The vast accumulated log data

would make it hard to be examined in time to

control the traffic, thereby the user might not be

satisfied by the non real time feedback [7][8].

In order for the BcN to be successfully diffused

to public, while competing with the traditional (best

effort) Internet, it should provide differentiated QoS

provisioning and accepted charging schemes. The

QoS-based charging system however requires a

practical traffic monitoring mechanism because

quality measurement and respective charging rely on

accurate, fast and manageable traffic monitoring

scheme.

The key requirements of the profitable TE for

BCN are as follows:

1) practical and commonly acceptable QoS

measurement

2) real time feedback mechanism

3) handling vast amounts of traffic analysis data

In the paper, Rule Based Capture(RBC)/ User

Satisfaction Parameter (USP) scheme is suggested to

fulfill the key requirements of the BcN TE. The USP

is designed to be used across all subnetworks

composing the BcN. By using the USP, user

satisfaction is examined in a short time to meet the

real time requirement of the TE. The RBC algorithm

is introduced to continuously delete unnecessary

data which has been already used to measure the

user satisfaction i.e., the USP level. The RBC is

expected to prevent the vast accumulation of traffic

analysis (log) data.

The proposed RBC/USP is implemented on the

Linux platform and its performance is investigated.

We compared the average sizes of the log files for

the cases of adopting the RBC/USP and of

conventional traffic monitoring.

368

Kim H., Jeong M. and Kim J. (2007).

A REAL TIME TRAFFIC ENGINEERING SCHEME FOR BROADBAND CONVERGENCE NETWORK (BCN).

In Proceedings of the Second International Conference on Signal Processing and Multimedia Applications, pages 358-360

DOI: 10.5220/0002137003580360

 SciTePress

Following the introduction, key requirements of

the BcN TE is investigated in Section 2. The

operation of the RBC/USP scheme is described in

Section 3, and its operation and performance are

investigated in Section 4. Conclusion follows in

Section 5.

2 KEY REQUIREMENTS OF

REAL TIME TE FOR BCN

2.1 Practical and Commonly

Acceptable QoS Measurement

Conventional QoS measurements are expressed

mostly in numerical value, such as delay, bandwidth,

or error rates. The numerical value is convenient for

representing network status or service statistics, but

it is hard to express appropriately the level of user

satisfaction because a collection of numerical values

of QoS parameters is hard to interpret[7]-[13].

Furthermore, cascading the QoS parameters through

many subnetworks to measure the Quality of

Experience(QoE) would be more complex to

interpret. To alleviate this problem, a nonnumeric

satisfaction measurement scheme, the USP is

suggested in the paper.

It is noted that BcN is composed of various

subnetworks such as conventional (best-effort)

Internet, Multi Protocol Label Switching(MPLS),

wireless LAN, and telephone network. For

successful commercialization of the BcN, an

acceptable quality measurement scheme that can be

used commonly in the heterogeneous BcN

subnetwork is required, and the USP is suggested for

this purpose.

2.2 Real Time Feedback Mechanism

The meaning of “real time” here is that traffic can be

controlled while the service is being provided. For

example, if you are not satisfied with the quality of a

streaming service, you may pay more to increase the

video quality during the streaming service. On the

contrary, if you are fully satisfied with a service then

you may not need a premium service, and want to

pay less. In some situations, moreover, just the best

effort Internet would be sufficient for the user.

Therefore, BcN TE should provide wide range of

options from free service to premium service with

instantaneous (real time) feedback. It is noted that

conventional feedback or claiming based on the

Service Level Agreement (SLA) would be possible

only after the service is ended.

3 RBC/USP ALGORITHM

3.1 Basic Operations

The rationale of the RBC/USP algorithm is that the

traffic analysis data which was already used

successfully in calculating the quality measurement

(i.e., the USP) is thrown away. The operation of the

RBC/USP TE algorithm is shown in Fig. 1.

Figure 1: Operation of RBC/USP TE algorithm for BCN.

The captured data is first saved at the Traffic

Data Buffer (TDB), and is analyzed at flow basis to

extract some QoS parameters to calculate the USP

level. The USP level is obtained from a pre-defined

combination of typical QoS parameters, such as

bandwidth and delay. The extracted USP level is

then compared with the current USP levels of the

service (or a flow) in order to update the USP Table.

The USP Table contains all the service quality

information of a user.

