Monthly Archives: July 2014

Cautions of Insufficient OSN 8800 Cross-connect Capacity on the U2000

Summary:
A U2000 earlier than V100R006C02 can properly manage cross-connect capacity subracks of the OSN 8800. After it is upgraded to the U2000 V100R006C02 or a later version, an alarm is generated indicating insufficient number of license items for subracks with a certain type of cross-connect capacity.

Insufficient OSN 8800 Cross-connect

1.2 [Problem Description]
On the U2000 V100R006C02 and later versions, the license control mode for cross-connect capacity subracks of the OSN 8800 is changed from the common deduction mode to stepped deduction mode based on the sales mode of cross-connect capacity subracks. As a result, when either of the trigger conditions is met, an alarm indicating insufficient number of license items for cross-connect capacity subracks is generated.
Sales mode of cross-connect capacity subracks: If a customer sets the cross-connect capacity to 1.28 TB for an OTN subrack, the customer needs to purchase a 360 GB, a 360 GB-720 GB, and a 720 GB-1.28 TB license item respectively.
Common deduction mode: If the cross-connect capacity of a subrack is set to 1.28 TB, the U2000 deducts only a 720 GB-1.28 TB license item without deducting any 360 GB and 360 GB-720 GB license items.
Stepped deduction mode: If the cross-connect capacity of a subrack is set to 1.28 TB, the U2000 deducts a 360 GB, a 360 GB-720 GB, and a 720 GB-1.28 TB license item respectively.
Trigger conditions:
A U2000 earlier than V100R006C02 uses the common license deduction mode. If either of the following scenarios occurs, an alarm will be generated indicating insufficient number of license items:
Scenario 1: A customer only purchases a 720 GB-1.28 TB license item and sets the cross-connect capacity of a subrack to 1.28 TB. The U2000 earlier than V100R006C02 can properly manage the subrack NE. After being upgraded to the U2000 V100R006C02 or a later version, the U2000 uses the stepped license deduction mode. As a result, the U2000 cannot manage the subrack NE due to the lack of a 360 GB and a 360 GB-720 GB license item.
Scenario 2: A customer purchases a 360 GB, a 360 GB-720 GB, and a 720 GB-1.28 TB license item respectively. The U2000 earlier than V100R006C02 uses the common license deduction mode and allows the customer to set three types of cross-connect capacity, 360 GB, 720 GB, and 1.28 TB, for an OTN subrack. After being upgraded to the U2000 V100R006C02 or a later version, the U2000 uses the stepped license deduction mode and allows the customer to set only 1.28 TB cross-connect capacity for the OTN subrack. As a result, the U2000 cannot manage the subracks with 320 GB and 720 GB cross-connect capacity due to the lack of the 360 GB and 360 GB-720 GB license items.

Symptom:
The following two figures show the changes of license item consumption after a U2000 is upgraded to the U2000 V100R006C02 or a later version.

Identification method:

  •  Method 1:

Based on the actual number of consumed license items on the U2000 earlier than V100R006C02, manually calculate that after the upgrade and compare it with the number of purchased license items.
The calculation formula is as follows:
Number of consumed 360 GB license items (after the upgrade) = Actual number of consumed 360 GB license items (before the upgrade) + Actual number of consumed 720 GB license items (before the upgrade) + Actual number of consumed 1.28 TB license items (before the upgrade) + Actual number of consumed 2.56 TB license items (before the upgrade)
Use Figure 1-1 and Figure 1-2 as an example.
Number of consumed 360 GB license items (after the upgrade: 38) = Actual number of 360 GB license items (before the upgrade: 1) + Actual number of 720 GB license items (before the upgrade: 5) + Actual number of 1.28 TB license items (before the upgrade: 4) + Actual number of 2.56 TB license items (before the upgrade: 28)
In the same way, use the following formulas to calculate the number of consumed 360 GB-720 GB, 720 GB-1.28 TB, and 1.28 TB-2.56 TB license items after the upgrade:Number of consumed 360 GB-720 GB license items (after the upgrade) = Actual number of consumed 720 GB license items (before the upgrade) + Actual number of consumed 1.28 TB license items (before the upgrade) + Actual number of consumed 2.56 TB license items (before the upgrade)

