Zte umts Load Monitoring and expansionguide 141112012330 Conversion Gate02

UMTS Network Load Monitoring and Expansion Guide R1.0

UMTS Network Load Monitoring and Expansion Guide

Internal Use Only▲

LEGAL INFORMATION By accepting this certain document of ZTE CORPORATION you agree to the following terms. If you do not agree to the following terms, please notice that you are not allowed to use this document. Copyright © 2013 ZTE CORPORATION. Any rights not expressly granted herein are reserved. This document contains proprietary information of ZTE CORPORATION. Any reproduction, transfer, distribution, use or disclosure of this document or any portion of this document, in any form by any means, without the prior written consent of ZTE CORPORATION is prohibited. and are registered trademarks of ZTE CORPORATION. ZTE’s company name, logo and product names referenced herein are either trademarks or registered trademarks of ZTE CORPORATION. Other product and company names mentioned herein may be trademarks or trade names of their respective owners. Without the prior written consent of ZTE CORPORATION or the third party owner thereof, anyone’s access to this document should not be construed as granting, granting, by implication, estopped or otherwise, any license or right to use any marks appearing in the document. The design of this product complies with requirements of environmental protection and personal security. This product shall be stored, used or discar ded in accordance with product manual, relevant contract or laws and regulations in relevant country (countries). This document is provided “as is” and “as available”. Information contained in this document is subject to continuous update without further notice due to improvement and a nd update of ZTE CORPORATION’s products and technologies.

ZTE CORPORATION Address:

NO. 55 Hi-tech Road South ShenZhen P.R.China 518057

Website:

http://support.zte.com.cn

Email:

[email protected]

ZTE Confidential Proprietary

© 2013 ZTE CORPORATION. CORPORATION. All rights reserved.

I



LEGAL INFORMATION By accepting this certain document of ZTE CORPORATION you agree to the following terms. If you do not agree to the following terms, please notice that you are not allowed to use this document. Copyright © 2013 ZTE CORPORATION. Any rights not expressly granted herein are reserved. This document contains proprietary information of ZTE CORPORATION. Any reproduction, transfer, distribution, use or disclosure of this document or any portion of this document, in any form by any means, without the prior written consent of ZTE CORPORATION is prohibited. and are registered trademarks of ZTE CORPORATION. ZTE’s company name, logo and product names referenced herein are either trademarks or registered trademarks of ZTE CORPORATION. Other product and company names mentioned herein may be trademarks or trade names of their respective owners. Without the prior written consent of ZTE CORPORATION or the third party owner thereof, anyone’s access to this document should not be construed as granting, granting, by implication, estopped or otherwise, any license or right to use any marks appearing in the document. The design of this product complies with requirements of environmental protection and personal security. This product shall be stored, used or discar ded in accordance with product manual, relevant contract or laws and regulations in relevant country (countries). This document is provided “as is” and “as available”. Information contained in this document is subject to continuous update without further notice due to improvement and a nd update of ZTE CORPORATION’s products and technologies.

ZTE CORPORATION Address:

NO. 55 Hi-tech Road South ShenZhen P.R.China 518057

Website:

http://support.zte.com.cn

Email:

[email protected]



I



Revision History Product Version

Document Version

RNC V3.09

Serial Number

R1.0

Reason for Revision First published

Author Date 2011-3-15 2011-3-15

Document Version R1.0


Prepared by Qiao Bin, Jin Zhengtuan, and Xu Yi

Reviewed by

Ma Wei


Approved by

Wang Zhenhai

II



Intended audience: UMTS network optimization engineers

Proposal: Before reading this document, you had better have the following knowledge and skills. SEQ 1

Knowledge and skills Null

Reference material Null

2 3

Follow-up document: After reading this document, you may need the following information. SEQ 1

Reference material Null

Information Null

2 3



III



About This Document Summary Chapter

Description

1

Overview

Briefly introduces the background and the main contents of high-load network monitoring and optimization.

2

High-Load Network Monitoring

Describes the classification of UMTS network elements (NEs) and the key performance indicators (KPIs) for network load monitoring.

3

High-Load Network Optimization

Describes the process of network load optimization.

4

High-Load Network Expansion

Describes the thresholds, judgment, and implementation of capacity expansion for high-load networks.


© 2013 ZTE CORPORATION. All rights reserved.

IV



TABLE OF CONTENTS 1 1.1 1.2

Overview ......................................................................................................... 1 Background ...................................................................................................... 1 Main Contents .................................................................................................. 2

2 2.1 2.2 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5

High-Load Network Monitoring ...................................................................... 6 WCDMA NE Classification ................................................................................ 6 Network Load Monitoring Indicators .................................................................. 7 Key Indicators ................................................................................................... 7 Average Utilization of Non-HSDPA Code Resource.......................................... 7 Average Utilization of Non-HSDPA Carrier Transmit Power.............................. 9 Uplink Noise Rise ........................................................................................... 10 Average Throughput of HSDPA Cell ............................................................... 11 Average Throughput of HSDPA Single User................................................... 11

3 3.1 3.2

High-Load Network Optimization ................................................................ 13 Network Load Optimization Stages ................................................................. 13 Network Load Optimization Process ............................................................... 14

4 4.1 4.1.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4

High-Load Network Expansion .................................................................... 16 Expansion Process ......................................................................................... 16 Expansion Analysis Process ........................................................................... 16 Expansion Criteria and Methods ..................................................................... 17 Cell Expansion ................................................................................................ 17 Node B-CE Expansion .................................................................................... 26 IUB Transmission Expansion .......................................................................... 30 RNC Expansion .............................................................................................. 33



V



FIGURES Figure 1-1

High-Load Network Monitoring and Optimization ................................................ 3

Figure 1-2

High-Load Network Monitoring Process .............................................................. 4

Figure 1-3

High-Load Network Optimization Process ........................................................... 5

Figure 2-1

Relationship Between the Uplink Capacity and Noise ....................................... 10

Figure 3-1

Flowchart of Network Load Optimization ........................................................... 13

Figure 3-2

High-Load Network Optimization Process ......................................................... 15

Figure 4-1

Expansion Analysis Process ............................................................................. 17

Figure 4-2

Cell Expansion Decision Process ...................................................................... 18

Figure 4-3

Combination Chart of Cell Expansion Decision Formula ................................... 21

Figure 4-4 Relationship Between the Code Resource Utilization and Traffic ...................... 23 Figure 4-5 Relationship Between the Cell Carrier Transmit Power Utilization and TCP Limited Proportion ................................................................................................................ 24 Figure 4-6 Average Utilization Rates of Uplink and Downlink NodeB CE Resources in Shapingba, Chongqing, China .............................................................................................. 28 Figure 4-7 Maximum Utilization Rates of Uplink and Downlink NodeB CE Resources in Shapingba, Chongqing, China .............................................................................................. 29

TABLES Table 2-1

Code Resource Distribution of Code Channel ...................................................... 8

Table 4-1

Cell Expansion Thresholds ................................................................................. 19

Table 4-2

Cell Expansion Implementation Rules ................................................................ 25

Table 4-3 Node B CE Expansion Thresholds and Expansion Methods .............................. 27 Table 4-4


Table 4-5

Iub Transmission Expansion Thresholds ............................................................ 31

Table 4-6


Table 4-7

Monitoring Indicators of RNC Hardware Expansion ............................................ 36

Table 4-8

Observation Indicators of RNC Hardware Expansion ......................................... 37

