Browse Prior Art Database

NETWORK SEGMENTATION AND SLICING: METHODS FOR NETWORK SLICING IN 802.11AX-TYPE OF SYSTEMS IN 5G NETWORKS

IP.com Disclosure Number: IPCOM000250415D
Publication Date: 2017-Jul-12

Publishing Venue

The IP.com Prior Art Database

Related People

Mukesh Taneja: AUTHOR

Abstract

Methods are provided to support network slicing in IEEE 802.11ax type of 5G systems. Per-slice (coupled with an enhanced per-client) proportional fair metric is defined and methods are provided to compute this by taking into account various factors such as urgency indicators for that client and slice, throughput deficit for that slice, buffer indices for the clients, channel conditions, instantaneous and weighted average of data rates and MCS values across Resource Units (RUs). Methods are also presented for resource allocation.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 16% of the total text.

Copyright 2017 Cisco Systems, Inc. 1

NETWORK SEGMENTATION AND SLICING: METHODS FOR NETWORK SLICING IN 802.11AX-TYPE OF SYSTEMS IN 5G NETWORKS

AUTHOR: Mukesh Taneja

CISCO SYSTEMS, INC.

ABSTRACT

Methods are provided to support network slicing in IEEE 802.11ax type of 5G

systems. Per-slice (coupled with an enhanced per-client) proportional fair metric is defined

and methods are provided to compute this by taking into account various factors such as

urgency indicators for that client and slice, throughput deficit for that slice, buffer indices

for the clients, channel conditions, instantaneous and weighted average of data rates and

MCS values across Resource Units (RUs). Methods are also presented for resource

allocation.

DETAILED DESCRIPTION

5G network slicing will allow operators to split single physical network into

multiple virtual networks (and 802.11ax could be part of this end-to-end network). Each

such logical slice to meet certain performance goals. An example is given below:

Slice A) eMBB: Video / AR / VR (High throughput, low latency, etc.).

Slice B) uMTC (ultra-high reliable IoT applications): Low latency, high reliability.

Slice C) mMTC: Large number of devices, low / sporadic data per device, etc.

Slice D) No strict delay requirements: May be higher throughput per-device but

(much) smaller number of devices (compared to Slice C).

IEEE 802.11ax supports Downlink / Uplink (DL / UL) Orthogonal Frequency

Division Multiple Access (OFDMA) in addition to other features for High Efficiency

Wireless Local Area Network (WLAN) operation in dense scenarios. With OFDMA, it

supports Resource Units (RUs) of sizes such as 26 / 52 / 104 / 242 / 484 / 996 / 2x996 sub-

carriers. As an example, if channel bandwidth is 20 MHz, each client could be assigned

RUs of sizes 26 / 52 / 104 / 242 sub-carriers (resulting in approximate bandwidth allocation

Copyright 2017 Cisco Systems, Inc. 2

of 2 / 4 / 8 / 20 MHz to some of these clients). 802.11ax system also supports varying MCS

(or data rate) values for each client for each RU.

The challenge is how to ensure that an 802.11ax network can provide required

resources for a group of devices that are part of a given logical network slice? How many

different slices (with different performance goals) can an 802.11ax network support at any

given time?

In a static allocation scheme, certain RUs (i.e. 802.11ax Resource Units or group

of sub-carriers) can be reserved for each slice. Due to varying channel conditions and

interference, DL / UL data rate (or MCS) that can be achieved for each RU for each client

can keep varying. With this approach, stations belonging to a network slice may or may

not get their desired QoS and observed system capacity may also be lower than expected.

A Proportional fair (PF) QoS scheduler at an access point (AP) considers

instantaneous channel condition for each user and observed throughput for that user to

compute a PF metric. It picks up a user with maximum value of PF metric to serve.

Fo...