Changbin Liu, Lei Shi, Bin Liu

Department of Computer Science and Technology, Tsinghua University

Proceedings of the Fourth European Conference on Universal Multiservice Networks (ECUMN’07)

Chen Bin Kuo (20077202)

Young J. Won (20063292)

DPNM Lab.

Introduction NGN traffic classifications and their utility functions Network utility maximization (NUM) Numeric results and analysis Discussion Conclusion

04/19/23 2

DPNM Lab.

Next generation network (NGN) must natively support triple-plays.

How to schedule traffic and allocate bandwidth at both backbone and access links.

Designing a scheduling (bandwidth allocation) algorithm is exactly the issue this paper tries to settle.

04/19/23 3

DPNM Lab.

In industry designing NGN [13][14], the strict-priority scheduling is mostly adopted. Rigidly favors the voice and video traffic without flexibility

Utility-based solutions Shenker [1] discussed traffic classifications in IP network

from the viewpoint of user utility Kelly et al. [5][6] applying utility-based methods to

scheduling and bandwidth allocation in the objective of Network Utility Maximization (NUM)

04/19/23 4

DPNM Lab.

No single work has emphasized on the practical issue of scheduling triple-play services under the background of NGN.

Translating this issue into a nonlinear maximization problem with inequality constraints.

04/19/23 5

a) VoIP trafficb) IPTV trafficc) TCP elastic trafficd) HTTP traffice) Other UDP traffic

04/19/23 6

DPNM Lab.

Due to remarkable distinction of QoS requirements in NGN Classifying NGN traffic into five categories

User utility function is introduced To measure network performance and user satisfaction

degree Determined by the QoS metrics received in the user end Including packet delay, jitter and loss rate

04/19/23 7

DPNM Lab.

Sensitive to packet delay and loss caused by bandwidth insufficiency Utility function falls into the category of hard real-time kind

[1][2][10], with a minimal bandwidth requirement of Bmin1

04/19/23 8

DPNM Lab.

Utility function is similar to VoIP’s but with some differences Tolerate occasional delay-bound violations and packet drops Minimal encoding rate, denoted as Bmin2 is independent of

network congestion Logistic model is used

04/19/23 9

DPNM Lab.

Generated by delay-tolerant TCP applications Such as file transfer and email

Utility function have been studied by Kelly et al. [6] and other researchers [11][12]

04/19/23 10

DPNM Lab.

TCP traffic which concerns packet delay Mainly contains the HTTP traffic generated by web services

Utility function is different from TCP elastic traffic, has a minimum tolerable bandwidth Bmin4

04/19/23 11

DPNM Lab.

DNS packets, other streaming media traffic, and on-line gaming traffic [17][18] Delay-sensitive Every application type has a utility function

The shape of utility function resembles IPTV traffic

04/19/23 12

DPNM Lab. 04/19/23 13

VoIP IPTV TCP elastic HTTP UDP

Bmin 64Kbps 100 Kbps 24Kbps

Bmax 10Mbps 10Mbps 10Mbps 500Kbps

ɛ 0.001 0.001

a) KKT methodb) Lagrange multipliers method without KKT conditions

04/19/23 14

DPNM Lab.

Based on NGN traffic’s utility functions, we can solve the congestion-phased bandwidth allocation issue while conforming to NUM.

Total utility gained on the link is:

Bandwidth allocation is restricted by:

04/19/23 15

N : the number of NGN users utilizing this link

pi : traffics classes

C : the bandwidth of a link (set to 10Gbps)

N : the number of NGN users utilizing this link

pi : traffics classes

C : the bandwidth of a link (set to 10Gbps)

DPNM Lab.

Lagrange Multiplier method with KKT (Karush-Kuhn-Tucker) conditions

Solving the nonlinear optimization problem

Accurate and comprehensive solution requires substantial complicated computations Applying simplified form which is enough to ravel NUM

problem for triple-plays

04/19/23 16

DPNM Lab.

Observing NGN traffic’s utility functions VoIP/IPTV/other UDP traffic’s utility functions are

relatively smoother in some points It is not cost-effective to allocate bandwidth to

VoIP/IPTV/other UDP traffic without booming the utility Turning point (TP)

04/19/23 17

BandwidthBandwidth

IPTVIPTV HTTPHTTP

DPNM Lab.

After finding the TP, we can apply the Lagrange Multipliers method without KKT conditions to solve the NUM problem in (10)

Subject to:

04/19/23 18

a) Data-dominated networkb) IPTV-dominated network

04/19/23 19

DPNM Lab.

Two network scenarios Current Internet, where HTTP and TCP elastic traffic still

dominate the volume Prospective NGN, where the emerging services, especially

the IPTV traffic, will dominate the network For each scenario, calculate in two situations

Maximal Utility Equalization (MUE) Maximal Utility In-equalization (MUI)

04/19/23 20

V1 (VoIP) V2 (IPTV) V3 (TCP elastic)

V4 (HTTP) V5 (other UDP)

MUE

1 1 1 1 1

MUI 1 9 1 1.5 2

DPNM Lab.

Data-dominated network According to recent trace observation [15]

IPTV-dominated network

04/19/23 21

VoIP IPTV TCP elastic HTTP other UDP

Traffic proportions

10% 10% 10% 50% 20%

VoIP IPTV TCP elastic HTTP other UDP

Traffic proportions

10% 50% 10% 20% 10%

DPNM Lab. 04/19/23 22

DPNM Lab. 04/19/23 23

DPNM Lab.

Previous bandwidth allocation schemes for triple-play services mostly adopt the strict-priority scheduling Highest priority to VoIP traffic Second highest priority to IPTV and lowest priority to others

In this paper Highest priority to VoIP traffic Assigning IPTV traffic with second-highest priority is not well

supported from the objective of NUM Suggesting that ISP charges more about IPTV services (future

work)

04/19/23 24

DPNM Lab.

Studied the problem of scheduling and bandwidth allocation for triple-play services in the objective of NUM.

Presenting theoretical method to compute bandwidth allocation results

Results: VoIP and other low-throughput UDP traffic can always be

guaranteed of sufficient bandwidth As congestion becomes severer, IPTV’s bandwidth

decreasing quickly TCP elastic and HTTP traffic experience exponential

bandwidth degradations when congestion degree increases

04/19/23 25

04/19/23 26