Table 4-9

Observation Indicators of RNC Hardware Expansion ......................................... 39

Table 4-10

RNC Expansion Implementation Rules............................................................. 41



VI


1

Overview

1.1

Background


To monitor and optimize the high load and performance of UMTS network is one of the key tasks in the network operation and maintenance stage. With the increase of the subscriber number and service application, especially with the rapid development of the wireless broadband service, the network load will keep increasing. When the network load reaches a certain level, the network resources will be congested and the network performance will deteriorate, affecting the end users ’ service experience. To provide the users with high-speed access experience and keep the competitiveness of the UMTS network, the operators should perform real-time monitoring to the load and performance of the whole network, alarm the network element (NE) exceeding the load threshold, take timely measures of optimization and expansion to meet the requirement of service development. In the narrow sense, the load refers to the traffic loaded by the network or channel. In the broad sense, except for the network traffic, the operators need to consider the resource utilization of the software and hardware of each NE in the network. The higher the utilization rate is, the heavier the load will be. Compared with the 2G network, the monitoring and management of the UMTS network is more complex. The reasons are as follows: The UMTS is a soft capacity system. Its capacity is not only constrained by the hard resources such as the CE and Iub configuration bandwidth, but also constrained by the soft resources such as the OVSF code, uplink interference, and downlink power. Subject to the requirements for the network coverage and service quality, the system capacity is not a fixed value. UMTS is a hybrid multi-service system. The system capacity is different due to different service structure and different service model, so we cannot simply use the traffic of a certain service to monitor the system capacity.



1



The UMTS may use the hybrid carrier strategy of R99+HSPA. R99 and HSPA share the system resources, making it more complex to monitor the capacity of R99 and HSPA. The UMTS is a network focusing on the data service. To judge the data network congestion, we cannot simply follow the processing of the traditional voice-centric network, that is, we cannot judge the network congestion according to whether there is a admission rejection, but should judge the network congestion by combining the HSPA

user ’s real-time experience rate with the network resource occupation. Based on the network management counter of RNC V3.09, this manual gives the definition of the monitoring indicators of UMTS network load and the suggestions for the monitoring threshold.

1.2

Main Contents The high-load network monitoring and optimization guide shall apply to the communication network in the UMTS commercial operation and maintenance phase. As shown in Figure 1-1, the high-load network monitoring and optimization transversely aim at three levels of NEs: the cell of radio access network (RAN), Node B and RNC. Longitudinally, there are three phases: high-load network monitoring, high-load network optimization and high-load network expansion, respectively corresponding to the three parts of this guide.



2


Figure 1-1


High-Load Network Monitoring and Optimization

Horizontal: RNC, NodeB, and cell Vertical

1. High-load network monitoring User perception evaluation

Network resource evaluation

2. High-load network optimization RF optimization

Parameter optimization (handover threshold, congestion control, and load balancing)

3. Expansion decision and implementation Expansion decision -> Expansion i mplementation

Part 1 corresponds to Chapter 2 of this guide, mainly describing the indicators needed to be monitored for the network load. As shown in Figure 1-2, the network load monitoring of WCDMA system mainly aim to three levels of NEs: the cell of wireless access network, NodeB and RNC. Each NE corresponds to different RAN. Mainly involving the air interface resources such as the code resource and power resource, the cell NE closely relates to the user ’s experience rate and focuses on the user ’s feeling. The Node B NE mainly involves the transmission resource and CE resource. According to the RNC configuration, the RNC NE mainly involves the indicators such as the occupation of RCP and CPU as well as the use of RUP and CE resources.



3


Figure 1-2


High-Load Network Monitoring Process

Yes Cell-levelmonitoring, analysis, and alarm

High-load network performance monitoring and evaluation

NodeB-level monitoring, analysis, and alarm

RNC-level monitoring, analysis, and alarm

Perform optimization to solve the problem of high load

No

Meet the cell capacity expansion requirements or not?

Yes

Cell expansion Add carriers or power amplifiers

Meet the NodeB capacity expansion requirements or not?

Yes

NodeB expansion Add BPC boards or transmission resources

Meet the RNC capacity expansion requirements or not?

Yes

RNC expansion Software: Add licenses Hardware: Add RUB or RCB boards

No

Part 2 corresponds to Chapter 3 of this guide, mainly describing the optimization of the high-load network performance. There are mainly 2 aspects: the high-load network optimization process and common optimization methods. As shown in Figure 1-3, in the optimization process of high-load network, you should optimize the RF and wireless parameters according to the actual network situation. The wireless parameters optimization mainly includes the parameters such as the handoff, congestion control, load equalization, DRBC, power control and HSPA, so as to reduce the consumption of various resources.



4


Figure 1-3


High-Load Network Optimization Process

High-load NE Cell-level NEs involve false load rises and real load rises. Vertical

RF optimization Neighbor optimization, only for cell-level NEs RF neighbor optimization

Primary pilot optimization

Reduce the soft handover ratio

Reduce soft handoff overheads

Parameter optimization Congestion control

Load balancing

DRBC

Power control and HSPA paremters

Part 3 corresponds to Chapter 4 of this guide, mainly introducing the high-load network expansion. As shown in Figure 1-2, the high-load WCDMA network will be respectively expanded in the three levels of cell, Node B and RNC. The content contains the expansion analysis process, expansion criteria, expansion methods and implementation details. Reading guide: If you want to understand the high-load network optimization measures given in this guide, please directly go to Chapter 3. If you want to understand the expansion criteria and methods, please directly go to Chapter 4. Any question about the counter or indicators, you can directly refer to Chapter 2 or understand by other means.



5



2

High-Load Network Monitoring

2.1

WCDMA NE Classification The NE level of WCDMA system can be classified into the RAN cell, Node B, RNC and CN. We mainly focus on the load monitoring and evaluation of the three NE levels of the RAN cell, Node B and RNC. For all NEs, we need to consider various scenarios of service behaviors, find the reasonable monitoring indicators and set the monitoring threshold, so as to perform the monitoring, alarm and load control. For the same service behavior, different NE has different monitoring parameters. As to the air interface, we mainly study the factors such as the cell throughput, single-user throughput, downlink power, uplink interference and downlink code resource. As to the NodeB, we mainly consider the utilization of hardware resources. 1.

Cell level The monitoring parameters of the NE in cell level mainly aim at the air interface, such as the cell throughput, average throughput of HSDPA users, average utilization of non-HSDPA carrier transmission power, average utilization of non-HSDPA code resources and the uplink noise rise.

2.

Node B level The NE in Node B level mainly monitors the utilization of Node B hardware resources, such as the utilization of uplink/downlink Node B CE resources, and the utilization of Iub interface uplink/downlink bandwidth.

3.

RNC level The NE in RNC level mainly monitors the utilization of hardware resources, including the CPU load (control plane), CE resource utilization (user plane) and bandwidth utilization (interface board).



6



The RNC will also observe the traffic operation indicators of the existing network, including the Erl, traffic, BHCA and the quantity of online users.

2.2

Network Load Monitoring Indicators According to the classification of three NE levels of WCDMA system, each NE respectively corresponds to different monitoring indicators and thresholds of the network load performance. Please refer to the document attached below.

Expansion Monitoring Indica

Unlike the non-flexible resource and load of Node B and RNC, the cell-level load monitoring indicators are the most complex, so following we will mainly introduce the cell-level load monitoring indicators.

