Agilent - Actix DV For Agilent GSM GPRS
Agilent - Actix DV For Agilent GSM GPRS
(GSM/GPRS)
September 2003
Contents
Page 1
1 Introduction to Actix DV for Agilent
Page 1
1.1 Benefits of Actix DV
Actix DV has been designed to seamlessly process and analyze all data
collected by Agilent’s Nitro drive test equipment for GSM, GPRS, and
UMTS networks, quickly and easily, enabling users to enhance the
network performance and improve their operational efficiency.
Several sets of predefined reports, combined with an intuitive user-
interface, provide a one-stop analysis of crucial statistical data allowing
users to investigate the overall condition of a network within minutes of
data being collected. Actix DV saves time and resources in determining
and resolving network problems.
The business benefits of Actix DV for Agilent are that it:
The key functional benefits of Actix DV for Agilent are that it:
Page 2
1.2 Customers Addressed by Actix DV
Actix DV is targeted at users interested in ensuring the quality of voice
and data services on wireless networks and companies seeking to
maximize the return on their investments in wireless network
infrastructure. Actix DV is specifically focused on these organizations:
• Wireless System Operators
- Network Design Groups
- Network Performance Groups
- Network Operations Groups
• Wireless Network Infrastructure Providers
- Support Services
- Trial Teams
• Subscriber Unit Vendors
- Quality Assurance
- Field and Acceptance Testing
• Wireless Design and Optimization Consultancies and Turn-key
Solution Providers
- Optimization Specialists
- RF Design Engineers
Page 3
1.3 Operational Tasks and Processes
Typically, a new technology rollout goes through several phases,
including Lab Testing and Field Trails, Planning, Initial Rollout, Immature
Buildout and Mature Growth. The Lab Testing and Field Trials phase
begins first with pure design and laboratory testing and progresses to
increasingly more complex field trails.
Actix DV GSM/GPRS
Solution
Page 4
1.3.1 Rollout of Services
GPRS is typically rolled out as an overlay on an existing GSM network
where optimization is performed over a single air-interface medium.
During GSM initial rollout, operators often begin with site verification to
ensure basic operational state after each individual BTS is switched on-air.
This is followed by optimization of cluster areas and then a final network-
wide KPI (Key Performance Indicator) Drive, before the network is
commercially launched.
During the overlay of the GPRS service onto the GSM network, in
addition to measuring the data layer performance, the performance of the
voice layer needs to be closely monitored to ensure there has been no
degradation. This recursive process between voice and data requires
intensive data collection, efficient analysis of the collected data, and
accurate isolation of network problems in order for smooth service
upgrade.
Page 5
2 System Requirements
The following section provides the recommended and minimum system
requirements for successfully running Actix DV for Agilent.
If you install Actix DV on a system that does not meet the recommended
configuration requirements, you may experience the following limitations:
• You may be unable to process large amounts of data.
• You may experience delays when trying to perform specific tasks.
• You may not be able to take advantage of some mapping
capabilities.
• Actix DV’s advanced correlation and analysis capabilities may not
perform as desired.
Available Hard Disk Space 15 Gigabytes (about 80MB for DV; most of
this space is for drive test data files, map information, and virtual memory
needs).
Page 6
2.2 Minimum System Configurations
Processor (CPU): Pentium II 333-MHz or higher.
Available Hard Disk Space: 1 Gigabyte (for DV, drive test data files,
map information, and virtual memory needs).
Page 7
3 Feature Overview
3.1 Actix DV Features
Key measurement data is organized in predefined modules on the opening
screen of Actix DV providing users with an immediate comprehensive
overview of the network’s condition. This allows the visual management
of data in customized forms of graphic representation such as maps,
charts, and user-controlled panels for selection of specific events. Users
have the flexibility of displaying pertinent statistics simultaneously or only
showing pinpointed areas of interest.
Page 8
3.1.1 Event Navigator
Using advanced event detection, the Event Navigator displays the events
in the selected device on a timeline of the entire drive. It is the principle
device for 'driving' DV and allows users to either jump straight to an
event, or step along the timeline using the Replay panel. Each event has a
tool tip that displays the cause and radio conditions at that point.
A bar along the top shows (in this case) serving cell and traffic channel.
