Technical White Paper

for Visualized IP
Network Operation &
Management
Contents
Technical White Paper for Visualized IP Network O&M..............1
1 Overview...............................................................................2
2 Three Invisible Elements in IP Network O&M..........................3
3 Huawei Visualized IP Network O&M Solution.........................6
4 Service Quality Monitoring.....................................................9
5 Network Quality Monitoring..................................................13
6 Deployment Modes...............................................................16
6.1 Deployment of Probes on the U2520 and Basic Networking Modes............16
6.1.1 Deployment of Probes on the U2520.......................................................16
6.1.2 Basic Networking Modes of Probes on the U2520....................................17
6.1.3 Probe Topology and Distributed and Hierarchical Management on the
U2520..............................................................................................................19
7 Indicator-based Evaluation System.........................................20
7.1 SLA Evaluation............................................................................................20
7.2 Hierarchical Measurement...........................................................................22
7.2.1 KPI...........................................................................................................22
7.2.2 KQI..........................................................................................................23
8 Reference Standards..............................................................26
 8.1 MOS...........................................................................................................26
8.2 RFC4445-MDI.............................................................................................27
8.3 VMOS.........................................................................................................27
9 Typical Applications...............................................................29
9.1 Flow-by-flow MDI and VMOS Video Monitoring Scheme.............................29
10 Interworking........................................................................31
10.1 EANTC Test (Video Monitoring Test)..........................................................31
10.2 Compliant Standards and Protocols...........................................................32
11 Acronyms and Abbreviations...............................................35

Figures
Figure 3-1 Huawei full-lifecycle IP network O&M solution..................................6
Figure 3-2 Network location of the service quality monitoring solution..............7
Figure 4-1 Service architecture of the U2520.....................................................9
Figure 4-2 SAP management.............................................................................11
Figure 5-1 Network architecture supported by the U2520.................................13
Figure 7-1 Key points of SLA evaluation.............................................................20
Figure 7-2 Hierarchical indicator system.............................................................22
Figure 7-3 Three-dimensional KQI aggregation..................................................24
Figure 8-1 User satisfaction corresponding to different MOS scores...................26
Figure 9-1 Typical application of the flow-by-flow IPTV service monitoring scheme.......29
Figure 9-2 Flow-by-flow MDI and VMOS video monitoring scheme...................30
Figure 10-1 Physical network topology used in EANTC tests...............................31
Figure 10-2 Topology view associated with a video monitoring test...................32

Tables
Table 4-1 Service types, test types, and test indicators supported by the U2520
V200R001........................................................................................................12
Table 5-1 Network types, test types, and test indicators supported by the U2520
V200R001........................................................................................................15
Table 10-2 Ethernet service standards and protocols..........................................34
Table 10-1 Network standards and protocols.....................................................34
Technical White Paper for
Visualized IP Network O&M

Abstract:
All IP is definitely the major technology for future network and service
development. In addition, end users are always pursuing for better
service experience. Nevertheless, the service quality is invisible to
end users on an all-IP network. Carriers will thus have to face two
troubles. One is how to shorten the time of configuring services and
locating faults on a large-scale IP network to achieve highly efficient
operation and maintenance (O&M). The other is how to satisfy users'
quality of experience (QoE) requirements by monitoring and managing
the network quality. Huawei provides the visualized IP network O&M
solution, through which the network O&M department and the service
department can obtain the same QoE indicators. This document
describes the technologies related to visualized IP network O&M
and the typical applications of these technologies in the visualized IP
network O&M solution from Huawei.

Keywords:
U2520, VoIP, IPTV, HSI, O&M

1
 1 Overview
At present, voice over IP (VoIP), Internet protocol television (IPTV),
high speed Internet (HSI), mobile, and virtual private network (VPN)
services, games, and value-added services are running on IP and multi-
protocol label switching (MPLS) networks. The following challenges
are imposed on carriers:

•• How to manage QoE indicators of end users

•• How to monitor the service quality in real time
•• How to quickly and accurately locate faults
•• How to perform capability evaluation before deploying a new
service
•• How to evaluate the comprehensive service quality around the
clock

Though the integrated network management system (NMS) on

an IP/MPLS network where various services are carried is available
for service configuration, management, and maintenance, how to
determine service quality degradation, how to evaluate the service
bearing capability of the network, how to monitor the network status,
especially how to quickly and accurately locate faults through tests
and how to identify whether a fault is a service platform fault or a
network fault, are now the new challenges for IP network O&M.