2.3

Key Indicators

2.3.1

Average Utilization of Non-HSDPA Code Resource Average occupancy of cell code resources = Quantity of code resources occupied by all cell services/Total number of code resources It basically reflects the overall utilization status of the cell code resources. Average occupancy of cell non-HSDPA code resources = Quantity of code resources occupied by non-HSDPA service/Total number of non-HSDPA service code resources To some extent, it reflects the utilization status of R99 service code resources. The background network management can directly calculate the average availability of the cell code resources and the average occupancy of the HSDPA code resources. According to the two indicators, we can get the statistics formula of the average occupancy of non-HSDPA code resources, as shown below:



7



Average occupancy of non- HSDPA code resources = (1 – Average availability of the cell code resources – Average occupancy of HSDPA code resources)/(1 – Average occupancy of HSDPA code resources) For the HSDPA code channel, our system can perform dynamic adjustment according to the R99 service requirement. When the R99 traffic grows, the system can dynamically reduce the HSDPA code channels to be the minimum distribution value. So the average occupancy of non-HSDPA code resources can also be expressed as: Average occupancy of non- HSDPA code resources = (1 – Average availability of the cell code resources – Average occupancy of HSDPA code resources)/(1 – Minimum HSDPA code channel) For the hybrid carrier cell of R99+HSPA, the maximum code channel of R99 service is affected by the configuration parameters of HSDPA code channel. When the quantity of code channels that will be occupied by the R99 service exceeds the maximum of available code channels of the R99 service, the R99 service will refuse to receive due to insufficient DCH code resources. Maximum of available code resources of R99 service = 256 – Quantity of code channels occupied by common channel – Minimum of HS-PDSCH code channels × 16 – Quantity of HS-SCCH code channels × 2 – Quantity of E-AGCH code channels × 1 – Quantity of E-RGCH code channels × 2 Among which, the quantity of code channels occupied by common channel, the minimum of HS-PDSCH code channels, the quantity of HS-SCCH code channels, the quantity of E-AGCH code channels and the quantity of E-RGCH code channels come from the background network management configuration. Suppose the code channel parameter configuration of the HSPA and common channel is as shown in Table 2-1.

Table 2-1

Code Resource Distribution of Code Channel

Channel

Spreading Code

Quantity of Code Channel

HS-PDSCH

16

8 at least

HS-SCCH

128

2



8


E-AGCH

256

1

E-RGCH

128

1

CPICH

256

1

PCCPCH

256

1

SCCPCH

64

1

PICH

256

1

AICH

256

1


We can see that the maximum of available code resources of R99 service = 256 – 8 – 8 × 16 – 4 – 1 – 2 = 113. Because the quantity of code resources occupied by common channel is fixed and small, usually we can ignore it for calculation. So, when the utilization of cell code resources is very high, and even a congestion of code resource occurs, we suggest you reduce the minimum of HSDPA code channel to be 1.

2.3.2

Average Utilization of Non-HSDPA Carrier Transmit Power The average utilization of cell non-HSDPA carrier transmit power can be calculated by the statistic indicators in the network management. Here is its definition: Average utilization of cell non-HSDPA carrier transmit power =Total downlink transmit power of cell non-HSDPA code / Total downlink available power of cell non-HSDPA For the hybrid carrier cell of R99+HSPA, when it determines to control the acceptance of the DCH based on downlink power, one of the determination conditions is: NOHSDSCHPower + deltaP ≤ R99 admission threshold Among which, NOHSDSCHPower is the Transmitted carrier power of all codes not used for HS-PDSCH or HS-SCCH transmission reported by the NodeB. Currently, the acceptance threshold of R99 is usually set to be 85%, that is, 85% of the maximum transmit power of the cell. When the RNC determines to accept the DCH based on downlink power, if there are multiple requests of establishing connection at the determination time, the system will add the predictive power of all the newly-established connections based on the existing NOHSDSCHPower, and then compare with the power



9



acceptance threshold. When the predictive power is bigger than the acceptance threshold, the system will reject all the requests of establishing connection. If there are many requests of establishing connection, the predictive power deltaP will be large and it is easy to refuse to accept.

2.3.3

Uplink Noise Rise In WCDMA system, all the users share the same carrier, and the users are distinct from each other by different spreading code and scrambling code. For the uplink, due to the non-orthogonality of the user scrambling code, each user signal becomes a noise (interference) to other user signals. Therefore, each signal is included in the broadband interference background generated by other users. To access a call, the mobile station power must be large enough to overcome the noise of other mobile stations in the bandwidth. The relationship between the uplink capacity and noise rise is as shown in Figure 2-1.

Figure 2-1

Relationship Between the Uplink Capacity and Noise



10



From the figure you can see that, there is a non-linear relationship between the NodeB uplink noise rise and uplink capacity (load). When the uplink capacity (load) reaches a certain threshold, the noise rise will increase sharply. Therefore, th e UMTS radio network planning is based on certain uplink load planning. Generally the uplink load is designed to be 50%, corresponding to 3db of noise rise. When the uplink load is too large, both the system uplink coverage and performance will obviously deteriorate due to the sharp noise rise The indicator of cell uplink noise rise cannot be calculated directly from the network management. It formula is defined as follows: Cell uplink noise rise = Average value of cell carrier received power – System noise floor

2.3.4

Average Throughput of HSDPA Cell Mainly from the perspective of the total HSDPA throughput, we use the average throughput of HSDPA cell to evaluate whether the cell is busy, and determine whether the cell needs to be expanded by considering the Average Throughput of HSDPA Single User at the same time. Average throughput of HSDPA cell = Amount of user data confirmed by HSDPA MAC. The unit is Kb. It indicates the average t hroughput of HSDPA cell in the data transmission time. If the average throughout of HSDPA cell is small, you need to analyze whether it is because of poor coverage or insufficient transmission, or because the service demand of the cell user is small, such as QQ online service. If the small data amount of the user scheduling is caused by poor coverage or insufficient transmission, you need to optimize in the perspective of coverage so as to improve the overall cell throughput. Only when the HSDPA cell average throughput is relatively large, you need to further assess the Average Throughput of HSDPA Single User.

2.3.5

Average Throughput of HSDPA Single User For the HSDPA data service, except for the traditional indicators such as call connection rate and call drop rate, there is another more important indicator used to measure the



11



user experience, that is, user average download rate. When the user experience rate of the HSDPA users is below expectations, we need to optimize and expand the network. When the average experience rate of the HSDPA users cannot meet expectations, except for the possible causes mentioned above that the network coverage is poor or the transmission bandwidth is insufficient, there is another cause that too many users initiate the data transmission at the same time. If the low HSDPA user rate is caused by too many users initiating the data transmission at the same time, we need to optimize and expand the network. HS-PDSCH is a shared physical channel, and the transmission bandwidth is shared by all the HSDPA users. If too many users initiate the data transmission at the same time, the real-time transmission rate of each HSDPA user will reduce. Therefore, except for the indicator of HSDPA user real-time experience rate, the system should also provide the indicator of real-time transmission HSDPA user quantity, which is used to judge whether the low real-time transmission rate of each HSDPA user is caused by too many HSDPA users initiating the data transmission at the same time. The Average Throughput of HSDPA Single User is defined as follows: Average throughput of HSDPA single user (Kbps) = Amount of user data confirmed by HSDPA MAC/Data transmission time of HSDPA users



12



3

High-Load Network Optimization

3.1

Network Load Optimization Stages High-load network will cause many problems such as the access failure, handover failure, call drop and HSPA low rate, badly affecting the user experience and thus needing to be optimized or expanded urgently. Figure 3-1 shows the high-load network optimization stage, that is, after the network load rise and before the network expansion. When the network load is monitored to be high, we first need to optimize to reduce the network load. If the load is still relatively high after the network optimization, we need to prepare for expansion.