Number of consumed 720 GB-1.28 TB license items (after the upgrade) = Actual number of consumed 1.28 TB license items (before the upgrade) + Actual number of consumed 2.56 TB license items (before the upgrade)
Number of consumed 1.28 TB-2.56 TB license items (after the upgrade) = Actual number of consumed 2.56 TB license items (before the upgrade)

Compare the calculated number of license items with the number of license items purchased by the customer, and determine whether the license quantity meets post-upgrade requirements. If the number of consumed license items for subracks with a certain type of cross-connect capacity exceeds the number of purchased license items, rectification is required.

  • Method 2:

Before the upgrade, use the UEsay (V100R009C00SPC200 or later versions) to check the U2000. If the UEsay prompts that the number of licenses will be insufficient after the upgrade, rectification is required.

  •  Method 3:

After the U2000 is upgraded, if the U2000 license information indicates insufficient number of subracks with a certain type of cross-connect capacity on the OSN 8800, rectification is required.

1.3 [Root Cause]
On the U2000 V100R006C02 and later versions, the license control mode for cross-connect capacity subracks of the OSN 8800 is changed from the common deduction mode to stepped deduction mode based on the sales mode of cross-connect capacity subracks. As a result, when either of the trigger conditions is met, an alarm indicating insufficient number of license items for cross-connect capacity subracks is generated.

1.4 [Impact and Risk]
On the U2000 V100R006C02 and later versions, if the number of license items is insufficient but the customer does not purchase required license items in time:

  •  The U2000 reports alarms.
  • You cannot set cross-connect capacity for the OTN subracks that have insufficient number of license items.

1.5 [Measures and Solutions]
Workarounds
No workaround is provided.
Preventive measures
No preventive measure is provided.
Solutions
If your office encounters the problem and meets the trigger conditions, follow the rectification instructions to obtain required license items to resolve the insufficiency of cross-connect capacity subracks.

1.6 [Rectification Instructions]
The way of obtaining a license file by following the software license rectification process is the same as that of obtaining a license based on the contract. After an onsite engineer submits a batch license rectification application, the engineer obtains a new contract number and PO number. Then the engineer can activate and download the desired license file based on the contract number or PO number in the FlexNet Operations System (FNO).
1. Workflow of batch license replacement

2. Onsite personnel fill in the Fixed Network U2000 Product_XX Site_Software License Replacement Delivering List Template with the following information:
− Customer Name: indicates the full name, not short name of the customer. This name information must be the same as the name contained in the license file that has been used onsite.
− Quantity: indicates the quantity of required license items for U2000 rectification. You can use the license calculation formula to calculate the quantity of required license items.
− ProductVersion: indicates the target U2000 software version. Fill this item with the version of the U2000 used on the live network. The format is “VXXXRXXX”.
− Packing Instructions(Site_Info): indicates the name of the site where the live-network U2000 is located. Fill this item with the actual site name. If multiple U2000s require rectification, ensure their site names are unique.
Quantity calculation method
Obtain the quantity of required license items after the upgrade from the prompt information of the U2000 or manually calculate the quantity. Based on the quantity and the actual quantity of license items purchased by the customer, use the following formula to calculate the quantity of license items for subracks with a certain type of cross-connect capacity to be bestowed to the customer:

Quantity of license items for subracks with a certain type of cross-connect capacity (bestowment) = Quantity of license items consumed (after the upgrade) – Quantity of license items purchased
For cross-connect capacity subracks that have sufficient license items, fill 0 for bestowment quantity.
3. Fill in the rectification e-flow.
− Fill the regional division, department, purpose, and product name based on the actual situation of the site.