2
2 Three Invisible Elements in IP
Network O&M
Three invisible elements exist in IP network O&M, which reduces the
O&M efficiency.

•• The quality of services carried on IP networks is invisible.

Carriers cannot learn end users' experience on the services.
The traditional NMS provides the function of viewing network
performance, but the quality of carried services is invisible.
Network performance is separated from service quality.
The service department and the network department have
different understandings on faults because there is no uniform
measurement. Therefore, experts of different departments need
to work together to locate faults. This poses high requirements
for skills but results in low fault location efficiency.

•• Service trails cannot be viewed because routes are invisible.

Dynamic routes are imported to IP networks, and service trails
on Layer 3 networks are invisible. As routes are invisible, faults
reported by end users usually disappear when O&M engineers
perform fault location, and fault causes cannot be located
because the faults cannot be reproduced and no associated
historical information is available. All of this results in the difficulty
of eliminating the potential troubles, and skilled datacom experts
need to participate in fault location. In addition, O&M engineers
cannot prevent network-wide faults caused by route flapping.
Route flapping has a catastrophic impact on networks, and even
causes the networks to break down.

•• End-to-end (E2E) channels are invisible.

The process of creating E2E channels is complex and the status
of the channels is invisible. Cross-domain deployment is required
for the creation of E2E channels and the configurations are rather
complex. Nevertheless, traditional single-domain NMSs cannot
achieve visualized and efficient service deployment and the status
of IP channels is invisible after service deployment. O&M engineers

3
 need to consider the relationships between service deployment
parameters on every node because incorrect parameter settings are
hard to discover and rectify. Therefore, high skills of O&M engineers
are required.

The preceding troubles really afflict O&M departments of carriers. Is

the reliability of IP networks really low?

Seeking a Solution is fairly urgent.

Actually, the reliability of IP networks can be guaranteed. Reactive IP

network O&M must be changed to proactive IP network O&M.

In reactive O&M, O&M engineers locate and rectify faults only after
receiving complaints from end users. This cannot satisfy IP network
O&M requirements. Carriers need to forecast weak parts of networks
according to daily network running information and take proper
measures. Therefore, the change of the O&M mode is much important
for the evolution towards all IP. Then, what proactive O&M is suitable
for IP networks?

Proactive O&M requires the monitoring on service experience of end

users. Therefore, O&M personnel need to periodically collect network
performance and service quality data, analyze the data, extract the
trend information, and pre-locate possible fault points and weak
parts. In this manner, carriers can learn the service experience of end
users, check whether the service quality will be degraded, and solve
problems caused by the degradation in advance. In this manner,
complaints from end users will be reduced, customer loyalty will be
improved, and O&M costs will be decreased.

The distribution of carriers' investment drive on the O&M system

proves the necessity of proactive O&M. The result of a survey
conducted by Gartner shows that the first O&M investment drive
of carriers is proactive preventing network performance faults. This
drive accounts for 27% of the total. The second and third are quickly
troubleshooting network fault s and meeting application performance
service level agreement (SLA), which account for 15% and 12%
respectively.

IP network O&M imposes the following requirements:

4
•• Quick troubleshooting
Network and service faults should be rectified quickly to implement
fast fault identification and location.

•• Proactive fault prevention

Service quality and network performance should be monitored in
real time so that faults can be discovered in time; faults of the IP
bearer network should be associated with affected services, and
alarms should be generated based on trend analysis.

•• Routine network management

The network quality monitoring and health evaluation system
should be established. The KQI/KPI indicator system reflects and
manages user experience, and helps to learn the actual network
running status, thereby continuously improving end users' loyalty.

Measurability is the basis of manageability, and manageability is the

basis of improvability. After all IP is implemented, measurability must
be ensured first, that is, IP network O&M must be visible.

5
 3 Huawei Visualized IP Network
O&M Solution
Figure 3-1 shows Huawei full-lifecycle IP network O&M solution. The
U2520 is responsible for visualized O&M of IP networks and carried
services.

Services Mgmt. ●N2510- ●U2520-

(Customer Fiber/Copper IPTV/VOIP/HIS/VPN
QoE Assurance) Line Assurance Services Assurance

●U2560- ●U2000-IP/Transport/
Network Mgmt. Home Network Microwave/PON/
Unified Mgmt. DSLAM unified Mgmt

PE
●MDS 6600- FTP Youtube

Network Backbone IPTV

Web
VolP

Planning and Metro

PE
qame
Designing

Home Access line Network (access/metro/backbone) Service Centers

Figure 3-1 Huawei full-lifecycle IP network O&M solution

Though IP networks can be managed in many ways, network reliability

can be guaranteed and services can be better carried only if network
management is fully associated with services, QoE indicators are
concerned, and visualized IP network O&M is implemented. Huawei sets
an unprecedented example of visualized IP network O&M in the industry
by implementing the visualization of IP service quality, trail, and deployment
and completely solving the "black box" problem of IP network O&M.