Figure 3-1

Flowchart of Network Load Optimization

Horizontal: RNC, NodeB, and cell Vertical

1. High-load network monitoring User perception evaluation


2. High-load network optimization RF optimization

Parameter optimization (handover threshold, congestion control, and load balancing)

3. Expansion decision and implementation Expansion decision -> Expansion i mplementation



13


3.2


Network Load Optimization Process The network load optimization process is as shown in Figure 3-2. Actually the optimization of high-load network aims to the cell air interface resources. The cell load rise can be solved by RF optimization and parameter optimization. The RF optimization mainly aims to the coverage, neighbor cell and interference optimization, so as to reduce the excessive resource consumption resulted from overshooting, pilot pollution and high-proportion switching. The parameter optimization includes the switching threshold optimization, control methods of congestion acceptance (refuse and preempt), load control, DRBC, power control and HSPA related parameters, as well as the intra-system and inter-system cell load equalization. These optimizations can not only reduce the cell load, some optimization methods can also reduce the Node B and RNC load, such as the switching and DRBC downspeeding. Relatively speaking, the load rise of Node B and RNC belongs to the consumption of its own hardware resource. Please note that, some optimization methods are especially for some kind of resource or indicator, but may have a negative impact on another resource or indicator. For example, by reducing the HSDPA code resource we can reduce the non-HSDPA code resource utilization and R99 service congestion, but meanwhile the HSDPA service rate and user experience will also be reduced. So during the optimization, we need to comprehensively consider the balance of various optimization methods and assessment indicators. If there are still some indicators being limited after the optimization, we need to prepare or implement the expansion.



14


Figure 3-2


High-Load Network Optimization Process

High-load NE Cell-level NEs involve false load rises and real load rises. Vertical

RF optimization Neighbor optimization, only for cell-level NEs RF neighbor optimization

Primary pilot optimization

Reduce the soft handover ratio

Reduce soft handoff overheads

Parameter optimization Congestion control

Load balancing

DRBC

Power control and HSPA paremters



15


4


High-Load Network Expansion According to the development experience of the fixed broadband network, the data service will grow explosively when it comes to a certain stage. But when will the explosive turning point come relates to the tariff policies, terminal development status, network quality and user behaviors, and thus it is difficult to predict. Therefore, we suggest that the expansion indicator threshold setting can be divided into two stages: the monitoring threshold and expansion threshold. In this way sufficient space can be left for the expansion. The monitoring threshold means that, when the indicator reaches this threshold, you need to prepare related expansion resources. When the expansion threshold is reached, you need to implement corresponding expansion actions. We also suggest you pay attention to relevant factors such as the tariff, terminal, network quality and publicity. When relevant strategies change, you should consider the possibility of expanding the network in advance.

4.1

Expansion Process

4.1.1

Expansion Analysis Process For the network expansion, you can begin with the network load monitoring, respectively perform corresponding monitoring, analysis and alarm for each level of NE, and expand the NEs meeting the expansion criteria by proper expansion methods. The expansion analysis process is as shown in Figure 4-1.



16


Figure 4-1


Expansion Analysis Process

Yes

Cell-level monitoring, analysis, and alarm

High-load network performance monitoring and evaluation

NodeB-level monitoring, analysis, and alarm

Perform optimization to solve the problem of high load

RNC-level monitoring, analysis, and alarm

No

Meet the cell capacity expansion requirements or not?

Yes

Cell expansion Add carriers or power amplifiers

Meet the NodeB capacity expansion requirements or not?

Yes

NodeB expansion Add BPC boards or transmission resources

Meet the RNC capacity expansion requirements or not?

Yes

RNC expansion Software: Add licenses Hardware: Add RUB or RCB boards

No

4.2

Expansion Criteria and Methods For the WCDMA system, the high-load network expansion needs to respectively aim to three NEs of the RAN cell, Node B and RNC. The cell load level only reflects the load status of the cell itself to some extent. A Node B can have many cells and the different quantity of cell results in different load of Node B. If a Node B contains too many cells, although the cell itself does not have too much load, the Node B ’s load may exceed the limit. Similarly, the RNC load is affected by the quantity of its Node B and cells. So you need to assess all the three NEs, and formulate different expansion criteria and methods correspondingly. The expansion criteria mainly include the expansion threshold and expansion assessment formula, and the expansion methods respectively correspond to the limit of different NEs and resources.

4.2.1

Cell Expansion Among the three NEs, th e WCDMA system cell is the NE closest to the actual us ers and the minimum unit used to assess the network load. The cell load and performance level directly affects the user experience, so the cell load monitoring and assessment will be the key point in our daily monitoring and assessment, and the cell expansion is also the core content of the WCDMA network expansion.



17


4.2.1.1


Cell Expansion Decision Process As shown in Figure 4-2, the cell load decision focuses on the user experience, and decides the cell load by combining the utilization of network resource indicators.

Figure 4-2

Cell Expansion Decision Process Network load monitoring To evaluate user perception: Average packet user-perceived rate; Uu interface congestion conditions

No Low transmission rate and congestion Yes

Low resource utilization rates

Resource utilization evaluation: code resources and power resources

To evaluate network resources: Utilization rates of UL and DL power resources; Utilization rate of code resources High resource utilization rates

False load rise

No

Real load rise

Optimization

High-load decision Yes Expansion

The indicators of assessing the user experience are mainly the data user experience rate and cell resource congestion level. The network resources mainly refer to the air interface code resource and power resource. For details please refer to the cell indicators mentioned in Section 2.2.

4.2.1.2

Expansion Thresholds and Methods The user experience is the most direct and effective reflection of the network load level. In the past, the user experience was mainly assessed by some traditional indicators such as the call connection rate and call drop rate. But for the 3G network, the increase of data service users is an inevitable trend and the data service proportion will be bigger and



18



bigger. Therefore, the user experience of data service will also become the most important factor to measure the 3G network load, and the best indicator to assess the user experience of data service is the user download experience rate of data service. According to the expansion principle of ―Focus on the user experience‖, we will regard the HSDPA user average experience rate (throughput) as the core to assess the cell load, and try to accurately assess the cell load by combining the air interface. Each monitoring indicator of the cell load assessment has been set an expansion indicator number. The expansion indicator SPI is a logical indicator, and the value can only be 0 or 1. When the expansion indicator SPI reaches the threshold, the value will be 1, or else 0. We also provide corresponding expansion methods when each indicator reaches the expansion threshold, for your reference.

Table 4-1

Cell Expansion Thresholds

Expansion Indicator

Indicator Name

No.

Alarm

Expansion

Threshold

Threshold

≤1 Mbps

≤512 Kbps

Expansion Method

Average SPI1

Throughput of HSDPA Single

HSPA+/Multi-carrier/add

User

NodeB HSDPA cell SPI2

average

≥100 MB

≥150 MB

≥60%

≥70%

≥60%

≥70%

≥6 dB

≥8 dB

throughput Non-HSDPA code SPI4

resource average

Multi-carrier/add NodeB

occupancy Average utilization of SPI5

non-HSDPA carrier transmit

Expand power amplifier/add NodeB

power SPI6

Uplink noise rise




19



Admission SPI8

rejection

Set to be a

proportion due to

fixed

limited downlink

value: 1

≥2%


code resources Admission SPI9

rejection

Set to be a

proportion duo to

fixed

limited downlink

value: 1

≥2%

Expand power amplifier/add NodeB

power TCP According to the threshold setting of the cell load monitoring indicator SPI and the cell expansion assessment process, we can get the combination chart of the cell expansion decision formula, as shown in Figure 4-3.



