Principle of QoS Planning

The QoS planning is an E2E planning, and it can be divided into following policies: traffic classification, marking, and scheduling policy; and traffic monitoring and DBA policy.

Traffic Classification, Marking, and Scheduling Policy

Service Type	802.1p Priority	Queue Scheduling Method	OLT Queue ID (8 Queues)	ONT Queue.ID
				When Eight Queues Are Supported	When Four Queues Are Supported
Management service	6	PQ	6	6	3
VoIP service	5	PQ	5	5	2
IPTV service	4	PQ	4	4	2
Wi-Fi access service	1	PQ	1	1	1
Internet access service	0	PQ	0	0	0

Note: The service priorities in this table are the recommended values. The service priorities are arranged according to the operators's actual plan.

Traffic Monitoring and DBA Policy

Items	Management service	Internet access service	Wi-Fi access service	VoIP service	IPTV service
GEM port	11 [Remark 1]	14	15	12	13
TCONT	Select a DBA type based on service types. For example, select Type1 for the management service, Type4 for the Internet access service, and Type2 for the VoIP and IPTV services.
DBA type	Type 3 (DBA profile: assured bandwidth + maximum bandwidth. Users are allowed to preempt the bandwidth on condition that the users' assured bandwidth is guaranteed. However, the total bandwidth cannot exceed the maximum bandwidth.)
DBA bandwidth planning	Configure the DBA bandwidth according to the user's bandwidth package. The assured bandwidth is the maximum bandwidth required by management packets, VoIP, and IPTV upstream packets. The maximum bandwidth is larger than or equal to the maximum bandwidth that users apply.
Rate limit on OLT downstream	No rate limit	Configure rate limit by a traffic profile as required. [Remark 2]	Configure rate limit by a traffic profile as required. [Remark 2]	No rate limit	No rate limit
Rate limit on ONU upstream port	No rate restriction
Rate limit on ONU downstream port	No rate restriction

Remark 1: GEM port value depends on the planning of the service provider, but in principle, use different GEM ports for different services.

Remark 2: Table 1 shows the reference service bandwidth of each service for each user.

Table 1 Reference service bandwidth of each service for each user
Service Type	Upstream Bandwidth	Downstream Bandwidth	Bandwidth Description
Internet access service/Wi—Fi access service	Determined based on the service tariff package.	Determined based on the service tariff package.	Available bandwidth of Internet access service or Wi—Fi access service = Committed bandwidth of the service tariff package - VoIP bandwidth - IPTV bandwidth
VoIP service	200 kbit/s	200 kbit/s	The upstream bandwidth and the downstream bandwidth of VoIP service are symmetrical. The actual bandwidth is related to the coding format. This bandwidth is calculated for two POTS ports.
IPTV service (standard definition program)	/	2.5 Mbit/s per channel	IPTV service mainly occupies the downstream bandwidth. The actual bandwidth depends on the coding format, the picture in picture information, 10% bandwidth burst traffic, and the number of programs that can be concurrently watched by one user (in the case of multiple STBs). The upstream bandwidth is mainly used for transmitting IGMP packets, which requires little bandwidth. Therefore, the bandwidth occupied by IGMP packets can be neglected.
IPTV service (high definition program)	/	9.7 Mbit/s per channel

Note:

· The rate restriction on the BRAS or SR is recommended. OLTs and ONTs do not restrict the rate for service streams.

· If BRAS does not support rate restriction, OLTs can restrict the rate for service streams through the traffic profile.

· Different service packets on the ONT are distinguished by different VLAN IDs. The service packets are mapped to GEM ports based on VLAN IDs so that different service packets are transmitted to different GEM ports. Each GEM port (each service) corresponds to a T-CONT or all GEM ports share a T-CONT.

· The sum of the assured bandwidth of all ONTs connected to a PON port and the fixed bandwidth of OMCI management channel is less than the GPON upstream bandwidth. Some bandwidth must be reserved for the future service expansion.