The U2520 is the IP network evaluation system launched by Huawei.

As a key device for visualized IP network O&M, it consists of the service
monitoring system and the network evaluation unit, and supports the
operation of full broadband services.

The network monitoring system provides the functions of network

quality monitoring, service quality monitoring, on-demand test,
threshold crossed alert (TCA) management, probe management,
system management, and report management.

The network evaluation unit is a case-shaped external probe that

can be deployed at the network access layer and convergence layer
to collect network and service performance data and report it to the
monitoring system.
6
 Visualize

Service Standus & Inregraclon

SL A Roaporting OSS Ord Party Analyze

Center
Center
Measure
HSI
VoIP
IPTV Access Metro Edge Internet
GSM/Base Station
Softx
BTS BSC MGW
Access Backhaul Backbone IPTV

Figure 3-2 Network location of the service quality monitoring solution

The U2520 service quality monitoring solution is based on the

distributed hierarchical architecture that consists of the probe
measurement layer, server data analysis and collection layer, and
user Webpage display layer. As shown in Figure 3-2, with the probe
measurement layer, the service quality monitoring solution can fully
function in the VoIP, HSI, IPTV, and Multi-Play scenarios. The server
data analysis and collection layer of the U2520 center is used to
manage the NEU100 probe network and collect the KPIs gathered
by probes. At the access layer, the U2520 works with the N2510 to
achieve E2E service quality monitoring. At the user Webpage display
layer, the U2520 implements deep KPI data analysis and provides the
customized report system.

The U2520 is acclaimed as the dashboard of an IP network. It has the

following technical advantages:

•• Visualized IP forwarding plane

The U2520 detects the quality of IPTV, VoIP, HSI, an VPN services
on IP networks and reflects end users' service experience. By
comparing the quality of services segment by segment, the U2520
implements service-based fault identification and responsibility
clarification, which helps O&M personnel to quickly rectify faults
on the IP forwarding plane. The IP network O&M department and
the service operation department have the same QoE indicators
and thus fault identification is easy to achieve.

•• Visualized IP control plane

7
 Trails are available for IP services and the U2520 automatically
calculates and displays service trails and dynamically monitors and
analyzes route changes to prevent network-wide faults caused by route
flapping, which helps O&M personnel to quickly analyze, forecast, and
locate faults on the IP control plane.

Visualized IP forwarding plane and visualized IP control plane can be

combined. The forwarding plane proactively simulates the service packet
test by using service trails. If Layer 3 service trails change, the test
simulated by the forwarding plane adjusts dynamically to implement
automatic monitoring management on service bearing. If the service
bearing quality is degraded, the intelligent diagnosis assistant is
automatically started to locate the degradation cause in a while and
provide troubleshooting suggestions, which greatly improves IP network
O&M efficiency and end users' satisfaction.

In addition to completely monitoring the IP forwarding plane and IP

control plane, the U2520 provides a powerful indicator system, a network
health evaluation system, and O&M evaluation rules to improve the
IP network O&M capability. According to the analysis statistics of the
indicator system, O&M personnel not only can rectify faults, but also can
know fault causes and improve accordingly to prevent certain faults that
result from human factors, thus improving the IP network O&M efficiency.

The U2520 can also implement network-wide service monitoring and

provide the service quality monitoring solution for the networking scenario
where multi-vendor devices are used.

The U2520 works with Huawei U2000 that provides the following
functions:

•• Managing the resource pool for E2E service parameters

•• Automatically creating E2E services after users click source and sink nodes
•• Dynamically allocating service parameters
•• Automatically managing logical relationships between E2E service
parameters on NEs
•• Automatically checking configurations
•• Automatically checking connectivity
•• Deploying services at a time
•• Supporting one-stop batch NE configuration
•• Deploying VPN data for E2E services
•• Supporting visualized E2E operations
•• Supporting visualized display of channel quality
8
4 Service Quality Monitoring
IP services have the characteristics of wide coverage, complicated
service features, and large service traffic. This causes the difficulty
in monitoring the quality of IP services. In addition, carriers face
challenges regarding the following problems:

