-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
Product Name Confidentiality
eLTE Confidential
Product Version 22 pages in total
V2.3
Technical White Paper for System Reliability
Prediction of eLTE 2.3
(For internal use only)
Prepared by Liu Hao (employee ID: 00273140) Date 2014-06-03
Reviewed by Date
Approved by Date
Huawei Technologies Co., Ltd
All rights reserved.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 2 of 22
Change History
Date Issue Description Author
2014-06-20 1.01 Added the eCNS610-related
information.
Liu Hao (employee ID:
00273140)
2014-06-03 1.00 Completed the draft. Liu Hao (employee ID:
00273140)
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 3 of 22
Technical White Paper for System Reliability
Prediction of eLTE 2.3 Key words:
eLTE 2.3, reliability, eCNS600, eCNS610, DBS3900
Abstract:
This document describes the methods for calculating the
reliability indicators of the network elements (NEs)
in the eLTE 2.3 solution.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 4 of 22
1 Reliability Modeling
The reliability modeling involves the following steps:
1. Defining functions and physical components
2. Defining faults
3. Defining fault properties
4. Constructing reliability models
5. Determining the number of simulations and the life cycle
6. Determining assumptions in reliability prediction
1.2 Defining Functions and Physical Components
The first step for calculating system reliability is to
determine the physical components used for reliability
modeling. Not all components in the system need to be
incorporated in reliability modeling. The
components that do not impact key services, such as maintenance
terminals, can be ignored in reliability
modeling if not required. Component selection for reliability
modeling also depends on actual applications.
For example, if the voice broadcast system is used for
management assistance, which is not a key service,
the components in the system can be ignored in reliability
modeling. However, if the voice broadcast
system is involved in security tracing, the components need to
be incorporated in reliability modeling.
In system-level reliability modeling, typical components include
the eCNS, eNodeB, terminals involving
security services (for example, a vehicle-mounted terminal for
train control), transmission devices, antenna
feeder devices, and power supply modules. Sometimes, a mobile
terminal is ignored in system-level
reliability modeling, but its reliability can be independently
calculated.
1.3 Defining Faults
System reliability indicators have a close relationship with
system fault definitions. Different fault
definitions lead to different calculation results of reliability
indicators.
When system or subsystem faults are defined for NEs or boards
working in redundancy mode, the
redundancy type needs to be considered. For example, an eNodeB
fault is determined for an eNodeB with
three carriers working in 2+1 backup mode only when two or more
carriers are faulty. For a co-site network
hosting two eNodeBs working in 1+1 backup mode, a fault in only
one eNodeB is not considered as a
system fault.
However, the duration for active/standby switchover needs to be
considered in the preceding situations. For
details, see section 1.4 "Defining Fault Properties."
Besides hardware faults, other fault types may be involved in a
project, such as software faults, faults
caused by human errors, and faults caused by external factors
(for example, lightning stroke and violent
damage).
1.4 Defining Fault Properties
Common fault properties are defined as follows:
Failure distribution curve, failure rate, and
mean-time-to-failure (MTTF)/mean time between failures
(MTBF)
Mean time to repair (MTTR) and definition of repairs and spare
parts
Fault impact
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 5 of 22
Failure Distribution Curve, Failure Rate, and MTTF/MTBF
A failure rate is the average number of failures for a component
per unit of time after the component begins
to fail. Generally, a failure rate is expressed as .
Accordingly, the failure rate function is expressed as (t)
because the failure rate is a function of the time t. The
failure rate function corresponds to the failure
distribution curve, the value of which may vary according to the
component. The failure rate for an
electronic component within its life cycle is generally a
constant, and exponential distribution is used for
such a failure rate. The failure rate for a mechanical
component, however, varies depending on the time
segment. Weibull distribution is generally used for such a
failure rate.
MTTF indicates the average interval at which a fault occurs on a
component. MTTF is a reciprocal of a
failure rate expressed by exponential distribution.