TwitterLinkedInGoogle+FacebookPinterestTumblrStumbleUponRedditShare

Cautions on the Flash Memory Failure in DWDM Products

Summary: Because of defective incoming materials, the Spansion Flash memory units provided in a specific lot cannot erase Flash memory data. Huawei transmission boards equipped with these Flash memory units occasionally encounter a software uploading failure, patch upgrade failure, or log recording failure; or the boards sometimes reset unexpectedly or fail to boot from a reset.

 DWDM Products
[Problem Description]

Trigger condition: The problem corresponding to this rectification notice is likely to occur when a board listed in the attachment is undergoing an upgrade/reset or backing up logs because of a Flash memory erasing failure. The probability increases with the frequency of erasing the Flash memory.

Symptom: Upgrade failure: When an involved NE is undergoing a software or patch upgrade, the U2000 displays that loading the software or installing the patch has failed.

1.Backup failure: An SCC board backs up the database every 30 minutes. When an involved SCC board attempts to back up the database after a configuration change occurs in the database, the attempt fails and the board reports a DBMS_ERR alarm. The new configuration is lost after the NE is reset.

2. Unexpected reset or failure to boot from a reset: An involved board is reset upon a component failure. During the restart from the reset, the board performs a Flash read/write check and fails to boot. Identification method: The boards in the attachment are at the risk of having this problem.

[Root Cause] The production process of Spansion is not well controlled. Due to production errors, the grating oxide layer on the Flash chip wafer is uneven. Horizontally, the oxide layer has become thinner because of corrosion. In a high voltage application (programming or Flash erasing), the oxide layer will be punctured and consequently erasing the Flash memory will fail. For a board equipped with a Spansion Flash memory unit, the Flash memory is erased when the board is undergoing an upgrade/reset or is backing up logs. The probability that the board fails to boot increases as the Flash memory is erased more often.

[Impact and Risk] 1.The boards listed in the attachment will occasionally encounter a software uploading failure, patch upgrade failure, or log recording failure; or the boards sometimes reset unexpectedly or fail to boot from a reset. 2.The following table provides the failure rate for each involved board.
[Measures and Solutions] Replace all the boards listed in the attachment.

Preventive measures: Replace all the boards listed in the attachment. The following table provides the board information provided in the part number library.

Please Note That Data Boards May Fail to Start Up After an Upgrade on OptiX OSN 550/500

Summary:
0.03% to 0.1% of certain models of data boards and central switching boards with the basic BIOS version of 1.21 or earlier on OSN 550/500 NEs fail to start up after the NEs are upgraded, which results in the rollback of the NE version. The issue is solved after the boards are replaced.

 OptiX OSN 550/500
Trigger conditions:

This problem may occur if the following conditions are met at the same time:
An NE is upgraded.
The NE uses one of the following boards: TNH1EM6T/F, TNM1EF8F, TNH1CSHDA/B, TNM1PCXLG, TNM1PCXX, TNM1PCXLX, TNM1PCXGB, or TNM1PCXGA.
The basic BIOS version of the board is 1.21 or earlier.
Symptom:
The faulty board stays in the BIOS state after an upgrade and does not start up. The fault is cleared after the NE version is rolled back.
Identification method:
1. Query the version

2. After an upgrade, the STAT indicator on the board is red and the other indicators are off, which indicates that the board stays in the BIOS state and fails to start up. The board starts up after the NE version is rolled back.

The following interfaces are displayed in the case of an automatic rollback after an upgrade failure on the U2000.When both of the preceding conditions are met, the fault can almost be identified.
[Root Cause]
On the basic BIOS of version 1.21 or earlier, the time sequence of the CPU adder is not the optimal configuration. As a result, there is a low possibility that the board fails to start up after an upgrade.
[Impact and Risk]
The board does not start up after an upgrade, which results in service interruptions. The fault is cleared after the NE version is rolled back.
[Measures and Solutions]

The fault is automatically cleared after the NE version is rolled back.

Solution:

The basic BIOS of boards on OSN 500/550 cannot be upgraded online. If the fault occurs on an NE during an upgrade, replace the board with a new board with the basic BIOS version of 1.22 or later.
Material handling after replacement:
Return the replaced board to Huawei, if the board supplied by thunder-link.com, then please contact your sales representative.