•• Evaluating services before service deployment

•• Collecting data to generate alarms when the service quality degrades
•• Identifying whether a service fault is caused by service problems or
network problems, and locating the faulty point of the service

IP services that can be monitored by the U2520 are classified into

HSI, IPTV, VoIP, and L3VPN services. The services can be monitored
in proactive and reactive modes. In proactive mode, user behaviors
are simulated through probes to evaluate the service quality and
monitor services in real time. In reactive mode, service flows on the
current network are received, and the quality of each service flow and
session is evaluated and monitored. In the former mode, users can
flexibly deploy services and easily locate faults. In the latter mode, all
service flows and sessions are monitored and the monitoring result
is accurate; however, this mode requires traffic mirroring and high
probe costs and causes the difficulty in deployment.

CCTV1
Service CCTV2
element
BTV1
SAP Service BTV2
element
Aggregation Aggregation Defined as: …
SAP
Service Service
SAP
SAP element www.sina.com
SAP
SAP Service www.sohu.com
element
www.google.com
The SAPs Service …
Aggregation
belong to element
different
SAP groups.

SAP group 1 SAP group 2 SAP group 3 SAP group 4

Star Full- mesh Star Star

When SAPs are organized in SAP groups, the required SAP topology can be established
for all SAPs of services according to the topology types of SAP groups.

Figure 4-1 Service architecture of the U2520

9
 A service on the U2520 consists of three parts: service element, service
access point (SAP), and SAP group.

•• Service element: As a sub-service or a part of a service, a service

element is used to evaluate the service quality. For example, from
a user's point of view, community antenna television (CATV)
providers and network providers are service providers (SPs), and
specific services provided by channels such as http://www.sina.com
and http://www.sohu.com are service elements. Service elements
work on the networks provided by SPs, and the service quality can
be evaluated.

•• SAP: A SAP is a conceptual point where a service is delivered to

users. SAPs are used to differentiate between users and SPs. SPs
push SLA-compliant services to all the SAPs. Therefore, each service
has at least one SAP, each SAP can belong to only one service, and
each user has to access a service through SAP(s).

•• SAP group: SAPs can be organized into different SAP groups

that use a star or full-mesh topology. In this manner, any type of
topology can be created for the SAPs that support various service
topology tests.

Managing a service is managing the SAPs of the service. SAPs must

be attached to physical networks. In service monitoring of the U2520,
all test instances are generated based on SAPs. A SAP needs to be
associated with an interface on an NE to indicate the location of
the SAP and with an interface on a probe to indicate the device that
initiates tests.

10
 SAP Network Network
SAP
edge edge

Access Access
devices devices

DSLAM DSLAM
Switch Carrier network Switch
Low -end (Multi -service network) Low -end
router router
…
…

Access network or Access network or

leased line leased line

Managed by carriers
Figure 4-2 SAP management

An IP service can have a single SAP or multiple SAPs.

•• Only one SAP is required to access a single-SAP service, such as a

broadband TV (BTV) service, a video on demand (VoD) service, or
an HSI service. These services that are provided by network servers
are accessed through only a single SAP.

•• Two or more SAPs are required to access a multi-SAP service, such

as a VoIP service or a point to point (P2P) service. These services
are accessed based on the interaction of SAPs.

The U2520 supports the ability to test various key indicators of end
users' services, such as the IPTV, VoIP, and HSI services. This facilitates
the quality evaluation of end users' services. Table 4-1 lists the
service types, test types, and test indicators supported by the U2520
V200R001.

11
 Service
Test Type Test Indicator
Type
BTV [channel switching time/MOS/packet
loss ratio]
Multicast Multicast receiving [jitter/DF/MLR]
BTV receive test RTSP [flow control time/connection time/
service RTSP test packet loss ratio/jitter/TCP retransmission]
VoD
RTSP [DF/MLR/MOS/TCPRT15/TCPRT24/
service
MLT24/MLT15]

Video RTP Video RTP [MLR/packet loss ratio/MOS/jitter/

Test DF]

FTP test FTP [download/upload rate]

DNS test DNS [parsing]

HSI
service HTTP [arrival time of the first packet/page
HTTP test
download duration]

PPPoE test PPPoE

DHCP test DHCP [IP address acquisition time]

VoIP
VoIP RTP VoIP RTP [packet loss factor/R value/MOS/
service
test delay/jitter factor/jitter/packet loss ratio]