20


Figure 4-3


Combination Chart of Cell Expansion Decision Formula

High-load cell decision

User perception evaluation SPI1 × SPI2 = 1 The cell has a heavy PS service load and a low userperceived transmission rate; the cell may be a high-load cell.

SPI8 × SPI9 = 1 The cell has severe congestion and user perception is bad; the cell may be a high-load cell.


SPI × SPI2 = 1

SPI4 × SPI8 + SPI5 × SPI9 + SPI6 × (SPI4 + SPI5) > 1

Yes

Yes

The cell has a high network resource utilization rate and a high PS service load; user perception about PS service is bad. Hence this cell is a highload cell.

The cell has a high resource utilization rate and a high nonHSDPA service load; user perception about access is bad. Hence this cell is a high-load cell.

No False load rise

Real load rise

Optimization

From above we can get the general formula of the high-load cell decision: S_cell = SPI1 × SPI2 + SPI4 × SPI8 + SPI5 × SPI9 + SPI6 × (SPI4 + SPI5) Formula description: 1.

S_cell is the cell load index.

2.

SPI1 × SPI2 is mainly used to filter the high-load cell focusing on the data service, that is, need to meet the requirements of low user rate and high cell throughput.



21


3.


SPI4 × SPI8 + SPI5 × SPI9 + SPI6 × (SPI4 + SPI5) is mainly used to filter the high-load cell focusing on the non-HSDPA ser vice. The purpose of SPI1 × SP I2 is to perform mutual correction of two counters. Two SPIs meeting the criteria can basically determine that the cell is in a high-load state. SPI4 × SPI8 means the average utilization of non-HSDPA code resources is relatively high and the situation of refusing to accept is serious. If the average utilization of non-HSDPA code resources is relativel y high but there is no situation of refusing to accept, it means that, although the load is high, it does not meet the expansion criteria. If the situation of refusing to accept is serious but the average utilization of non-HSDPA code resources is not high, it may be caused by the virtual load rise due to improper resource allocation. SPI5 × SPI9 means the same as SPI4 × SPI8. SPI6 × (SPI4 + SPI5) means at least two indicators meet the criteria. There are two causes, one is that the uplink may be interfered, and the other is that the automatic noise floor update is false. So we use the two indicators of the code resource and power resource to correct, so as to ensure the cells we filtered are really the cells with relatively high load.

4.

When S_cell > 0, it means that the cell enters a high-load state and needs to be expanded, and we need to perform monitoring optimization and load assessment.

5.

The bigger value of S_cell means the heavier load of the current cell. The minimum of S_cell is 0 and the maximum is 5.

4.2.1.3

Threshold Setting Methods and Foundations For different networks, the expansion thresholds may be different. Following is the brief introduction to the threshold setting of each indicator.

4.2.1.3.1

Average Occupancy of Non-HSDPA Code Resources As shown in Figure 4-4, for the cell with the CS traffic of the whole network in a certain area greater than 1Erl, the average occupancy of non-HSDPA code resources reflects the cell R99 traffic level to some extent, and is in proportional to the cell CS traffic. So in



22



the cell load assessment, when the indicator of average occupancy of non-HSDPA code resources is used, it means the indicator of CS traffic is indirectly used too. The average occupancy of non-HSDPA code resources not only reflects the occupancy of cell R99 code resources and the situation of refusing to accept, but also reflects the cell CS traffic load level.

Figure 4-4

4.2.1.3.2

Relationship Between the Code Resource Utilization and Traffic

Average Utilization of Non-HSDPA Carrier Transmit Power In some networks, when the average utilization of non-HSDPA carrier transmit power is greater than 40%, there will be a situation of refusing to accept due to the limited downlink power. It relates to the measurement and decision cycle of refusing to accept — 2 ms. If too many services are accepted in 2 ms at the same time, it will cause the situation of refusing to accept.



23


Figure 4-5


Relationship Between the Cell Carrier Transmit Power Utilization and TCP

Limited Proportion

4.2.1.3.3

Uplink Noise Rise Definition of uplink noise rise: Total average received power of cell uplink-RTWP NodeB noise floor Currently ZTE uplink acceptance control switch is closed, but the HSUPA scheduling is controlled by the parameter of MaxRTW P. The default configuration of MaxRTWP is 6dB. We suggest setting the expansion threshold of the uplink noise rise to be 8dB (corresponding to 85% uplink loads). Theoretically, 6dB means the cell has 75% uplink loads, obviously not indicating a high-load load. But 8dB cor responds to 85% uplink loads. So we suggest setting the expansion threshold of the uplink noise rise to be 8dB and setting the alarm threshold to be 6 dB.

4.2.1.3.4

HSDPA User Average Throughput & HSDPA Cell Average Throughput When the real-time experience rate of the HSDPA users cannot meet expectations due to the capacity reason, we need to expand the network capacity. The real-time experience rate of HSDPA users can be directly obtained from the network management background. Besides, the low HSDPA user rate may be caused by the poor coverage, insufficient transmission bandwidth and heavy network load. We need to expand the network capacity only when the low HSDPA user rate may be caused by the heavy network load. Therefore, except for monitoring the Average Throug hput of HSDPA Single User, we also



24



need to monitor the HSDPA cell average throughput, and use the two indicators to determine whether the network capacity needs to be expanded. When the Average Throughput of HSDPA Single User is less than 512 Kbps, we need to make the next-step decision of the capacity monitoring. The HSDPA cell average throughput indicates the service volume of cell data transmission. The HSDPA cell average throughput is too low may be because the application layer flow is not enough or the cell coverage is poor. In this situation, we should not perform the expansion. Therefore, we suggest considering the expansion when the HSDPA cell average throughput > 150 MB. In general, when the Average Throughput of HSDPA Single User is less than 512 Kbps and the HSDPA cell average throughput is greater than 150 MB, the cell capacity should be expanded.

4.2.1.3.5

Admission Rejection Proportion When the call congestion ratio is over 2%, the user experience will be badly affected. Therefore, we set the alarm and expansion threshold of this KPI as 2%.

4.2.1.4

Expansion Implementation Rules The NodeB expansion implementation rules mainly set the hour as the granularity. The monitoring and assessment cycle is 1 week. Because each cell has different user behavior and different busy hours, we recommend implementing 7 × 24 hour monitoring mode. The implementation rules are as shown in Table 4-2.

Table 4-2

Cell Expansion Implementation Rules

Monitoring

Monitoring Mode 1

Mode Monitoring Object Monitoring Granularity

Monitoring Mode 2

Cell of the whole network


Hour

Hour



25



A week (7 × N: N refers to the fixed Monitoring Cycle

A week (7 × 24)

busy hours of each da y, and the busy hour is set according to the existing network state.)

Utilization alarm threshold: If in 1 week, S_cell > 0, N ≥ 10, perform the monitoring Alarm

optimization and expansion

Monitoring

assessment.

Trigger

Utilization expansion threshold:

Condition

If in 1 week, S_cell > 0, N ≥ 10, add to the cell list of monitoring optimization and expansion assessment.