How to Configure Protection Group on BWS 1600G or Metro 6100

Summary: If a protection group configured using an above-mentioned board with hardware version E8 or CB experiences both SF state and SD state, the SF or SD state may not end after the related link fault is cleared, which is abnormal.

 Metro 6100

[Problem Description]

Trigger condition:

  •  A board with hardware version E8 or CB is used to configure inter-board wavelength protection, inter-subrack wavelength protection, client 1+1 wavelength protection, or extended intra-board wavelength protection.
  • Both SF and SD events occur in the protection group. For example, the protection group enters the SD state due to bit errors and then enters the SF state due to a fiber cut.

Symptom:
No abnormal alarm is reported for the boards in the protection group, but the protection group is in the SF or SD state on the NMS.
Identification method:

Step 1 Check whether the following conditions occur.
1. An above-mentioned board with a software version earlier than 8.24 is used to configure inter-board wavelength protection, inter-subrack wavelength protection, client 1+1 wavelength protection, or extended intra-board wavelength protection.
2. No abnormal alarm is reported for the boards in the protection group, but the protection group is in the SF or SD state on the NMS.
If the conditions occur, go to the next step.

If the conditions do not occur, no board is involved in this precaution notice and no further action is required.

Step 2 On the Navigator, run ops-get-oppsevtlog:$pgID to query for protection switching logs and check whether SF_DETECTED and SD_DETECTED events match SF_CLEARS and SD_CLEARS events respectively. If they match, no board is involved in this precaution notice. If they do not match, a board is involved. Logs shown in the following figure are used as an example.


According to the logs, totally two SF events and one SF event are reported, but one SF clearance event and two SD clearance events are present. The board wrongly reports one SF clearance event as SD clearance event. Consequently, one SF event is retained and therefore the SF state persists for the protection group.

[Root Cause]
Board software uses the same flag bit to process SF and SD events and wrongly reports an SF/SD clearance event as an SD/SF clearance event. As a result, an SF/SD event persists.
[Impact and Risk]
The protection group is abnormal (in SF or SD state) when no alarm is reported. In addition, protection switching of the protection group fails.

[Measures and Solutions]
Recovery measures:
Reconfigure the protection group. You are advised to apply for an operation time window since services may be interrupted during the reconfiguration.

Preventive measures:
Upgrade board software to 8.24 or a later version.

Note: 1600GThe problem in this precaution notice is resolved since V100R006C02SPC800 for the OptiX BWS 1600G and since V100R008C01SPC630 for the OptiX Metro 6100.

Do you find any solution on rectification Cross-Connect BD_STATUS alarm on OSN 3500

Summary:
Creep occurs on SD585 soldered balls of some cross-connect boards of OptiX OSN 3500 so these boards repeatedly reset, fail to work, and report BD_STATUS alarms.
[Problem Description]
Trigger conditions:
There is a possibility that this problem occurs as a result of long-term exposure of the boards involved to high temperature.

Huawei OSN 3500
Symptom:

1. Boards are reset repeatedly and fail to work.
2. NEs may report BD_STATUS alarms or COMMUN_FAIL alarms on cross-connect boards. System control boards may report BIOS_STATUS alarms on cross-connect boards.
Identification method:
The problem can be identified if the following two conditions are met:
1. The boards are manufactured in Feb 2010, Apr 2010, May 2010, Jul 2010, Aug 2010, or Mar 2011.
2. The cross-connect boards are repeatedly reset and fail to work. The board BOMs are found in the attached Board Delivery Information.

[Root Cause]
The SD585 chip radiator uses the thick spring, which applies high levels of stress to the chip. The soldered ball of the SD585 chip may deform and short-circuit as a result of long-term exposure to high temperature. Therefore, the board repeatedly reset and fail to work.
[Impact and Risk]
1. Services are not influenced because of the 1+1 protection scheme is configured on the cross-connect boards. When a cross-connect board is faulty, services are switched over to the other cross-connect board.