UDP jitter
L3VPN L3VPN [UDP jitter/packet loss ratio/delay]
test
service
TCP test L3VPN [TCP delay]

Table 4-1 Service types, test types, and test indicators supported by the U2520 V200R001

12
5 Network Quality Monitoring
The U2520 can be used to comprehensively monitor the quality of
IP/MPLS networks in real time. Network quality analyse (NQA) is a
key to network monitoring. The NQA-enabled internal probe from
Huawei helps to detect and diagnose the network quality in real time,
and measure network performance through performance indicators
such as jitter, delay, and packet loss ratio. Users can use the NQA-
enabled internal probe to perform tests based on test instances.
Internal probes can be flexibly configured. For example, users can
configure internal probes through command lines, or remotely
configure internal probes and collect data by using the U2520. Each
test instance provides a series of test policy options through which
users can set parameters such as the length of sent packets, interval
for sending packets, and protocol type. Based on settings of the
parameters, the NQA-enabled internal probe can flexibly perform
tests according to the current network situation.

IP trail LSP

IP
connection IP site MPLS site

IP network MPLS network

Figure 5-1 Network architecture supported by the U2520

13
 Networks that can be monitored by the U2520 are classified into
physical networks and logical networks. A physical network consists of
NEs and physical links between NEs. It manages elements associated
with physical NEs, such as interfaces, boards, and frames. The U2520
evaluates the quality of a physical network according to indicators
such as the CPU usage, memory usage, interface packet loss ratio, and
interface bandwidth usage of physical NEs.

A logical network consists of network sites and logical connections. A

physical network can be abstracted as many different network sites.
As shown in Figure 5-1, the route connecting the IP network and the
MPLS network is abstracted as two logical sites: an IP site and an MPLS
site. The two sites belong to the IP network and the MPLS network
respectively. Different network sites have different network interfaces.
For example, the MPLS site has only MPLS interfaces. The U2520
evaluates the quality of a logical network according to indicators, such
as delay, jitter, packet loss ratio, connectivity, and trace, associated
with logical links and logical trails. In addition, the U2520 supports the
abilities to locate and identify faults on logical networks.

The features of network monitoring through the U2520 are as follows:

•• The bearer network provides service-sensitive functions of

diagnosing network faults and monitoring the network quality,
which ensures that the bearer network can meet users' SLA
requirements.

•• The U2520 provides abundant fault diagnosis tools. With these

tools, users can quickly locate faults according to the relation
between networks and service layers, thus greatly shortening the
fault location time.

•• The data of typical services on the bearer network can be simulated

according to traffic characteristics. In this manner, users can obtain
information about the bearing of simulated services, thus accurately
understanding the bearing capability of the bearer network.
This helps to dynamically adjust policies according to the current
situation to continuously improve network status.

The U2520 can be used to test various key indicators on IP, MPLS, and
physical networks, which helps to accurately evaluate the network
quality. Table 5-1 lists the network types, test types, and test indicators
supported by the U2520 V200R001.
14
Network
Test Type Test Indicator
Type

UDP jitter and packet loss ratio

UDP jitter test
UDP delay
L3VPN
TCP jitter test TCP delay

ICMP jitter test ICMP delay, packet loss ratio, and jitter

LSP jitter test LSP delay, packet loss ratio, and jitter
MPLS
network LSP TE jitter LSP TE delay, packet loss ratio, and
test jitter

Card statistics Network [load]

Physical
Physical port Interface [packet loss ratio]
network
statistics Port [bandwidth usage]

IP
UDP jitter test UDP jitter and packet loss ratio
network

Table 5-1 Network types, test types, and test indicators supported by the U2520 V200R001

15
 6 Deployment Modes

6.1 Deployment of Probes on the U2520 and Basic

Networking Modes

6.1.1 Deployment of Probes on the U2520

Probes on the U2520 can be deployed at the access layer of end users,
the convergence layer, and the core layer.

Deploying Probes at the Access Layer of End Users

Active probes on the U2520 are equipped with FE interfaces. Probes
deployed through FE interfaces are nearest to end users, thus ensuring
the most accurate service monitoring results. The scale of probes
deployed on the user side is usually large. Therefore, in the case
where a lot of probes are deployed and a lot of services need to be
monitored, the U2520 occupies many bandwidth resources, which
affects actual services. If probes need to be deployed at the access
layer of end users, it is recommended that representative locations (for
example, locations of key customers) be selected to deploy probes, or
an active probe be mounted to a downstream modem of a DSLAM.

