Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Backhaul-Aware Scheduling Algorithms for GPRS/EDGE Networks

IP.com Disclosure Number: IPCOM000126466D
Original Publication Date: 2005-Jul-19
Included in the Prior Art Database: 2005-Jul-19
Document File: 7 page(s) / 291K

Publishing Venue

Motorola

Related People

Rajeev Agrawal: INVENTOR [+5]

Abstract

In this paper, we consider the problem of scheduling for GPRS and EDGE networks, where the backhaul bandwidth is aggregated across all data timeslots in a carrier. This is in contrast to a circuit backhaul where the bandwidth is provisioned for each individual data timeslot. To achieve statistical multiplexing gains, the aggregate backhaul bandwidth for each carrier is allowed to be smaller than that needed for all the data timeslots to operate at the highest modulation and coding scheme. We design scheduling algorithms that take into account the provisioned backhaul bandwidth. We show that not taking the backhaul bandwidth into account can yield sub-optimal performance. Our algorithms variously trade-off performance, computational complexity, and the amount and accuracy of information used.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 17% of the total text.

Backhaul-Aware Scheduling Algorithms for

GPRS/EDGE Networks

Rajeev Agrawal*, Aniruddha Diwan#, Pranav Joshi*,

Suresh Kalyanasundaram#, and Vinod Ramachandran#

*Motorola,

Arlington Heights

,

USA

#Motorola,

Bangalore

,

India

E-mail: {Rajeev.Agrawal, Aniruddha, Pranav.Joshi,

Suresh.Kalyanasundaram, vinodkumar}@motorola.com

Abstract

In this paper, we consider the problem of scheduling for GPRS and EDGE networks, where the backhaul bandwidth is aggregated across all data timeslots in a carrier. This is in contrast to a circuit backhaul where the bandwidth is provisioned for each individual data timeslot. To achieve statistical multiplexing gains, the aggregate backhaul bandwidth for each carrier is allowed to be smaller than that needed for all the data timeslots to operate at the highest modulation and coding scheme. We design scheduling algorithms that take into account the provisioned backhaul bandwidth. We show that not taking the backhaul bandwidth into account can yield sub-optimal performance. Our algorithms variously trade-off performance, computational complexity, and the amount and accuracy of information used.

1 Introduction

Cellular networks employ hierarchical network architecture with multiple network elements that accomplish different functionalities. Figure 1 shows the General Packet Radio Service (GPRS) network where all the elements work together to perform different functions such as radio resource management, session management, and mobility management, that are needed for the successful operation of a cellular network. Adaptive modulation and coding is one of the main techniques by which 2.5G and 3G systems enhance end user data throughput. The fundamental idea in adaptive modulation and coding is to reduce the redundant bits, i.e., increase the code rate, and/or go towards higher modulation schemes when the channel is good, thereby increasing the data rates. The channel coding is typically done at the BTS (Base Transceiver Station) to avoid having to transport the redundant bits over the wired backhaul links, where such robust error protection is not required. Channel coding is only required for the wireless channel, and hence it is reasonable that the channel coding is performed at the BTS in the downlink direction. The radio link control (RLC) layer in GPRS/EDGE (Enhanced Data Rates for Global Evolution) handles functions such as block segmentation, retransmission, and reliable transfer [2]. The RLC protocol is terminated at the packet control unit (PCU) and the GPRS/EDGE standards allow the PCU to be either located along with the BTS, BSC (Base Station Controller), or SGSN (Serving GPRS Support Node) [3]. One of the main reasons to have the PCU located along with the BSC or SGSN is that the need for transferring of buffers and the RLC context between different network elements is reduced. For example, RLC buffers and contexts need to be transferred for all inter-BTS handoffs when the PCU is co-located with the BTS, while it...