Browse Prior Art Database

LTE scheduling and Load Balancing using QMAN

IP.com Disclosure Number: IPCOM000241209D
Publication Date: 2015-Apr-03
Document File: 5 page(s) / 104K

Publishing Venue

The IP.com Prior Art Database

Abstract

We describe a flexible and scalable software architecture for accelerating Long Term Evolution (LTE) Layer2 mobile wireless stack on Data Path Acceleration Architecture (DPAA) enabled multicore platforms for high performance and capacity. LTE air interface scheduler is a key and highly compute intensive module of LTE Layer2. LTE air interface scheduler schedules the radio bearer data for transmission on an air interface based on LTE QoS assigned to them. This solution describes the offload of prioritized radio bearer handling to hardware using DPAA. Application of this solution accelerates the LTE air interface scheduler operation. We also describe a solution for flexible multicore partitioning of LTE Layer2 components using DPAA.

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 31% of the total text.

LTE scheduling and Load Balancing using QMAN

Abstract

We describe a flexible and scalable software architecture for accelerating Long Term Evolution (LTE) Layer2 mobile wireless stack on Data Path Acceleration Architecture (DPAA) enabled multicore platforms for high performance and capacity. LTE air interface scheduler is a key and highly compute intensive module of LTE Layer2. LTE air interface scheduler schedules the radio bearer data for transmission on an air interface based on LTE QoS assigned to them. This solution describes the offload of prioritized radio bearer handling to hardware using DPAA. Application of this solution accelerates the LTE air interface scheduler operation. We also describe a solution for flexible multicore partitioning of LTE Layer2 components using DPAA.

Problem description

There is a need to devise flexible and scalable software architecture for LTE Layer2 on DPAA enabled multicore platforms for high performance and capacity. LTE requires prioritized traffic processing based on LTE QCI values. To address this a solution needs to be devised for acceleration of LTE Layer2 Scheduler by hardware offload of prioritized LTE traffic processing. LTE traffic is grouped into LTE Service Data Flow (SDF) aggregates. Hence there is a need to offload LTE SDF processing to hardware for LTE Scheduler acceleration. Conventional multicore partitioning of software is generally rigid and involves software based inter-process and inter-core communication. This is wasteful of CPU cycles and generally uses locks to achieve synchronization between multiprocessing elements. This solution explores DPAA based offload of multicore processing for LTE use case. DPAA (accelerator) assisted multicore scheduling of LTE Layer2 components provides lockless and flexible partitioning and load balancing scheme for LTE Layer2 stack.

Solution overview

This solution describes multicore scheduling of LTE Layer2 modules using QMAN for a single sector downlink processing. This solution can be scaled up to support multiple LTE sectors as well. The scheme is laid out assuming that dequeues from QMAN are carried out by processing cores in a non-frame queue (FQ) selective manner.

Association of modules to cores is flexible and is controlled by appropriate initialization sequence at the time of LTE Layer2 library initialization. The QCI/ Packet Delay Budget (PDB) of a bearer will be decided at the time of bearer allocation. The pool channels Pool Channel 1 (PCH1) and Pool Channel 3 (PCH3) both will have identically prioritized Frame Queues (FQs). Downlink Packet Data Convergence Protocol (PDCP) module (PDCP-DL) enqueues the frames oblivious to their QCI/PDB value. Downlink scheduler dequeues the frames and puts them for downlink Radio Link Control (RLC-DL) processing based on their QCI/PDB. Priority of various FQs in the system can be fine-tuned to adjust the system performance and stability.

Detailed solution description

The diagram in Figure 1 describes the...