DSLAM Fiber FE/GE

To Metro Ethernet

Copper Line
Copper Line

FE
Modem

IPTV
PC

Deploying Probes at the Convergence Layer of a

Network
Probes deployed on nodes at the convergence layer of a network can
be equipped with FE and GE interfaces. When services are simulated
in active mode to monitor the service quality, only common GE or
16
FE interfaces are required. When the service quality is monitored in
reactive mode, a mirrored port needs to be configured.

Fiber FE/GE Fiber FE/GE

Fiber/LAN FE/GE
Switch

Deploying Probes on Core or Edge Routers at the IP

Bearer Layer
Deploying probes on core or edge routers at the IP bearer layer is
similar to deploying probes on nodes at the convergence layer. Users
can connect an active probe to such a router directly through a fiber,
a LAN, an FE interface, or a GE interface. To deploy a passive probe
to monitor the audio and video signals on the network in real time,
users can connect the probe to the router through a mirrored port,
without considering whether the load on the router is increased.

Fiber FE/GE Fiber FE/GE

Fiber/LAN FE/GE
Router
Mirror Port/Common port

17
 6.1.2 Basic Networking Modes of Probes on the U2520

Loopback Through the Cooperation of Two Probes

GRS

CDMA
Source NTP Destination

HSI Service Monitoring in C/S Mode

Server

Client

Monitoring of Audio and Video Flows in Reactive Mode

Receiver
Media

Quality Monitoring in Trace Mode

Source

18
6.1.3 Probe Topology and Distributed and Hierarchical
Management on the U2520

Probes among different administrative regions or among different

districts in a city are connected in full-mesh mode. That is, a full-mesh
topology is established among all the nodes deployed with probes, as
shown in the following figures.

IP Network

The probes from administrative regions to the egress of the provincial

backbone network or from districts in a city to the egress of the city
network are connected in Hub&Spoken mode. That is, the probe on
the central node is connected to the probe on each edge node, as
shown in the following figures.

Responder
Controller
IP Network
Responder

Responder

Here, a hierarchical U2520 management system is formed.

19
 7 Indicator-based Evaluation
System
7.1 SLA Evaluation

As a formal agreement that is reached through the communication

and negotiation between two parties, the SLA is widely applied to
various fields. In the telecom industry, a group may sign SLAs with
its branch companies, a carrier may sign SLAs with SPs, and a carrier
may frequently sign SLAs with end users. An SLA can cover various
contents, such as service performance, problem rectification time, and
fault time.

With the rapid development of telecom services, the market value

chain becomes increasingly complicated, and more and more SPs
provide telecom services for end users. In that situation, the quality
of services provided to end users can be guaranteed only when SLA-
based monitoring is performed throughout the value chain. Any
misunderstanding of SLAs may affect the quality of E2E services. Figure
7-1 shows the key points of SLA evaluation.

To define overall standards for evaluating networks and services

To abstract networks and services

that are to be evaluated To establish an indicator system
SLA

Inventory Metric

Measurement
Measurement

To obtain data through tests

Figure 7-1 Key points of SLA evaluation

20
The U2520 is mainly used to perform SLA evaluation of IP services
and networks. For the U2520, a system of most concerned indicators
involved in the IP service development and IP network operation of
carriers is established, thus forming a series of SLA clauses. Users can
add required indicators to the SLA according to different situations.
The indicators fall into the following types:

•• HSI service indicators: for example, PPPoE dial-up duration, DNS

parsing time, and HTTP download time

•• VoIP service indicators: for example, mean opinion score (MOS),

delay, jitter, and packet loss ratio

•• IPTV service indicators: for example, v-MOS, MDI (DF, MLR, MR,
and MLT) delay, jitter, and packet loss ratio

•• Network indicators: for example, delay, jitter, packet loss ratio,

bandwidth usage, LSP delay, LSP jitter, LSP packet loss ratio, LSP TE
delay, LSP TE jitter, and LSP TE packet loss ratio

The U2520 adopts the SLA to evaluate services and networks. In

terms of service evaluation, the SLA can be used to evaluate a service,
a SAP, or a SAP group. When the SLA is used to evaluate a service,
all the SAPs of the service are evaluated, and the quality of each
SAP is assigned a weight that affects the service quality. In terms of
network evaluation, all the inventories, such as LSPs, connections,
trails, and interfaces, associated with the network are evaluated.
According to quality indicators that are added to the SLA by users
and inventories associated with the network or a certain service, the
U2520 automatically generates test instances in batches and modifies
test instances when inventories change, thus reducing the workload
of manual configuration.

