Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 27

Cell PCH and Fast Dormancy Introduction

OSP Trial Results

20th June 2012


Pawel Holubowski - NAD

1
Agenda

CELL PCH & FD


Introduction

OSP Network Overview

Trial Results

Summary
2
New Terminals

3
Smartphone Traffic Behavior
Idle time distribution
Live measurements, smartphone centric operator
Very frequent connection 100%
establishments
80%
Example from field
measurements 60%

50% of devices reconnect


40%
within 3 minutes
Often rather small amount of 20%
data transfer
Mostly network initiated 0%
0 1 2 3 4 5

Causes high signaling load Minutes

High degree of chattiness


4
Impact on the Network
It seems that the main issues are caused by two
behaviors/features:

5
Impact on system
behavior

6
Signaling improvements with
PCH HSPA Idle -> HSPA
FACH -> HSPA 200-500mA
Around 12 signaling Around 30 signaling
messages over Uu and Iub messages over Uu, Iub
FACH and Iu
100-150 mA

PCH -> FACH PCH Cell PCH is important to lower the


Only 3 signaling <5 mA
signaling load in the network
messages over Uu
IDLE Much less signaling from Cell PCH than
<5mA from idle to go to FACH or HSPA

Cell PCH state comparable with idle


from battery efficiency pointy of view

Using Cell PCH has a potential to reduce


signaling load for chatty devices

7
Device-triggered Fast
dormancy
Rel8 Fast Dormancy Pre-rel8 Fast Dormancy

With Rel8 fast Device triggered


HSPA connection release to
dormancy idle
Overrides the
The UE can indicate network controlled
that it has completed FACH switching
the data transfer
Used by almost all
Can be used to down smartphones
switch the UE to URA PCH
Invalidates PCH state
Idle

Rel8 Fast Dormancy returns the control to the network


9
Fast Dormancy in Huawei
RAN

10
Agenda

CELL PCH & FD


Introduction

OSP Network Overview

Trial Results

Summary
11
Trial objectives
FT Strategy
Objective: Secure the introduction of the Fast Dormancy feature in FT networks
while keeping stable & good QoS KPI.
Drivers:
Network signalling load decrease
Gain for baseband (#CE, #HSPA connections)
Latency reduction for the end user faster establishment
UE battery saving
Orange Spain network will be used to trigger T6 milestone (at Group level) in
case of non-degradation of the QoS KPI.

RNC of Murcia (MUR60R03)


341 NodeBs / 1543 cells
3G SW release: RAN 12
HSPA is enabled in all the cells. F1 is preferred for CS voice and F2 is
preferred for PS data. F3 is used where further capacity is required
12
RNC Murcia Smartphone behavior
analysis

Smartphone behavior definition: more than 10 PS RAB attempts for one user in

one busy hour.

Smartphone behavior users take more PS signaling load comparing with other

devices. The PS RAB Times produced by Smartphone behavior users is around

84% though the penetration of them is only 21%.


13
Smartphone Distribution Overview
The Smart Phone behavior user
penetration is 21% and produce
84% PS RAB attempts. Smart
Phone behavior user consumes
more signal load than general
user.

The Smart Phone behavior user


do not be contributed by one
brand terminal. More than one
brand terminal contribute this high
PS signal.

The expect gain of CELL_PCH is


only around 2.3%. The expect
gain of EFD is around 12.5%.

14
Agenda

CELL PCH & FD


Introduction

OSP Network Overview

Trial Results

Summary
15
Trial Timetable
Test Case Date Parameters
CELL PCH Activation
TC1 16/04/2012 DCH to FACH transition timer: 5s
FACH to PCH Transition timer : 15 s

DCH to FACH transition timer: 5s


TC2 24/04/2012 FACH to PCH Transition timer : 10 s
PS Inactivity time= 360 (6 min)

FastDormancyEnhancement: ON
FD DCH to FACH transition timer: 5s
TC3 09/05/2012
FD FACH to PCH Transition timer : 1 s
T323=D30

16.04.2012 24.04.2012 09.05.2012

T0 T1 T2 T3

START STOP
16
PS RAB Attempts

Cell PCH and FD caused 40% less of RAB attempts


17
PS Call Drop Rate

PS CDR
Delta from PS CDR Delta from
Period Nominator CDR
Reference Denominator Reference
Avg
CDR on stable
Reference 75,442 17,023,884 0.44%
TC1 89,104 18% 22,344,314 31% 0.40% level but 86% more
TC2
TC3
123,582
140,039
64%
86%
25,049,626
32,351,231
47%
90%
0.49%
0.43%
drops
18
PS Minutes per drop

35% higher
probability of drop

Minutes per Drop = [MeasuredTimeInMinutes]* [ERL_PS / ( #_of_PsDrops)]


where:
Erl_PS = sum(HS, PS R99, FACH) users (PCH not included)
#_of_PsDrops = sum(HS, PS R99, FACH) drops (PCH not included) 19
CS Call Drop Rate

KPI 06/05/2012 03/28/2012


Speech CDR 0.23 0.23
MRAB CDR 0.45 0.45 18% higher CDR most
Speech + MRAB CDR 0.32 0.27 probably due to higher MRAB
0.60 0.83
Speech Traffic Ratio
ratio
MRAB Traffic Ratio 0.40 0.17
Calculated CDR 0.32 0.27
20
PS users distribution

Huge number of users in PCH with minimal impact


on HS and FACH
CPU Load

8% for Max and 4% for Avg load reduction less than


expected
22
Channel Elements Usage

8%decrease od DL/UL CE usage due to lower


number of user in DCH state
23
Other indicators
23% Gain 20% Gain

24
RTT Measurements

Explanation provided by Huawei:


Big ping 1460Byte is higher than the traffic threshold for the transition FACH -> HS
In case of Web browsing SYN/SYN ACK analysis, the latency is lower because the packet s
are sent/ received in FACH state.

Achieved results below our expectation


Agenda

CELL PCH & FD


Introduction

OSP Network Overview

Trial Results

Summary
26
Summary
Achieved results might vary in different networks due to UE penetration and
parameters setting

Significant reduction (up to 50%) of signaling load over all interfaces Uu, Iub and Iu
(RRC, NBAP, RANAP)

Degradation of all major performance KPI CDR and CSSR for CS and PS

Lower max CPU load of the RNC boards managing the signalling part by 8%

Reduction of CE consumption and users in HS state by 8%

No gains for Web browsing and 15 to 20% for ping

There are still open issues addressed to Huawei:


Explanation and propose solution for QoS degradation

Investigation of lower than expected latency gain

27
Questions?

Thank you!
28

You might also like