2. In extra situations, both cross-connect boards configured in the 1+1 protection scheme become faulty in a short time. As a result, the NE fails to work and services are interrupted.

[Measures and Solutions]
Recovery measures:
Replace the faulty cross-connect board.
Solution:
Replace the faulty boards.
[Rectification Instructions] 

Replace the faulty boards.

SSN1SXCSA02 (03030KBM) boards are out of production, so replace SSN1SXCSA02 boards with SSN1SXCSA01 (03030DKF) when filling in an electric process application for board rectification in batches. The two kinds of boards can be mixedly inserted or completely replace each other. Active and standby boards of the same type are recommended.

How to solve STG Boards BD_STATUS Alarm on the OptiX OSN 8800

Summary: If the ejector lever on an STG board is not fully inserted during board installation or if a gap is introduced between the front panel and the surface of the board during front panel assembly, the micro switch on the STG board will not be closed and a BD_STATUS alarm will be reported for the STG board.

[Problem Description]
Trigger conditions:
The ejector lever on an STG board is not fully inserted during board installation, or a gap is introduced between the front panel and the surface of the board during front panel assembly.
Symptom:
A BD_STATUS alarm is intermittently or constantly reported for an STG board.
Identification:
You can determine that an STG board has the said problem when all the following conditions are met:
1. A BD_STATUS alarm is reported for the STG board.
2. The STAT indicator on the STG board is steady green.
3. Manufacturer information and logs of the STG board can be retrieved successfully.

[Root Cause]
Function of the micro switch:
The ejector lever and micro switch on an STG board are associated. When the active STG board is removed, the micro switch becomes open. When the board receives the open micro switch signal, the standby STG board takes over to replace the active STG board, ensuring a hitless switchover between the active and standby STG boards.
Root cause for the BD_STATUS

STG Boards

There are following situations in which this problem will occur:

1. The ejector lever is fully inserted, but the micro switch is not closed. In this situation, a stable BD_STATUS alarm is present on the STG board. (Usually this occurs during deployment.)
2. The ejector lever is fully inserted, but the micro switch state is unstable, alternating between closed and open. In this situation, transient BD_STATUS alarms are reported for the STG board.

[Impact and Risk]
The STG board is a clock board and provides a working clock for an NE. For each NE, two STG boards are configured to provide 1+1 backup. Therefore, there are three scenarios:

A. If the active STG board has the said problem, a BD_STATUS alarm is reported for the board and the standby STG board automatically takes over to replace the active STG board. Services on the NE are not affected.
B. If the standby STG board has the said problem, a BD_STATUS alarm is reported for the board. Services on the NE are not affected.
C. If both STG boards on an NE have the said problem, only the standby STG board is set to the offline state. The active STG board is operational and no BD_STATUS alarm is reported for the active STG board. Services are not affected in this situation.
In the preceding situations, there is no impact on services, but a switchover between the active/standby STG boards cannot be triggered. Therefore, in these situations, the defective STG boards must be replaced promptly. For details, see the following Measures and Solutions.

[Measures and Solutions]
Recovery measures:
None.
Workarounds:
Caution: Apply for a maintenance window for taking the following workaround.
Disable the micro switch function for defective STG boards by running commands in the script provided together with this precaution notice. The script file name is Micro Switch Function Disable Script for the OSN 8800 STG Boards.
For details, see the Guide to Disable the Micro Switch Function for the OptiX OSN 8800 STG Boards.

Note:

The micro switch enable commands are non-retainable; instead they will be lost after the STG boards are reseated or undergo a warm/cold reset. In this situation, the commands must be issued again.

The commands vary with product versions. Currently, the commands support only mainstream versions: V100R004C04SPC800, V100R006C01SPC200, V100R006C01SPC300, and V100R006C01SPC500. For any other Huawei product version, independently upgrade the STG board software and logic, or contact Huawei R&D engineers to perform the upgrade.