21
 7.2 Hierarchical Measurement
Service KQIs and
SLA KQI
network KQIs
KQIs
Intermediate associat ed
KQI with the SAP, LSP,
connection, path,
KPI board, and int erface

Original
data
Service test Service Network Network
result statistics test result statistics
Figure 7-2 Hierarchical indicator system

7.2.1 KPI

In general, carriers periodically report key performance indicators (KPIs)

of networks or services. As the most concerned data, KPIs actually
reflect the performance of a network or a service at a certain time
point. The KPIs, however, cannot provide information about the E2E
performance, the service quality or network quality in a certain period,
or the overall service status. Despite the preceding disadvantages, KPIs
still play an important role in the evaluation system because of its data
aboriginality and wide application.

KPI calculation method:

The KPI indicates the performance value of an object at a certain time

point. A KPI may be obtained by the aggregation of multiple levels of
KPIs. The KQI indicates the threshold-crossing rate of multiple KPIs of
an object in a certain period. A KQI is calculated based on the lower-
level KQIs or KPIs.

Unlike most KPIs that directly come from original test results, certain
KPIs are calculated based on multiple test results.

Example:
Assume that a VideoRTP test needs to be carried out, and
VideoUpJitterKPI needs to be collected.

The KPI calculation formula is as follows:

22
KPI = [AboveThPkts/(AboveThPkts + BelowPkts + BetweeenThPkts)] x
100%
The indicators to be collected are as follows:

AboveThPkts, BelowThPkts, BetweenThPkts, MaxJitter, and MinJitter

7.2.2 KQI

In consideration of KPIs' disadvantages, KQIs are developed to

comprehensively, accurately, and continuously evaluate network
status. KQIs are used to evaluate the quality of an object, such as a
product, product component, service, service element, or network,
thus reflecting the object's health. A KQI is aggregated by KPIs or
original data. During the aggregation, users can set the weights of
original indicators according to the importance of the indicators, thus
forming a hierarchical KQI system.

KQI calculation method:

For a service, a SAP KQI is obtained by calculating the threshold-

crossing percentages of all KPIs of a SAP; an SLA KQI (service KQI)
is obtained by averaging the KQIs of all SAPs. During averaging, the
number of KPIs used for SAP KQI calculation is taken as a weight.

For a network, a network KQI is obtained by aggregating the KPIs

and KQIs of connections and trails. A network KQI can also be
obtained by aggregating board KQIs and device KQIs, both of which

23
 are aggregated by interface KPIs and interface KQIs. During the
aggregation from lower-level KQIs to upper-level KQIs, all the lower- Time pe
Ty
ic
etr
level KQIs should be averaged, and the number of KPIs used for lower- M
level KQI calculation should be taken as a weight during averaging.

Note: In fact, an upper-level KQI can be obtained through KPI

Inven
calculation. The KQI value is the same as that obtained by aggregating tory I
nstan
ce
lower-level KQIs based on weights.

Certain KQIs, such as the SAP KQI and channel KQI, are used only in Figure 7-3 Three-dimensional KQI aggregation

reports or for trend viewing.

Example:

Assume that a VideoRtpUpJitterKQI test needs to be carried out, and

the video RTP [abnormal jitter packet rate] KQI needs to be collected.
The KQI calculation formula is as follows:

KQI = [∑(A(KPI))]/n

Required KPI: VideoUpJitterKPI

The U2520 adopts three-dimensional KQI aggregation. That is, KQIs

are aggregated by dimensions of time, metric, and inventory. On the
U2520, one indicator can be aggregated into a KQI by time, which
is a common aggregation mode. In addition, different indicators
or indicators of different inventories can be aggregated into a KQI
by weight, and indicators based on different dimensions can be
calculated at the same time.

As a basic part of an SLA, KQIs are basically used to calculate SLA

compliance. In general, one SLA contains multiple KQIs. The SLAs
applicable to the U2520 are classified into customer SLAs (also referred
to as service SLAs) and internal SLAs (also referred to as network SLAs).
Customer SLAs contain only service-related KQIs that indicate the
quality of services provided by SPs to customers; internal SLAs are used
by carriers to evaluate the quality of internal networks.

The U2520 helps to form a hierarchical network structure through an

inventory model and metric instances developed according to the SLA
definition.