Expansion Trigger Condition

4.2.2

Utilization alarm threshold: If in 1 week, S_cell > 0, N ≥ 3, perform the monitoring optimization and expansion assessment. Utilization expansion threshold: If in 1 week, S_cell > 0, N ≥ 3, add to the cell list of monitoring optimization and expansion assessment.



Suppose Sn is the expansion

Suppose Sn is the expansion index,

index, Sn = S_cell_1 + S_cell_2

Sn = S_cell_1 + S_cell_2

+ ……S_cell_n (n = 7 × 24).

+ ……S_cell_n (n = 7 × 24).

When S ≥ 10, expand the cell.


Actually, the formula means, Sn

Actually, the formula means, Sn is

is the sum of S_cell meeting the

the sum of S_cell meeting the

criteria (S_cell > 0).

criteria (S_cell > 0).

Greater Sn indicates greater cell

Greater Sn indicates greater cell

expansion demand.

expansion demand.

Node B-CE Expansion Node B NE lies in the intermediate level of the 3-level NEs, mainly providing the baseband resource pool for the cell and performing the data transmission of the cell. This section mainly introduces the NodeB CE resource expansion.

4.2.2.1

CE Expansion Thresholds and Methods The Node B expansion mainly inspects the shared resource utilization of the cell under Node B, for example, the CE resource and transmission resource. This section mainly introduces the CE resource.



26



The Node B CE resource load assessment is as shown in the table. An indicator number SPI is set for each monitoring indicator. The expansion indicator SPI is a logical indicator and its value can only be 0 or 1. When the indicator reaches the expansion threshold, the value will be 1, or else 0. Meanwhile, we provide the expansion method corresponding to each indicator reaching the expansion threshold. The expansion threshold and method are as shown in Table 4-3.

Table 4-3

Node B CE Expansion Thresholds and Expansion Methods

Expansion Indicator

Indicator Name

No. SPI11

Average utilization of uplink NodeB CE resource

Alarm

Expansion

Expansion

Threshold

Threshold

Method

60%

70%

BPC board

Average utilization of SPI12

downlink NodeB CE

60%

70%

80%

90%

80%%

90%

resource SPI13

Maximum utilization of uplink NodeB CE resource Maximum utilization of

SPI14

downlink NodeB CE resource

SPI15

SPI16

Admission rejection rate of uplink CE

Admission rejection rate of downlink CE

Expand the

Set as a fixed

≥2%

value: 1 Set as a fixed

≥2%

value: 1

Expand the BPC board Expand the BPC board Expand the BPC board

Expand the BPC board

Expand the BPC board

According to the expansion threshold setting shown in Table 4-3, we use the expansion decision formula to assess the Node B load and expansion demand. S_nodeb_CE = (SPI11 + SPI3) × SPI15 + (SPI12 + SPI4) × SPI16 Formula description: 1.

S_nodeb_CE is the Node B CE expansion index.



27


2.


(SPI11+SPI3) × SPI15 is mainly used to filter the uplink CE high-load cell, indicating that the uplink average or maximum utilization is relatively high, and meanwhile the uplink CE refuses to accept. If there is only high utilization but no CE admission rejection, it means the indicator does not reach the expansion threshold. If there is only admission rejection but no high utilization, it may be caused by uneven distribution of resources.

3.

(SPI12+SPI4) × SPI16 is the same as above.

When S_nodeb_CE > 0, it means the NodeB enters a high-load state and falls into our monitoring scope for monitoring optimization and expansion assessment. Greater value of S_nodeb_CE means greater Node B expansion demand. The minimum of S_node_CE is 0 and the maximum can be 2.

4.2.2.2

Expansion Threshold Setting Method and Foundation When the average utilization of CE resources is 70%, we think the NodeB is in a high-load state and needs to be expanded. But as shown in Figure 4-6 and Figure 4-7, due to the independence of uplink and downlink CE resource, in the high-load state, the uplink CE resource utilization of some networks is much larger (even twice) than downlink CE resource utilization. So we need to perform joint monitoring for the uplink and downlink CE resource indicators but expand the uplink and downlink CE resources respectively.

Figure 4-6

Average Utilization Rates of Uplink and Downlink NodeB CE Resources in

Shapingba, Chongqing, China



28


Figure 4-7


Maximum Utilization Rates of Uplink and Downlink NodeB CE Resources in

Shapingba, Chongqing, China

4.2.2.3

Expansion Implementation Rules The NodeB expansion implementation rules mainly set the hour as the granularity. The monitoring and assessment cycle is 1 week. Because each NodeB has different user behavior and different busy hours, we recommend implementing 7 × 24 hour monitoring mode. The implementation rules are as shown in Table 4-4.

Table 4-4


Monitoring

Monitoring Mode 1


Monitoring Mode 2



Hour

Hour A week (7 × N: N refers to the fixed

Monitoring Cycle

A week (7 × 24)

busy hours of each day, and the busy hour is set according to the existing network state.)



29


Monitoring

Monitoring Mode 1

Mode

Utilization alarm threshold: If in 1 week, S_nodeb_CE > 0, N ≥ 10, perform monitoring Alarm


Monitoring

assessment.

Trigger


Condition

If in 1 week, S_nodeb_CE > 0, N ≥ 10, add to the NodeB list of monitoring optimization and expansion assessment. Utilization expansion threshold: Suppose Sn is the expansion index, Sn = S_nodeb_CE_1 + S_nodeb_CE_2 + …S_nodeb_CE_n (n = 7 × 24).

Expansion Trigger Condition

When S ≥ 10, expand the cell. Actually, the formula means, Sn is the sum of S_nodeb_CE meeting the criteria (S_nodeb_CE > 0). Greater Sn indicates greater cell expansion demand.

4.2.3


Monitoring Mode 2

Utilization alarm threshold: If in 1 week, S_nodeb_CE > 0, N ≥ 3, perform monitoring optimization and expansion assessment. Utilization expansion threshold: If in 1 week, S_nodeb_CE > 0, N ≥ 3, add the NodeB list of monitoring optimization and expansion assessment.

Utilization expansion threshold: Suppose Sn is the expansion index, Sn = S_nodeb_CE_1 + S_nodeb_CE_2 + …S_nodeb_CE_n (n = 7 × 24). When S ≥ 3, expand the cell. Actually, the formula means, Sn is the sum of S_nodeb_CE meeting the criteria (S_nodeb_CE > 0). Greater Sn indicates greater cell expansion demand.

IUB Transmission Expansion The IUB expansion also belongs to the second level of the radio network, responsible for the data transmission. Its capacity constraint will directly affect each KPI.

4.2.3.1

IUB Interface Transmission Expansion Thresholds and Methods The IUB resource load assessment is as shown in the table. The expansion indicator number SPI is set for each monitoring indicator. The expansion indicator SPI is a logical indicator and the value can only be 0 or 1. When the indicator reaches the threshold, the value will be 1, or else 0. The corresponding expansion method is also provided here for



30



your reference when each indicator reaches the expansion threshold. The expansion threshold and methods are as shown in Table 4-5.

Table 4-5

Iub Transmission Expansion Thresholds

Expansion Indicator

Indicator Name

No.

Alarm

Expansion

Expansion

Threshold

Threshold

Method

80%

90%

Maximum forward SPI17

SPI18

accepted bandwidth proportion of IP

Expand the

Maximum backward

transmission.

accepted bandwidth

80%

90%

60%

70%

60%

70%

80%

90%

80%

90%

60%

70%

60%

70%

proportion of IP SPI19

Average forward accepted bandwidth proportion of IP Average backward

SPI20

accepted bandwidth proportion of IP Maximum forward

SPI21

accepted bandwidth proportion of ATM Maximum backward

SPI22

accepted bandwidth proportion of ATM Average forward accepted

SPI23

bandwidth proportion of ATM Average backward

SPI24

accepted bandwidth proportion of ATM

Expand the transmission. Expand the transmission.