Page 9
3.1.2 Console View
The Console appears on the left of the DV application window, and
manages all loaded files and devices.
Page 10
3.1.3 Map View
The Map is the same fully-featured component as used in other Actix A
Solutions, and is built on the industry-standard MapInfo® MapX®. This
allows the use of advanced Actix features such as multiple configurable
legends, user-definable cell displays, lines-to-cells, and advanced printing
support, with Layout and Preview.
Layer Control
Page 11
3.1.4 Data Views
Data views or 'forms' provide many different ways of examining aspects
of data for a particular technology, and are provided in a pre-defined
arrangement in the DV workspace.
The following forms are provided (the Event Navigator is common to all
technologies):
GSM
• GSM Current Channel – provides a summary of
important serving channel and serving cell radio
parameters, including BCCH frequency, TCH frequency,
Hopping Frequencies, Current TS, MAIO, HSN, Speech
Codec.
GPRS
• GPRS Data Session – provides a snapshot of important
data parameters for both received and transmitted packets.
Page 12
3.1.5 Charts
The pre-defined chart window in Actix DV allows users to monitor the
trend of particular measurements with the option of adjusting the timescale
to a specified duration or to the entire drive data.
Note that you can alter the timescale by clicking on the drop-down list in
the lower right hand corner of the chart.
Page 13
3.1.6 Reports
Standard reports are also available in DV, containing statistical
information summarizing the data in the data file.
Page 14
3.1.7 Protocol Stack Browser
The Protocol Stack Browser window's top panel allows users to view
summary information on each message of a particular type for the active
device, while the bottom panel displays the full decode for the currently
selected message.
Page 15
3.1.8 Replay Tool
The Replay panel is used to control the playback of the log file, starting
from the beginning of the file and displaying the data message by
message, either one message per click of the inner buttons or playing
through the messages using the outer buttons, at a speed set using the
sliding control. This allows variable-speed, bi-directional playback, and
can also be used to step single messages.
Page 16
3.2 Applications for GSM
The set of reports contained in this application gives a summary of the
performances of voice traffic, and allows a verification of the radio
network design.
Call Details
Summary
CC Messages Count
CC Setup 6 Call Setup Success Rate (CSSR) 83.33% 100.00%
CC Call Confirmed 0 Call Completion Rate (CCR) 83.33% 83.33%
CC Connect 6 Handover Success Rate (HSR) 85.71%
CC Alerting 6 Location Update Success Rate (LUSR) No LU
Formulae
RR Messages Count
RR Handover Command 7 CSSR 1: (Calls Completed + Calls Dropped) / Call Attempts
RR Handover Complete 7 CSSR 2: Outgoing and Incoming Call Setup / Call Attempts
RR Handover Failure 0 CCR 1: Calls Completed / Call Attempts
CCR 2: Calls Completed / Calls Setup OK
HSR: Handovers Successful / Handover Attempts
MM Messages Count LUSR: Location Update Successful / Location Update Attempts
MM Location Updating Request 0
MM Location Updating Accept 0
MM Location Updating Reject 0
Page 17
3.2.2 Drive Test Measurement Charts
Measurement Reports contain information concerning receive level and, in
dedicated mode, the quality and timing advance. The statistical
distribution of these elements allows engineers to assess the overall quality
of the radio conditions on the test route.
Peaks in the distribution of level and quality measurements may already
give some indications about particular problems or specific network
settings related power control and interference. Timing advance
distribution helps in identifying average serving distance (the distance to
the BTS serving the test mobile) and can also indicate abnormally distant
connections.
RxLevSub Distribution
120 100%
90%
100
80%
70%
80
60%
60 50%
#
40%
40
30%
20%
20
10%
0 0%
-110 -106 -102 -98 -94 -90 -86 -82 -78 -74 -70 -66 -62 -58 -54 -50 -46 -42
dBm
RxQualSub Distribution
800 100%
700 90%
80%
600
70%
500 60%
400 50%
#
300 40%
30%
200
20%
100 10%
0 0%
0 1 2 3 4 5 6 7
RxQual
TA Distribution
500 100%
450 90%
400 80%
350 70%
300 60%
250 50%
#
200 40%
150 30%
100 20%
50 10%
0 0%
0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63
TA
Page 18
3.2.3 Drive Test Data Call Quality Summary
The Drive Test Data Call Quality Summary supplies users with an
overview of the collected call quality data through a predefined panel
containing pertinent statistics such as Received Signal Quality and
Handover rates. This allows users to quickly investigate the overall
behavior of a network while the graphical displays present users with a
comparative analysis of the voice quality correlating to the coverage level.