More related:

Be Aware of SSN4SL64 Board ID on MSTP Products

Why all the boards on OSN 3500 become grey?

3G service not pass in STM-1 due to SNCP problem

Burst Optical/Electrical Technology

In 10G GPON upstream direction, Time Division Multiple Access (TDMA) is used. An optical network unit (ONU) transmits data only within the allocated timeslots. In the timeslots that are not allocated to it, the ONU disables the transmission of its optical transceiver to prevent other ONUs from being affected. The optical line terminal (OLT) then receives the upstream data from each ONU in a burst manner based on timeslots. Therefore, to ensure normal running of the 10G GPON system.

• Figure 1 shows the burst transmit function supported by ONU-side optical modules.
• Figure 2 shows the burst receive function supported by OLT-side optical modules.

Figure 1 Burst transmit function supported by ONU-side optical modules 

Ranging can be implemented to prevent cells transmitted by different ONUs from conflicting with each other on the OLT. However, the ranging accuracy is ± 1 bit and the cells transmitted by different ONUs have a protection time of several bits (not a multiple of 1 bit). If the ONU-side optical modules do not support the burst transmit function, the transmitted signals overlap and distortion occurs.

Figure 2 Burst receive function supported by OLT-side optical modules 

• The distance from each ONU to the OLT varies and therefore the optical signal attenuation varies for each ONU. As a result, the power and level of packets received by an OLT at different timeslots various.
• If the OLT-side optical modules do not support the burst receive function, the OLT may restore incorrect signals because only the level greater than the threshold is considered valid and the signals with the level lower than the threshold cannot be restored.

NOTE:

In the 10G GPON system, all data is broadcast downstream to ONUs. The transmission requires OLT-side optical modules to transmit optical signals continuously and ONU-side optical modules to receive optical signals continuously. Therefore, these optical modules are not required to support the burst receive and transmit function.

More related:

How to Troubleshoot Synchronous Ethernet Clocks For Huawei SDH

Be aware of SSN3PSXCSA replace Cross-connect Board on OptiX OSN 3500

3G service not pass in STM-1 due to SNCP problem

The Service Is Interrupted After the Protection Is Triggered

The service in the working channel is interrupted after the inter-board wavelength protection is triggered, because the 1+1 inter-board protection is mistakenly configured on the OptiX BWS 1600G.

Product

OptiX BWS 1600G

Fault Type

Service Interruption

Protection

Symptom

There are alarms on the SDH equipment interconnected to the DWDM network, but the client-side services are configured with the 1+1 inter-board protection and are not interrupted. After checking the entire network, it is found that there is no output power on the client side of the LWC1 board in slot 5 of the OptiX BWS 1600G, and the laser on the client side is shut down. In addition, reseating the board cannot solve this problem. After the LWC1 board in slot 5 is moved to another subrack of the local station, the laser on the client side of the LWC1 resumed normal operation.

Cause Analysis

After checking the entire network and analyzing the symptom, the cause for this problem is that the inter-board wavelength protection is mistakenly configured on the LWC1 boards in slot 5 and slot 6. On the existing network, the LWC1 boards in slot 5 and slot 6, however, are considered as two independent boards to carry separate services. When the customer adjusts the optical path, the inter-board wavelength protection is triggered. In this case, the client-side laser on the LWC1 board in slot 5 (the working board) is shut down. As configured, the protection switching mode is configured to the non-revertive mode. Therefore, although the fault that occurs on the LWC1 board in slot 5 is eliminated, the service in the channel cannot be restored, and thus the service is interrupted.

Procedure

1. After checking the entire network on site, the inter-board wavelength protection is deleted. This eliminates potential faults.

Reference Information

None.

More related:

when Ring MSP Configuration Failures on SSN1SLQ16 Boards on OSN 3500 Products

How to do when Occasional Hot Patch Installation Failures?

Why Broadcast Packets Cannot Be Transmitted Successfully on Several HMSTP Data Boards?