Expand the transmission.




For SPI17 –24 in the above table, we need to start the measurement on the OMCB for at least 1 week, and then close. According to the expansion threshold setting in the above table, we can use the expansion decision formula to assess the IUB transmission load and the expansion demand, as shown below:



31



S_trans = SPI17 + SPI18 + SPI19 + SPI20 + SPI21 + SPI22 + SPI23 + SPI24 Formula description: S_trans refers to the IUB transmission expansion index. When S_trans>0, it means the transmission enters a high-load state and falls into our monitoring scope for monitoring optimization and expansion assessment. A bigger value of S_nodeb means bigger expansion demand of the IUB transmission. The minimum of S_trans is 0 and maximum is 8.

4.2.3.2

Expansion Threshold Setting Methods and Foundations In a high-load state, the expansion threshold of the Iub interface uplink and downlink transmission bandwidth utilization should be 70% of the total transmission bandwidth.

4.2.3.3

Expansion Implementation Rules The IUB expansion implementation rules mainly set the hour as granularity, and the monitoring and assessment cycle is 1 week. Because each NodeB user has different behavior and different busy hour, we recommend implementing the 7 × 24 hour monitoring, and the monitoring implementation rules are as shown in Table 4-6.

Table 4-6


Monitoring

Monitoring Mode 1


Monitoring Mode 2



Hour

Hour A week (7 × N: N refers to the fixed

Monitoring Cycle

A week (7 × 24)

busy hours of each day, and the busy hour is set according to the existing network state.)



32


Monitoring

Monitoring Mode 1

Mode

Utilization alarm threshold: If in 1 week, S_trans > 0, N ≥ 10, perform the monitoring Alarm


Monitoring

assessment.

Trigger


Condition

If in 1 week, S_trans > 0, N ≥ 10, add to the NodeB list of monitoring optimization and expansion assessment. Utilization expansion threshold: Suppose the expansion index is Sn,

Sn = S_trans_1 +

S_trans_2 + ……S_trans_n (n = Expansion

7 × 24 hrs).

Trigger


Condition

Actually, the formula means, Sn is the sum of S_trans meeting the criteria (S_trans > 0). Greater Sn indicates greater cell expansion demand.

4.2.4


Monitoring Mode 2

Utilization alarm threshold: If in 1 week, S_trans > 0, N ≥ 3, perform the monitoring optimization and expansion assessment. Utilization expansion threshold: If in 1 week, S_trans > 0, N ≥ 3, add to the NodeB list of monitoring optimization and expansion assessment.

Utilization expansion threshold: Suppose the expansion index is Sn, Sn = S_trans_1 + S_trans_2 + ……S_trans_n (n = 7 × 24 hrs). When S ≥ 3, expand the cell. Actually, the formula means, Sn is the sum of S_trans meeting the criteria (S_trans > 0). Greater Sn indicates greater cell expansion demand.

RNC Expansion The RNC is at the highest level of the radio network, responsible for the work scheduling and processing of NodeBs and cells in its charge. In the perspective of software and hardware constraints, the RNC expansion can be divided into RNC hardware expansion and RNC sof tware expansion. RNC hardware expansion refers to the expansion triggered by the constraint of RNC hardware processing capability. The expansion can be performed by increasing the hardware boards.



33



RNC software expansion means that the software license is close to or reaches the committed capacity and thus the expansion is triggered. The expansion can be performed by increasing the software licenses. The RNC hardware expansion and software expansion may occur at the same time or occur respectively. Their association depends on the project hardware configuration mode and the software quotation mode. We need to monitor each project respectively according to related parameters of the RNC hardware expansion and RNC software expansion. In the perspective of modeling configuration, the RNC expansion can be divided into modeling expansion and non-modeling expansion. The modeling expansion means that the RNC hardware and software use the modeling configuration quotation. The expansion will be performed in the unit of model. The non-modeling expansion means that the RNC hardware and software do not use the modeling configuration quotation. The expansion will be performed in the unit of board.

4.2.4.1

Expansion Thresholds and Method

4.2.4.1.1

RNC Hardware Expansion The RNC hardware can be classified into the common hardware, capacity hardware and interface hardware. Expansion of common hardware: The common hardware mainly includes rack, frame and common board. 

Rack: The rack expansion depends on the quantity of frame. Each 4 frames need 1 rack.



Frame: including the control frame, resource frame and exchange frame



The expansion of control frame depends on the increase amount of the control plane processing board RCB. When there is the exchange frame, the main control frame can be inserted 6 RCBs, and the rest can be inserted 14. When



34



there is no exchange frame, the main control frame can be inserted 2 RCBs and the rest can be inserted 14.



The expansion of resource frame depends on the increase amount of the user plane processing board RUB and the interface board. Each resource frame can be inserted 15 RUBs and interface boards.



For the exchange frame, the system configures 1 exchange frame at most. When there are more than 2 resource frames, the exchange frame must be configured. When there are 2 or less resource frames, it is defaulted and recommended to configure the exchange frame.



Common board: including the global processing board and system exchange board



Global processing boards: including ROMB, CLKG and SBCX. The quantity is a fixed configuration and has nothing to do with the capacity, so usually there is no issue of expansion. If no active and standby boards are divided at the beginning, later we need to expand them to be active and standby according to the operation’s requirement.



System exchange boards: including THUB, GUIM, UIMC, PSN and GLI. Configuring a pair of THUB for the whole RNC is a fixed configuration. A pair of GUIM is configured for each resource frame. A pair of UIMC is configured for each control frame or exchange frame. A pair of PSN is configured for each exchange frame. A pair of GLI is configured for every 2 resource frames.

Expansion of capacity hardware: The capacity hardware can be divided into the control plane processing board RCB and the user plane processing board RUB. 

The monitoring indicators of control plane hardware expansion (RCB expansion) include: i.

RCP CPU load

ii.

NodeB quantity



35


iii. 


Cell quantity

The monitoring indicators of user plane hardware expansion (RUB expansion) include: RUP CE resource utilization

The monitoring indicators of the hardware expansion of RNC capacity hardware resource are as shown in Table 4-7.

Table 4-7

Monitoring Indicators of RNC Hardware Expansion

Indicator

Indicator Name

No. SPI31

SPI32

Average utilization of RUP CE resources Maximum utilization of RUP CE resources

Monitoring

Expansion

Expansion

Threshold

Threshold

Method

60%

70%

80%

90%

SPI33

RCP CPU average load

60%

70%

SPI34

RCP CPU peak load

80%

90%

NodeB quantity

140/pair of RCB

Cell quantity

420/pair of RCB

Expand the RUB board. Expand the RUB board. Expand the RCB board. Expand the RCB board. Expand the RCB board. Expand the RCB board.

For the NodeB quantity and cell quantity, we do not set the monitoring counter. When expand the NodeB, we need to assess whether the RCB needs to be expanded. For the utilization of RUB CE resources and the load of RCB CPU, we need to set the monitoring indicators. An expansion indicator num ber (SPI31 –34) is set for each monitoring indicator, and the value can only be 0 or 1. When the indicator reaches the expansion threshold, the value will be 1, or else 0.