Users can now focus on areas of data that are of interest such as high
dropped call rates or handover failure rates by using the Event Navigator.
Data can be easily updated with a single click by selecting a key event on
the Event Navigator.
Page 19
3.2.4 GSM Current Channel and Radio Parameters
Summary
This application has been tailored to display the most significant channel
configurations and radio parameters which provide users with a clear and
quick diagnosis of the serving channel.
The key parameters defined are:
• Serving Cell MCC, MNC, LAC, CI,
• Serving Cell BCCH, BSIC, TCH, TSC
• Serving timeslot number, sub-channel number
• Hopping layer configuration
• Cell Allocation List
• RxLev (full/sub)
• RxQual (full/sub)
• Current and maximum radio link timeout
• Timing advance
• Transmitting step size
• Indication of downlink discontinuous transmission
• Frame Error Rate
Page 20
3.2.5 Neighboring Cells
To troubleshoot the handover (in dedicated mode) or location update (in
idle mode) behavior of a network the neighbor chart monitors each
collected data point along the drive-tested route in correlation with the
serving channel and its neighboring channels.
Typically, a neighboring channel with a stronger signal by a defined
margin than the serving channel, would cause the call to be transferred to
the stronger signal and the (stronger) neighbor cell would become the new
serving cell. This is known as the handover process. However, in a
situation where the neighbor relationships and controlling parameters are
not correctly defined the serving channel may fail to handover to the
stronger signal and consequently the call may degrade in quality and
ultimately be dropped.
The Neighbor Chart allows users to trace any abnormal handover activities
between the serving channel and the neighboring channels.
Rx Level (dBm)
The far left bar displays the RxLev of the serving channel of interest in
relation to its neighboring channels, with their individual RxLev shown by
the bars on the right.
Page 21
3.2.6 Example of Troubleshooting Radio Network
Performance using Actix DV
One of the more powerful advantages of Actix DV is allowing users to
overlay multiple layers of attributes on a single map with the flexibility of
displaying the attributes desired by the user. For example, in radio
networks with downlink quality problems, a single attribute view may
provide users with a partial analysis of the problem area.
Displaying RxQual by
itself gives users a
partial analysis of the
area
Page 22
RxLev < -89 dBm RxQual > 4
Users can further drill down into the quality problem by studying the
implemented frequency plan using Actix DV map window. After
importing site configurations into the Actix “cellref” file, each sector can
be color coded according to their assigned frequency. This map quickly
identifies situations where sectors with the same color code are most likely
causing co-channel interference.
Co-Channel
Sectors
Page 23
3.3 Application for GPRS
The pre-designed applications available for analyzing GPRS data are:
Page 24
• Total session duration
• Mean session duration
• Total TCP data received
• Total TCP data sent
• Mean end-to-end round trip delay
• Maximum end-to-end round trip delay
Page 25
3.3.2 DT GPRS Radio Link Performance Analysis
This application is dedicated to the analysis of the radio link performance
and contains the following reports:
• Coding Scheme and Throughput per number of used Time Slots
• Timeslot allocation (percentage of usage of the Time Slots)
• Signal level (C value)
• Quality
• Interference analysis
• Throughput RLC and LLC
• Throughput and Coding Scheme
• Throughput and signal level
• Throughput and quality
• Throughput per TBF
• BLER and Retransmissions
Page 26
3.3.3 Design Validation
A key performance indicator of a GPRS network is the RLC throughput
which can be analyzed in relation to the coding scheme usage and the
number of GPRS data timeslots occupied. Combining these specific
elements into a single analysis will verify the quality of a network design
and determine the quality of the radio condition along the drive test route.
In this report, the following information elements have been combined:
• RLC Throughput: this is usually good when the timeslot allocation
and coding scheme usage levels are both high.