The workaround may introduce the following risk:
When two STG boards are deployed for 1+1 backup on an NE, services on the NE are occasionally interrupted after the active STG board is removed. Therefore, you are advised to check the active/standby status of the STG boards before the removal operation. If the STG board to be removed is in active state, switch it to the standby state on the U2000 and then remove it. Usually, it is required to remove an STG board in a switching test scenario. The board removal can be simulated by pulling the ejector lever outward to trigger a micro switch open event. Before performing the simulation, the micro switch function must be enabled again. For details on how to enable the micro switch function, see the Guide to Disable the Micro Switch Function for the OptiX OSN 8800 STG Boards.

Preventive measures:
Solution 1: Upgrade the affected NE to OptiX 8800 V100R007C02SPC300 or a later version. For the method of enabling the micro switch function, see the OptiX OSN 8800 STG Board Upgrade Guide.
Solution 2: Independently upgrade the STG board software and logic to a version matching the OptiX 8800 V100R007C02SPC300 product version. For details on how to perform the upgrade, see the OptiX OSN 8800 STG Board Upgrade Guide. After the upgrade, the micro switch function is disabled by default. If the micro switch function is required, run corresponding commands to enable it. For details, see the OptiX OSN 8800 STG Board Upgrade Guide.

 

How to identify High-Risk DWDM network

Summary: In recent years, the long spare part in-transit time has become a major factor leading to the long fault recovery time when major faults occur on WDM networks without protection. To ensure that live network faults are rectified in a timely manner, representative offices are required to assess the service robustness and spare part availability of WDM networks where unprotected long-span links, unprotected links traversing multiple countries,unprotected subnets or sites in remote areas, unprotected high rate services, or unprotected ROADM sites are present. In addition, representative offices are required to identify the high-risk networks that do not satisfy the service robustness and fault recovery SLA requirements, report these high-risk networks to the related regional departments and network product maintenance departments, and determine network enhancement measures with these departments together.

OSN8800-3

1. Criteria for Identifying High-Risk WDM Networks

Criteria for Identifying High-Risk WDM Networks
No. Identification Criteria Remarks
1 No protection is configured for services on the WDM network. If protection is configured for the WDM network or client services, identify whether the protection can effectively prevent faults on single boards from impairing services.
2 Key services cannot be switched to backup links in a timely manner in case of a fault.
3 Some subnets or links are deployed in remote areas. In case of a protection failure or faults on multiple boards, services will be impaired for a long time. Remote areas refer to islands, deserts, or country borders. Identify the areas where spare parts can be delivered only by means of flights, ferries, or international transportation. The arrival time of delivery in such modes is generally uncontrollable.
4 According to the service contract, the representative office is responsible for delivering spare parts to sites in case of a major fault.
5 In case of a major fault, the spare part delivery time is counted into the fault recovery time.
6 In case of a major fault, the maximum spare part delivery time does not satisfy the fault recovery SLA requirements.
7 In case of a major fault, the maximum spare part delivery time does not satisfy the spare part service SLA requirements.
8 Even the above-mentioned SLA requirements are satisfied, the impact of service impairment caused by a major fault is uncontrollable. Key industries, such as international business, finance, healthcare, and ticket booking, have high requirements on service robustness.

Representative offices can determine whether a network is high-risky based on the
preceding criteria and other similar information that has not been listed. If a network is
identified as high-risk, communicate the risks with the customer and promote the
customer to configure protection for services, reserve sufficient spare parts, and ensure
the timely availability of the spare parts.

Meanwhile, report the high-risk networks to the WDM maintenance contact persons of
the regional departments and network product maintenance departments to file the
networks, and determine network enhancement measures with these departments
together.

2. List of High-Risk Boards That Are Prone to Generate Single-Point Failures and
Cannot Be Easily Resolved

  • Top 1: Optical amplifier (OA) boards in unprotected networks

− Long-span dedicated boards such as HBA, ROP, or CRPC/RAU on unprotected links

− Two levels of OA boards such as OAU, OBU, or OPU at the receive end of the main optical path.

 − High-power OA boards such as OAU105 or OBU205 on the main optical path

For the specific critical boards in scenarios without protection, see the attachment List of Risky Boards in Scenarios Without Protection.