36



According to the expansion threshold setting in the table, we can assess the RNC load and expansion demand by the expansion decision formula, as shown below: S_hard_Ctrl = SPI31 × (1 + SPI32) S_hard_User = SPI33 × (1 + SPI34) Formula description: S_hard is the RNC expansion index, and the value can be 0, 1 and 2. When S_hard = 0, it means neither the peak nor average value meets the criteria. So we do not need to expand. When S_hard = 1, it means the average value meets the criteria but the peak does not, and the RNC enters a high-load state. So we need to expand the control plane or the user plane. When S_hard = 2, it means both the peak and average value meet the criteria, and the expansion is urgent. Except for the monitoring indicators mentioned above, we can also set some observation indicators to observe the actual network service state when the hardware is close to or reaches the expansion threshold, as shown in Table 4-8.

Table 4-8

Observation Indicators of RNC Hardware Expansion

Observation Indicator

Related Monitoring

Name

Indicator

Affected Board

BHCA

SPI33/SPI34

RCB board

CS traffic

SPI31/SPI32

RUB board

PS flow

SPI31/SPI32

RUB board

Quantity of online users

SPI31/SPI32/SPI33/SPI34

RCB board and RUB board

Expansion of interface hardware: There are several factors causing the interface hardware expansion, including: 1.

Capacity



37



2.

Separation of logical interface

3.

For example, in a Unicom project, the Iu/Iub interfaces in many provinces share the IP interface board in the beginning, later it is required that the interface boards of Iu/Iub interfaces should be separated. Therefore, we need to separately expand the interface boards without changing the capacity hardware. This kind of expansion is resulted from the operator ’s requirement and does not need any expansion foundation. We just need to re-calculate the flow of each interface after separation.

4.

Change of interface type

5.

For example, the ATM interface boards are used previously, now we need to increase the IP interface boards because the network develops to the all-IP technology. This kind of expansion is resulted from the operator ’s requirement and does not need any expansion foundation. We just need to re-calculate according to the new interface board algorithm.

6.

Quantity increase of NEs or Ports

7.

Many kinds of interface boards are related to the quantity of NEs and ports. For example, the interface boards will be increased by increasing the NodeB quantity, increasing the E1 quantity for each NodeB, increasing the Iur quantity, or increasing the Iu-flex function. In this case, we need to re-calculate the quantity of interface boards according to the new NodeB/port demands.

8.

Change of redundancy protection mode

9.

It is also triggered by the operator ’s demand. For example, we calculate the interface boards only according to the flow redundancy in the beginning, later the operator requires the port redundancy or board redundancy, or requires both the interfaces and boards are configured in 1+1 mode. We need to increase according to the demand.

For Case 2 to Case 5 mentioned above, we do not need to set the monitoring indicators, and perform the expansion correspondingly when it is necessary. For Case 1, we need to monitor the bandwidth usage of the interface boards. The monitoring parameters are as shown in Table 4-9.



38


Table 4-9

Observation Indicators of RNC Hardware Expansion Indicator Name

Average bandwidth utilization in the ATM interface board transmit direction

Average bandwidth utilization in the ATM interface board receiving direction Average bandwidth utilization in the IP interface board transmit direction Average bandwidth utilization in the IP interface board transmit direction Maximum bandwidth utilization in the IP interface board transmit direction Maximum bandwidth utilization in the IP interface board transmit direction

4.2.4.1.2


Monitoring

Expansion

Expansion

Threshold

Threshold

Method Expand the

70%

80%

ATM interface board Expand the

70%

80%

ATM interface board

70%

80%

70%

80%

70%

90%

50%

90%

Expand the IP interface board Expand the IP interface board Expand the IP interface board Expand the IP interface board

RNC Software Expansion The RNC software expansion is closely related to the quotation means of software feature. Each project has different quotation unit of the software feature, and in the same project, different software has different quotation unit. Software expansion in the dimensions of NodeB quantity and cell quantity: Some projects and some features are quoted in the units of NodeB quantity and cell quantity. If the NodeB quantity and cell quantity exceed the quotation quantity, we need to perform the software expansion. For example, in one project, the RNC hardware can support 210 cells and has 100 cells actually, and the software feature is quoted 100cells. Then, the increase of cell quantity will trigger the software expansion. When there are almost 210 cells, we need to trigger both the software expansion and hardware expansion at the same time. Software expansion in the dimension of NodeB CE:



39



In some projects, some features are quoted in the unit of NodeB ’s CE. So we need to monitor the CE operating indicators of all the NodeBs, and should trigger the software expansion when the indicators exceed the quotation unit. (For the monitoring indicators please refer to CE Monitoring Indicators of NodeB) Software expansion in the dimensions of Erl and flow: In some projects, some features are quoted in the unit of Erl and flow. So we need to monitor the indicators of Erl and flow, and should trigger the software expansion when the indicators exceed the quotation unit. (For the monitoring indicators, please refer to RNC Hardware Expansion Observation Indicators.) For example, in one project, the RNC hardware configuration is 250 Mbps and software is quoted 100 Mbps. When the flow indicator of existing network exceeds 100 Mbps, we need to expand the software license, that is, to quote the software feature for the increased flow. When the flow is almost 250 Mbps, we need to perform the hardware expansion.

4.2.4.1.3

Modeling Expansion and Non-Modeling Expansion For both the software expansion and hardware expansion, there are two methods: modeling expansion and non-modeling expansion. To use which method depends on whether the modeling method is used in the preliminary software and hardware configuration quotation. For the software expansion, the different between the modeling expansion and non-modeling expansion only lies in the expansion granularity. The modeling expansion must be based on the model granularity. For the hardware expansion, except for different expansion granularity, the difference between the modeling expansion and non-modeling expansion also lies in whether the user plane, control plane, and interface board are bound. If the non-modeling method is used, the control plane, user plane and interface board can be expanded respectively. For example, in one project, if the RCP CPU is monitored to be a little bit high-load but other indicators are normal, we only need to expand the RCB and do not need to add the RUB or interface board. Similarly, if the user plane and



40



interface board are monitored to be load-rising, we can also expand the RUB or interface board separately. If the modeling method is used, the control plane and user plane boards need to be linked according to the model. In some projects, the interface board is also contained in the model, so we need to link the control plane, user plane and interface board. For example, in one project, if the RCP CPU is monitored to be a little bit high-load but other indicators are normal, we need to expand the whole model to the upper capacity level but not only to expand the RCB.

4.2.4.2

Expansion Threshold Setting Foundation When the average utilization of RUP CE resources reaches 70% and the average load of CPU reaches 70%, it means the RNC becomes high-load and needs to be monitored and assessed for expansion.

4.2.4.3

Expansion Implementation Rules In the WCDMA network, the RNC’s reflection during busy hours is relatively obvious and uniform, so we can monitor the RNC load in two ways, as shown in Table 4-10.

Table 4-10

RNC Expansion Implementation Rules Implementation Rule 1

Monitoring Object Monitoring Granularity Monitoring Cycle Monitoring Trigger Condition


Implementation Rule 2

RNC of the whole network

RNC of the whole network

Hour

Hour

A week (7 × 24)

A week (busy hours of each day)

If in 1 week, S_hard ≥ 1, N ≥ 10,

If in 1 week, S_hard ≥ 1, N ≥ 3,

perform the monitoring

perform the monitoring



assessment.

assessment.


41

Zte umts Load Monitoring and expansionguide 141112012330 Conversion Gate02

Recommend Documents