• Timeslot Allocation: this is dependent on how many timeslots the
handset is capable of handling and the network has available for
allocation.
• Coding scheme: this is acceptable when it satisfies network
settings.
Based on these criteria, Actix DV generates a measurement report
indicating the coding scheme and throughput per number of used
timeslots. In the following example, the throughput is generally good
because it reveals that three timeslots are used with coding scheme 2 used
in each case.
Coding Schemes and Throughput per TS
File name: GPRS_Data_Test2.sd5
Logging Date: 11/3/2003
18 120
16
100
Mean DL RLC Throughput (kbps)
14
12 80
10
(%)
60
8
6 40
4
20
2
0 0
1 2 3
Number of DL Timeslots used
Page 27
3.3.4 Throughput Statistics
A statistical report calculates the minimum, average, and maximum RLC
and LLC throughput for both Uplink and Downlink activities allowing
engineers to evaluate the overall throughput quality of the collected data.
70.00
60.00 57.57
Throughput (kbps)
50.00
39.62
40.00
30.00
20.00 15.97
10.62
10.00
0.00 0.00
0.00
Min Mean Max
DL LLC DL RLC
50.00 46.27
45.00
40.00
Throughput (kbps)
35.00
30.00
25.00
20.00
13.38
15.00
10.00 5.81
3.81
5.00
0.00 0.00
0.00
Min Mean Max
UL LLC UL RLC
Page 28
3.3.5 Data Session Window
General information is found in the Data Session Window. Actix DV
provides a set of measurements outlining the throughput’s general
performance and presenting users with the appropriate information such as
Data Session Protocol and an indication of end-users perception. These
attributes are essential to analyzing maximum throughput and
troubleshooting a GPRS network.
The following attributes are available from the Data Session Window:
Page 29
Data Session Window showing both Uplink and Downlink signal activities
Page 30
3.3.6 Data Timeslot Window
To further investigate a network quality problem, users can shift to the
application “Data Timeslots”. For example, if there is a rapid degradation
in the RLC downlink throughput observed in the Data Session window,
users would typically be able to determine the possible cause by analyzing
the change in color scheme, timeslot usage level, and retransmission
percentile. Actix DV’s Data Timeslot application summarizes this detailed
data in addition to other related attributes of the serving channel, providing
users with more in-depth throughput analysis.
Data Timeslot Window showing one and three GPRS timeslots are utilized in Uplink and
Downlink respectively
Page 31
3.3.7 GPRS Throughput Chart
When troubleshooting the overall performance of a GPRS network, one of
the most effective ways would be through observing the trend of the
collected data measurements. Actix DV provides users with a multi-layer
Throughput Chart with scrolling capabilities, to enable monitoring of
throughput at a particular timeframe or for the entire duration of the
collected data. Users can also observe the trend of the downlink and
uplink throughput in respect to the allocated GRPS timeslots.
For example, users can observe the change in throughput quality during
downlink and uplink activities
Downlink throughput
went down
As Uplink activities
began
Page 32
3.3.8 Example of Troubleshooting GPRS Network
in Mapping Window
Utilizing Actix DV’s multiple attribute layering capability allows users to
quickly and easily identify the root-cause of downlink and uplink
throughput degradation. Typically this decrease is directly related to RF
conditions such as the Data Quality, Coding Scheme, and Retransmission
and can only be determined by tracing these multiple correlated attributes
simultaneously.
For example, during poor RF conditions, the BER may exceed the
threshold of a higher data rate coding scheme. This can force the data
session to switch to a lower data rate coding scheme to retain the integrity
of the link. Furthermore, an RF environment with interference may also
cause retransmission of the same data block to acquire acknowledgement
at the end-user terminal. Both of these cases will lead to a decrease in
RLC throughput.
Bad RxQual
Multiple
Retransmission
Coding
Scheme 1
Multiple attributes such as Data Quality, Coding Scheme, and Retransmission mapped
simultaneously illustrates the problem of increased retransmissions and the lowering of
the coding scheme due to poor RF conditions (see map above). The problem can be
solved by locating a cleaner frequency for the serving channel. However, if the GPRS
traffic is carried on the hopping layer, the MA list should be revised.
Page 33