To calculate the MTTF of a board to determine the reliability of
the board, the failure distribution curve (or
failure model) must be first determined, and then the failure
rate can be determined.
MTBF indicates the average time between consecutive failures of
a component, which can be calculated in
the following formula: MTBF = MTTF + MTTR.
MTTR and Definition of Repairs and Spare Parts
MTTR is closely related to service availability. The value of
MTTR equals the sum of Mean Active Repair
Time (MART) and Mean Logistical and administrative Delay Time
(MLDT). MART is related to factors
such as architecture design, fault management design, and board
startup time and is the intrinsic reliability
feature of components, while MLDT is the mean delay time to
ensure network maintenance, such as the
time spent on personnel dispatch, transportation, and
acquisition of spare parts. MLDT is related to the
staffing and equipment for the maintenance team.
The definition of repairs and spare parts involves resources and
spare parts required for a repair. With the
definition, the number of spare parts in the system and the
maintenance cost for the life cycle of a project
can be calculated. The transportation of spare parts also
affects MTTR.
Fault Impact
The fault impact determines the impact of a fault on the
functions and subsystem. The impact can be
classified into high, medium, and low levels. The impact of a
fault that affects the entire system and causes
a failure to provide services is a high-level impact. If a fault
leads to short-period service interruption in a
specific area served by multiple eNodeBs, the impact of the
fault is a medium-level impact. A low-level
impact indicates short-period service interruption for a single
service channel caused by, for example, ring
topology switchover after transmission interruption. An example
of a low-level impact is 2s voice service
interruption. The fault impact is defined so that it can be used
with the failure rate to determine the list of
critical items in the system, which facilitates the availability
optimization of the entire system.
1.5 Constructing Reliability Models
Huawei uses reliability block diagrams (RBDs) to construct
systematic reliability models for analysis and
calculation of reliability indicators.
The structure of an RBD indicates the logical relationships of
faults in a system. Each block indicates a
fault of a component, a subsystem fault, or an event that
affects the system fault. As for a subsystem fault,
the structure of another RBD can be used to indicate its
internal logical relationships. The logical flow of an
RBD starts from the input on the left and ends at the output on
the right. Between the input and the output
of the RBD, multiple blocks are arranged in serial or parallel
connections depending on the characteristics
of a system.
A system with serially connected blocks indicates that a fault
in any component will result in a system failure.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 6 of 22
Figure 1-1 System with serially connected blocks
Block 1 Block 2 Block 3
A system with parallel connected blocks indicates that the
components work in redundancy mode.
Figure 1-2 System with parallel connected blocks
Block 2
Block 1
Block 3
A complex system consists of both serial and parallel
connections, indicating that serially and parallel
connected subsystems exist in the system.
An RBD can also indicate the redundancy relationship of a
decision system (k out of n). As indicated by
the number 2 in Figure 1-3, at least two paths of the three
parallel paths must work properly. When two
paths are faulty, the system is faulty.
Figure 1-3 Decision system
Block 2
Block 1
Block 3
2
An RBD can also be used to analyze common cause failures. The
common cause failure indicates a fault
that can lead to failures of multiple functional components. If
there is a common cause failure in a system
with redundancy design, the failure must be expressed as a block
serially connected to other blocks in an
RBD. As shown in Figure 1-4, blocks 1 and 2 are failures of
independent components working in
redundancy mode, and the failure expressed by block 3 can lead
to simultaneous failures of the two
components.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 7 of 22
Figure 1-4 System with a common cause failure
Block 1
Block 2
Block 3
1.6 Determining the Number of Simulations and the Life Cycle
The number of simulations and the life cycle are involved as two
important parameters in the calculation of
the system unavailability based on the reliability model and
fault model. The number of simulations is the
number of times the system reliability model runs in simulation
mode. Each time a system reliability model
runs in simulation mode, the system determines whether it stably
runs for a specified period. The specified
period is the life cycle. For example, if the number of
simulations is 1000, the life cycle is one year, and the
number of times a fault occurs within the life cycle of the
entire system calculated based on the system
reliability model is 50, the system unavailability is 5%.