The extracted USP level is saved at the USP

Level Buffer, which is a temporary buffer used for

current service analysis. The service fulfillment is

then checked referring to the SLA Table. If the

captured data is from a new flow then the USP Table

is updated with the new level of USP for each

service (or flow). If the service satisfies the SLA, the

corresponding captured data is removed from the

TDB. By continuously throwing away the used

traffic monitoring data, the accumulation problem

A REAL TIME TRAFICC ENGENEERING SCHEME FOR BROADBAND CONVERGENCE NETWORK(BCN)

369

can be avoided. When the service does not satisfy

the SLA, a traffic control signal is sent immediately

to the related network devices (e.g., the routers). The

handling of the signal may depend on the SLA.

3.2 USP Levels

The USP level represents the service quality with a

nonnumeric level such as A, B, C, or D. For

example, if e-mail is defined to have two levels (e.g.,

A and B), ‘A’ may represent “good”, and ‘B’ a

“bad”service quality. The USP levels can be

extracted from a few typical QoS parameters. For

example, with streaming service, the combination

of: bandwidth of larger than 50Mbps, delay of less

than 20ms, jitter of less than 20ms will be graded to

be ‘A’ level. Other examples of USP levels are

given in Table 1 (the numerical values of QoS

parameters are given arbitrarily by authors). The

choice of key QoS parameters and their values

which will be used for calculating USP level should

be defined via heuristic evaluations.

Table 1: Mapping of USP level and QoS parameters.

Application

USP

level

bandwidth

(bps)

delay jitter

A > 1M < 500ms < 100ms

E-mail

B > 100K < 3s < 1s

A > 50M < 20ms < 20ms

B > 10M < 50ms < 20ms

C > 2M < 100ms < 20ms

D > 1M < 100ms < 50ms

Streaming

E > 500K < 200ms < 100ms

A > 10M < 100ms < 50ms

B > 1M < 200ms < 100ms WWW

C > 100K < 300ms < 100ms

A > 10M < 100ms < 50ms

B > 1M < 200ms < 100ms FTP

C > 100K < 300ms < 100ms

A > 500K < 200ms < 100ms

Telnet

/Rlogin

B > 100K < 300ms < 100ms

4 CONCLUSIONS

In the paper, a simple but practical TE scheme for

high speed networks such as BcN is proposed. The

RBC/USP algorithm is expected to overcome the

latent problems of BcN: providing acceptable QoS

agreement among vendors, real time feedback, and

accumulation of traffic analysis data. The RBC

solves the problem of massive accumulation of QoS

parameters, and the adoption of USP simplifies the

QoS negotiation among the heterogenous

subnetworks taking part in the end-to-end BcN

service. The USP can play an intermediary role of

service relay between service providers.

The RBC/USP was implemented in the Linux

platform and its performance is investigated. We

found that the average log size is reduced to 0.058%

for FTP, and 2.39% for streaming service by using

RBC/USP. The reduction of the log data means not

only practical usability of the algorithm, but also it is

expected to provide commonly acceptable end-to-

end QoS description. By simulations, we found that

the RBC/USP can be a good candidate of a real time

TE methodology for the emerging BcN.

Further study includes reasonable defining of

practical USP levels which can simply reflect

service requirements of users as well as network

service providers.

ACKNOWLEDGEMENTS

This work was supported in part by MIC, Korea

under the ITRC program (C1090-0603-0035)

supervised by IITA.

REFERENCES

Taeil Chae,“Introduction of KT -BcN,” BcN 2006, The 1st

International Workshop on April 2006, pp. 1 - 22.

ETRI, "BcN Vision for Ubiquitous Society," BcN 2006,

The 1st International Workshop on April 2006, pp. 1 -

38. .

Sang Ki Kim,et al., "Open Service Platform: A Service

Creation and Provisioning Environment for BcN

Infrastructure," BcN 2006. The 1st International

Workshop on April 2006, pp. 1 - 9.

Kyung Gyu Chun, et al., "Review on the Operation,

Administration and Maintenance (OAM) of BcN,"

2006 Asia-Pacific Conference on Aug. 2006, pp. 1 -

Jeongwook Won, et al., “How to succeed in NGcN of

Korea,” APCC 2003. The 9th Asia-Pacific Conference

on Volume 1, Sept. 2003, pp. 264 - 269.

Kyu Ouk Lee, et al., “QoS evaluation of KT-NGN,”

APCC 2003. The 9th Asia-Pacific Conference on Vol.

3, Sept. 2003, pp. 900 - 903.

SIGMAP 2007 - International Conference on Signal Processing and Multimedia Applications

370