[Impact and Risk]
If a critical board in an unprotected WDM network is faulty and the spare part cannot be provided in a timely manner, the fault may not be able to be resolved within the time specified in the fault recovery SLA signed with the customer, causing unmeasurable negative effects.

[Measures and Solutions]
Recovery measures:
Replace faulty parts in a timely manner.

Preventive measures:

  •  Identify high-risk networks according to the above-mentioned criteria, and report the high-risk networks to the WDM maintenance contact persons of the regional departments and maintenance departments to file the networks.
  •  Promote the customers to configure protection for services, reserve sufficient spare parts, and ensure the timely availability of the spare parts.

 

Huawei Announces 2014 H1 Sales Revenue

—Sales revenue reaches USD21.86 billion, a 19% increase year-on-year

huawei revenue Huawei today announced its operating performance for the first half of 2014. In the first six months of this year, Huawei generated sales revenue of USD21.86 billion, an increase of 19% year-on-year with an operating margin of 18.3%.

“Revenue and profit for the first half of 2014 are in line with our expectations. Driven by increasing investments in LTE networks worldwide, Huawei has further solidified its leadership position in mobile broadband. Rapid growth in software and services helped maintain steady growth in our carrier network business. Our efforts in the enterprise business have begun to pay off. We have enjoyed accelerated growth in this area. We also achieved quality and sustainable growth in our consumer business thanks to the increase of brand awareness and smart devices sales worldwide. Our flagship smartphone, the Ascend P7, is being sold in more than 70 countries and regions.” said Cathy Meng, Huawei’s Chief Financial Officer. “We are confident that in 2014 we will achieve sustainable growth, robust operations, and healthy financials.”

Looking ahead, Huawei will remain committed to open innovation and cooperation, creating greater value for our customers, advancing the ICT industry, and ultimately contributing to the building of a better connected world.

Although, according to revenue report, Huawei already overtake Ericsson, but, particularly in carrier business, Ericsson still ahead of Huawei, which business unit donate 70% of Huawei’s revenue. It is still hard to say who wins the telecom supplier’s world cup.

Be careful SS49EFS Interrupt your Metro1000 service

[Problem Description]
Trigger conditions:

An NE is powered cycled, and the SS49EFS board is cold reset.
Fault symptoms:
An OptiX Metro 1000 NE is interconnected with a third-party SDH device using SS49EFS boards. The alarms such as LP_TIM_VC12 and LP_SLM_VC12 are reported on the SS49EFS boards on the OptiX Metro 1000 NE. Consequently, the services on the SS49EFS boards interconnected to the third-party SDH device are interrupted.
Identification method:
All of the following conditions must be met at the same time:
SS49EFS boards earlier than V3.05 are configured on the OptiX Metro 1000.

Metro1000

[Root Cause]

After the SS49EFS boards are cold reset, the NE delivers overheard configurations to the boards. If no VCTRUNK is bound on the channels, the NE will read the default J2 and V5 bytes from configuration files and deliver them to the SS49EFS boards. The NE first delivers VC-12 overheads and then VC-3 overheads. The first three channels on an SS49EFS board are shared by VC-12s and VC-3s. The VC-12 overhead configurations will be overwritten by the VC-3 overhead configurations


[Impact and Risks]
After an NE is powered off or the SS49EFS board is cold reset, the services on the equipment interconnected to third-party SDH equipment may be interrupted.
[Measures and Solutions]
Recovery measures:
Check the received bytes on the SS49EFS boards. Reconfigure the to-be-transmitted and to-be-received J2 and V5 bytes on the SS49EFS boards to ensure that the to-be-received J2 and V5 bytes are the same as the received J2 and V5 bytes. (The problem will reoccur after the cold reset or power-off.)
Workarounds:
Do not use the one to three timeslot when configuring VC-12 services on the SS49EFS boards on an OptiX Metro 1000.
Solution
The SS49EFS board V3.06 resolved this problem.
[Rectification Scope and Time Requirements]