1.7 Determining Assumptions in Reliability Prediction
1.7.1 MTTR
MTTR is defined depending on the actual maintenance capability
in a project. For example, the in-transit
time for the repair of eNodeBs in a remote area is different
from that for the repair of eNodeBs near a city.
There is an assumption for the calculation of MTTR for each NE.
For details, see the reliability prediction
report of each NE.
1.7.2 Software Availability
Software reliability is the capability that the software
possesses to implement the required functionality
under specified conditions within a specified period. Software
reliability can be measured by availability.
Software availability is the probability for the software not to
cause system failures under specified
conditions within a specified period. Software reliability and
hardware reliability have many differences,
which are caused by the fault mechanisms of software and
hardware. Therefore, software reliability is
assessed differently from hardware reliability.
No mature quantitative measure is available for the analysis of
software availability in the industry.
Therefore, software availability is excluded from the
reliability models in this document. If software
availability needs to be considered in a project, an assumed
target value can be given to an associated NE.
1.7.3 External Factors
External factors are not considered in the reliability models in
this document.
In some projects, however, some external factors need to be
considered as required. For example, the
probability of optical fiber faults caused by rodents need to be
considered in areas where no rodent-free
measure is used for optical fibers. In areas flooded with
violence, the probability of violent damages to the
outdoor equipment needs to be considered.
Lightning protection measures are taken for Huawei outdoor
eNodeBs, and therefore the probability of damages caused by
lightning stroke can be ignored.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 8 of 22
1.7.4 Human Factors
Human factor analysis or human error analysis is used to measure
the impact of human errors on system
availability. Human errors are errors leading to system faults
caused by misoperations of system operation
personnel. Humans are flexible in executing tasks. Humans can
handle exceptions and faults at any time,
but human errors may also be introduced. Human errors mainly
include the failure to execute a specified
function, incorrect execution of a specified function, and
execution of an unspecified function.
Human errors are not considered in reliability models in this
document. In actual projects, the system
operation behaviors of humans can be determined based on the
application scenario of the system to
analyze the impact of human errors so as to calculate the
probability of human errors that may lead to
system faults. A large number of methods can be used to analyze
the probability of human errors, among
which there is a simple method that mainly depends on the
judgment of experts, that is, human error
assessment and reduction technique (HEART).
2 Reliability Prediction Methods
2.1 Board Reliability Prediction Methods
2.1.1 Component Reliability Prediction Method
The formula for calculating the component failure rate is as
follows:
TiSiQiGiSSi
where,
Gi is the generic steady-state failure rate for the ith
component.
Qi is the quality factor of the ith component.
Si is the stress factor of the ith component.
Ti is the steady-state temperature factor for the ith component
under normal working temperature.
Under 40C and 50% stress, S equals T, which has a value of 1.0,
and therefore the formula can be simplified as:
Ssi = Gi Qi
2.1.2 Method for Calculating the Board Failure Rate
The board failure rate is the sum of the component failure rates
and can be calculated in the following
formula:
n
i
SSiiESS N1
where,
n is the number of component types.
Ni is the number of components of the ith component type.
E is the board environment factor. For the fixed and controlled
environment on the ground, E has a value of 1.0.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 9 of 22
2.2 System Reliability Prediction Methods
The calculation of system reliability depends on the system
fault definitions to a large extent. Different
system fault definitions lead to use of different algorithms.
For example, some board faults in the system
lead to a failure of the entire system, whereas some board
faults lead to unavailability of only some
functions in the system. According to Bellcore SR-TSY-001171,
there are two kinds of system-level faults:
total system downtime (TSD) and partial system downtime
(PSD).
2.2.1 System Fault Definitions and Formula
TSD is an approximate time period during which the entire system
breaks down and fails to handle any
requests. TSD is generally expressed as minutes per year.
PSD is the weighted mean of the downtime during which the system
only partially fails. The weight factor
is the number of lines impacted by a specific failure mode.
The calculation formulas for the TSD and PSD are as follows:
Ti
LiNTNL
iDTSD
Li NL
Lii NLDPSD0
))/((
where,
Li is the number of subscriber lines impacted by failure mode
i.
Ti is the number of trunk lines impacted by failure mode i.
NT is the total number of trunk lines in the system.
NL is the total number of subscriber lines in the system.
Di is the predicted downtime of failure mode i, expressed as
minutes per year.
2.2.2 Calculation of Di
According to the preceding formulas, the calculation of system
reliability or system availability depends on
the calculated Di value of each board.
The impact of a failure mode is first analyzed to determine
which boards in the system will cause TSD and
which will cause PSD. Then, the Di values of these boards are
calculated.
During Di calculation, boards with redundancy and boards without
redundancy are differentiated.
Calculation of Di for Boards Without Redundancy
The availability of a board without redundancy can be calculated
in the following formulas according to the
previously calculated failure rate and determined recovery rate
(reciprocal of MTTR):
Availability (A) = MTBF/(MTBF + MTTR)
Downtime (Di) = 525,600 x (1-A) (expressed as minutes/year)
Failure rate (1 Failures In Time or FIT) = 10-9
/hour
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 10 of 22
Calculation of Di for Boards with Redundancy
The availability of boards with redundancy can be calculated by
using the Markov model for repairable
products.
Boards with redundancy can be differentiated by work mode:
active/standby backup and load sharing. The
availability of the two types of boards with redundancy can be
calculated in one of the following formulas:
The availability of boards working in N+1 backup mode can be
calculated as follows:
})1()1()]1(1[{1 NaCNaNNaCA NN
The availability of boards working in N+1 load sharing mode can
be calculated as follows:
)}1()1()1()]1(1[{1 1 NaCNaNNaCA NAN
A
where,
A is the availability of boards working in N+1 redundancy
mode.
C is the switchover ratio or the probability of successful
switchover times.
CA is the probability of successful switchover times for the
active board.
CS is the probability of successful switchover times for a
standby board.
CA x CS indicates the switchover ratio of boards working in
active/standby backup mode.
CA indicates the switchover rate of the boards working in load
sharing mode.
a is the availability of each board working in N+1 redundancy
mode, which can be calculated using
MTBF/(MTBF+MTTR).
Di can be calculated based on the availability.
In actual applications, the calculation of Di is complicated
with many conditions considered. For example,
the backup mode can be warm backup or hot backup, and the time
for an active/standby switchover cannot
be calculated using a simple formula. The lifetime distribution
of different components can neither be
calculated using a simple formula.
To accurately calculate the system availability, Huawei adopts
professional reliability simulation software.
The software creates system reliability models based on RBDs and
then analyzes the system availability
and reliability based on Monte-Carlo simulation to simulate the
availability indicators close to the actual
projects.
2.3 Other Relevant Parameters
The MTTR mentioned in this document refers to the onsite repair
time and does not include the time
required for personnel transfer or logistics.
In this document, the MTTR of each board and equipment is
determined to be 1, 2, 4, or 24 hours
according to the MIL-HDBK-472, engineering experience, and field
data.
In addition, according to the reliability engineering baseline
of Huawei, the failure detection rate of active
boards is 95%, the failure detection rate of standby boards is
90%, and the switchover success rate is 99%.
3 eCNS600 Configuration and Reliability Prediction
3.1 eCNS600 Reliability Prediction Models
eCNS600 reliability models can be constructed based on boards
configured with or without redundancy.
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 11 of 22
3.1.1 eCNS600 Reliability Model Constructed with Boards
Configured Without Redundancy
PEM EPC FAN
SMU
SMU
SDM
SWU
SWU
SWI OMU USI
ISU OXI
3.1.2 eCNS600 Reliability Model Constructed with Boards
Configured with Redundancy
PEM
PEM
FPC
FPC
FAN
FAN
SMU
SMU
SWU
SWU
SWI
SWI
OMU
OMU
SDM
SDM
USI
USI
ISU
ISU
QXI
QXI
3.2 Typical Configurations of eCNS600 Reliability Models
3.2.1 Typical Configuration of eCNS600 Boards Without
Redundancy
Board/Module Description Quantity
PEM Power Entry Module 1
FPC Flexible Printed Circuit 1
FAN Fan 1
SMU Service Management Unit 2
SDM Service Data Management 1
SWU Switch Unit 2
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 12 of 22
Board/Module Description Quantity
SWI Switch Interface Unit 1
OMU Operation and Maintenance Unit 1
USI Universal Service Interface 1
ISU Integrative Session Unit 1
QXI QUAD 10GE Interface Unit 1
3.2.2 Typical Configuration of eCNS600 Boards with
Redundancy
Board/Module Description Quantity
PEM Power Entry Module 2
FPC Flexible Printed Circuit 2
FAN Fan 2
SMU Service Management Unit 2
SDM Service Data Management 2
SWU Switch Unit 2
SWI Switch Interface Unit 2
OMU Operation and Maintenance Unit 2
USI Universal Service Interface 2
ISU Integrative Session Unit 2
QXI QUAD 10GE Interface Unit 2
3.3 eCNS600 Board Reliability Indicators
Board/Module Failure Rate (FITs) MTBF (Hours) MTBF (Years)
OMU 2219.61 450529.60 51.43
USI 403.19 2480220.24 283.13
ISU 1973.01 506839.80 57.86
QXI 403.19 2480220.24 283.13
SMM 683.53 1462993.58 167.01
SDM 101.1 9891196.83 1129.13
SWU 2357.99 424090 48.41
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 13 of 22
Board/Module Failure Rate (FITs) MTBF (Hours) MTBF (Years)
SWI 518.08 1930203.83 220.34
PEM 487 2053261.6 234.4
FAN 1000 1000000 114.2
FPC 272.7 3667033.4 418.6
3.4 eCNS600 Reliability Prediction
3.4.1 Reliability Prediction for eCNS600 Boards Without
Redundancy
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
15.73 1 99.99928% 3.81
15.73 2 99.99854% 7.62
3.4.2 Reliability Prediction for eCNS600 Boards with
Redundancy
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
40.0 1 99.99972% 1.50
40.0 2 99.99943% 3.00
4 eCNS610 Configuration and Reliability Prediction
4.1 eCNS610 Reliability Prediction Models
Figure 4-1 eCNS610 reliability prediction model with eight hard
disks, fans working in 7+1 backup mode, and power supply modules
working in 1+1 backup mode
Mother
board
RAID
controller
card
CPU I/O card
Backplane
connecting
hard disks
Hard disk 1
Hard disk 8
...
Fan 1
Fan 8
...
Power supply
module 1
Power supply
module 2
8 7
Memory 1
Memory 24
... 24
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 14 of 22
4.2 Typical Configurations of eCNS610 Reliability Models
Board/Module Quantity
Mother board 1
RAID controller card 1
CPU 1
Memory 24
I/O card 1
Backplane connecting hard disks 1
Hard disk 8
Fan 8
Power supply module 2
4.3 eCNS610 Board Reliability Indicators
Board/Module Failure Rate (FITs) MTBF (Hours) MTBF (Years)
Mother board 3721.0 268744.96 30.7
RAID controller card 559.2 1788268.96 204.1
CPU 40.0 25000000 2853.8
Memory 5000.0 200000 22.83
Backplane connecting hard disks 82.9 12062726.2 1377.0
I/O card 82.6 12106537.5 1382.0
Fan 582.0 1718213.06 196.1
Hard disk 3425.0 291970.80 33.33
Power supply module 2000.0 500000 57.1
4.4 eCNS610 Reliability Prediction
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
12.29 1 99.999071% 4.88
12.29 2 99.998143% 9.76
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 15 of 22
5 DBS3900 Configuration and Reliability Prediction
A DBS3900 consists of two independent modules: baseband (BB)
module and radio frequency (RF)
module, which are connected over the common public radio
interface (CPRI) through an optical cable.
5.1 BBU3900/BBU3910
The BBU3900/BBU3910 is an indoor baseband unit, which fits into
2U space of any standard 19-inch
cabinet. The BBU3900/BBU3910 occupies small space and can be
easily installed.
The BBU3900/BBU3910 includes boards such as UMPT, UPEU, and UBFA
and is fed with 48 V DC
power.
5.2 RRU3251/RRU3232/RRU3252/RRU3253
The RRU3251, RRU3232, RRU3252, and RRU3253 are remote radio
units used outdoors, which can be
installed on a pole or a wall near the antenna.
Figure 5-1 RRU
5.3 DBS3900 Reliability Prediction Models
5.3.1 DBS3900 Reliability Model Constructed with Boards
Configured Without Redundancy
Figure 5-2 DBS3900 reliability model 1
UPEUc FANc UMPT LBBP
RRU
RRU
...
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 16 of 22
5.3.2 DBS3900 Reliability Model Constructed with UPEUs and LBBPs
Configured with Redundancy
Figure 5-3 DBS3900 reliability model 2
UPEUc
UPEUc
FANc UMPT
LBBP
LBBP
RRU
RRU
...
5.3.3 DBS3900 Reliability Model Constructed with UMPTs
Configured with Redundancy
Figure 5-4 DBS3900 reliability model 3
UPEUd
UPEUd
FANd
UMPT LBBP
LBBP
RRU
RRUUMPT
...
5.4 Typical Configurations of DBS3900 Reliability Models
5.4.1 Typical Configuration 1 for the DBS3900
Board/Module Description Quantity
UPEUc Power and Environment interface unit 1
FANc BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 1
RRU3251 Remote Radio Unit 1-3
5.4.2 Typical Configuration 2 for the DBS3900
Board/Module Description Quantity
UPEUc Power and Environment interface unit 1
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 17 of 22
Board/Module Description Quantity
FANc BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 1
RRU3232/RRU3252/RRU3256 Remote Radio Unit 1-3
5.4.3 Typical Configuration 3 for the DBS3900
Board/Module Description Quantity
UPEUc Power and Environment interface unit 2
FANc BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3251 Remote Radio Unit 1-3
5.4.4 Typical Configuration 4 for the DBS3900
Board/Module Description Quantity
UPEUc Power and Environment interface unit 2
FANc BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3232/RRU3252/RRU3256 Remote Radio Unit 1-3
5.4.5 Typical Configuration 5 for the DBS3900
Board/Module Description Quantity
UPEUc Power and Environment interface unit 2
FANc BBU Fan Module 1
UMPT Main Processing & Transmission unit 2
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3251 Remote Radio Unit 1-3
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 18 of 22
5.4.6 Typical Configuration 6 for the DBS3900
Board/Module Description Quantity
UPEUc Power and Environment interface unit 2
FANc BBU Fan Module 1
UMPT Main Processing & Transmission unit 2
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3232/RRU3252/RRU3256 Remote Radio Unit 1-3
5.4.7 Typical Configuration 7 for the DBS3900
Board/Module Description Quantity
UPEUd Power and Environment interface unit 1
FANd BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 1
RRU3251 Remote Radio Unit 1-3
5.4.8 Typical Configuration 8 for the DBS3900
Board/Module Description Quantity
UPEUd Power and Environment interface unit 1
FANd BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 1
RRU3232/RRU3252/RRU3256 Remote Radio Unit 1-3
5.4.9 Typical Configuration 9 for the DBS3900
Board/Module Description Quantity
UPEUd Power and Environment interface unit 2
FANd BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 19 of 22
Board/Module Description Quantity
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3251 Remote Radio Unit 1-3
5.4.10 Typical Configuration 10 for the DBS3900
Board/Module Description Quantity
UPEUd Power and Environment interface unit 2
FANd BBU Fan Module 1
UMPT Main Processing & Transmission unit 1
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3232/RRU3252/RRU3256 Remote Radio Unit 1-3
5.4.11 Typical Configuration 11 for the DBS3900
Board/Module Description Quantity
UPEUd Power and Environment interface unit 2
FANd BBU Fan Module 1
UMPT Main Processing & Transmission unit 2
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3251 Remote Radio Unit 1-3
5.4.12 Typical Configuration 12 for the DBS3900
Board/Module Description Quantity
UPEUd Power and Environment interface unit 2
FANd BBU Fan Module 1
UMPT Main Processing & Transmission unit 2
LBBPd2 Baseband Process and Radio Interface unit 2
RRU3232/RRU3252/RRU3256 Remote Radio Unit 1-3
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 20 of 22
5.5 DBS3900 Board Reliability Indicators
Board/Module Failure Rate (FITs) MTBF (Hours) MTBF (Years)
UPEUc 380 2631578.95 300.4
UPEUd 560 1785714.29 203.8
FANc 940 1063829.79 121.4
FANd 950 1052631.58 120.1
UMPT 1550 645161.29 73.6
LBBPd 1120 892857.14 101.9
RRU3253/3259 5710 175131.35 19.9
RRU3251 2460 406504.07 46.4
RRU3232/RRU3252/RRU3256 3100 322580.65 234.4
5.6 DBS3900 Reliability Prediction
5.6.1 Reliability of the DBS3900 Configured in Mode 1
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
28.99 1 99.99961% 2.07
28.99 3 99.99882% 6.21
5.6.2 Reliability of the DBS3900 Configured in Mode 2
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
28.90 1 99.99960% 2.08
28.90 3 99.99881% 6.23
5.6.3 Reliability of the DBS3900 Configured in Mode 3
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
46.20 1 99.99998% 1.30
46.20 3 99.99926% 3.90
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 21 of 22
5.6.4 Reliability of the DBS3900 Configured in Mode 4
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
46.12 1 99.99997% 1.30
46.12 3 99.99926% 3.90
5.6.5 Reliability of the DBS3900 Configured in Mode 5
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
123.54 1 99.999908% 0.48
123.54 3 99.999723% 1.44
5.6.6 Reliability of the DBS3900 Configured in Mode 6
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
122.22 1 99.999907% 0.49
122.22 3 99.999720% 1.47
5.6.7 Reliability of the DBS3900 Configured in Mode 7
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
27.77 1 99.999589% 2.16
27.77 3 99.998767% 6.48
5.6.8 Reliability of the DBS3900 Configured in Mode 8
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
27.54 1 99.999585% 2.18
27.54 3 99.998756% 6.54
5.6.9 Reliability of the DBS3900 Configured in Mode 9
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
46.00 1 99.999752% 1.30
-
Technical White Paper for System Reliability Prediction of eLTE
2.3
INTERNAL
2014-06-20 Huawei confidential. No spreading without permission.
Page 22 of 22
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
46.00 3 99.999255% 3.90
5.6.10 Reliability of the DBS3900 Configured in Mode 10
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
45.86 1 99.999751% 1.31
45.86 3 99.999253% 3.92
5.6.11 Reliability of the DBS3900 Configured in Mode 11
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
121.03 1 99.999906% 0.49
121.03 3 99.999717% 1.48
5.6.12 Reliability of the DBS3900 Configured in Mode 12
MTBF (Years) MTTR (Hours) Availability Interruption Duration
(Minutes/Year)
120.61 1 99.999905% 0.50
120.61 3 99.999716% 1.50