24
In the hierarchical network structure, metric instances are associated
with each other, indicating the rules for indicator calculation. These
rules are called indicator aggregation calculation rules. The U2520
supports the following types of indicator aggregation calculation:

•• Indicator aggregation calculation by inventory (hereinafter referred

to as I calculation)

This type of calculation aggregates the indicators of lower-level

inventories into indicators of upper-level inventories according to
the aggregation relation or including relation between inventories.
According to whether the weights of inventories are considered in
calculation, I calculation is subclassified as follows:

− I calculation without considering the weights of inventories

Examples: protocol interface traffic, protocol site traffic, and
network traffic

− I calculation taking the weights of inventories into

consideration

Example: SAP KQIs are aggregated into a service KQI on the

condition that SAP weights are considered.

•• Indicator aggregation calculation by metric (hereinafter referred to

as M calculation)

•• Indicator aggregation calculation by time (hereinafter referred to

as T calculation)

Test results displayed on the U2520 are original data. After important
original data is calculated, a KPI that reflects the performance of an
object at a certain time point can be obtained. A KQI is aggregated
by one or more KPIs, and KQIs can be aggregated into an upper-level
KQI. Only the upper-most KQI is displayed in SLA clauses and takes
part in SLA compliance calculation.

25
 8 Reference Standards
8.1 MOS

The MOS indicator defined by ITU-T is used to evaluate the service

quality through scores from 1 to 5. It is the most popular quality
indicator that is used to show user satisfaction. In addition, it is a main
indicator that is used to monitor VoIP signaling.

R-value ( lower limit ) MOSCQE ( lower limit ) User satisfaction

90 4.34 Very satisfied

80 4.03 satisfied
70 3.60 Some users dissatisfied
60 3.10 Many users dissatisfied
50 2.58 nearly all users dissatisfied

Figure 8-1 User satisfaction corresponding to different MOS scores

Based on ITU-T G.107E-Model, the MOS indicator on the U2520 is

developed in compliance with the ETSI 101329-5 standard recognized
by European Telecommunications Standards Institute (ETSI).

Jitter buffer is another important indicator that affects user satisfaction.

During the processes of generating, sending, receiving, and processing
voice packets, problems such as delay, jitter, and packet out-of-order
may occur. To reduce the impact of these problems on packets, the
decoder temporarily stores packets in a jitter buffer so that the packets
can be played at an even rate. The jitter buffer size of a user's decoder
directly affects user satisfaction in voice quality. The U2520 can simulate
the jitter buffer size of the user's decoder, which helps the measured voice
quality be consistent with the experience of end users.

26
8.2 RFC4445-MDI

The MDI indicator defined in the RFC4445 standard is used to

evaluate the quality of transmitting IPTV video flows on an IP
network. This indicator involves the delay factor (DF) and media loss
rate (MLR) parameters.

As a very important evaluation indicator for the VoIP network, MDI

can be used to precisely measure and monitor network jitters and
delays that affect the video transmission quality. In addition, it can
accurately reflect the quality of a lot of concurrent media flows, and
provide measurement results that are more accurate than the results
of subjective observation. Network evaluation helps to determine
how many IPTV users are supported by a network so as to provide
a basis for network design and device deployment. It can also help
to analyze potential problems on a network so that users can take
correct protection measures before faults occur on the network.

The DF value, expressed in milliseconds, indicates the delay and jitter

of tested video flows. With the DF, jitter changes of video flows are
converted into requirements for video transmission and decoder
buffering. A higher jitter level of tested video flows indicates a greater
DF value. If the time of video contents stored in the network device
or the decoder's buffer is no less than the DF value of tested video
flows, the playing quality of the video contents does not degrade.

The MLR, expressed by the number of media packets lost per second,
indicates the rate of media packet loss during the transmission of
tested media flows. The loss of encapsulated video packets directly
affects the playing quality of video contents. Therefore, during the
transmission of IP video flows, the desired MLR value should be zero.

8.3 VMOS

The VMOS model developed by Huawei is used to monitor the video

quality. Focusing on users' QoE, this model can be used to estimate
and quantify users' experience of video quality degradation according
to network factors of packet loss and jitter.

If the video playing quality is degraded but the video flows are
properly sent by the head end, packet loss or jitter during transmission
is the major cause. Packet loss leads to shrink of media information;

27
 jitter causes playing discontinuity. As a result, the playing quality
becomes poorer when more packets are lost or the jitter becomes
higher. Packet loss or jitter, however, does not directly affect users'
experience because users' experience is also affected by the video
type, basic compression damage, and forms of packet loss and jitter.
Compared with the MDI algorithm, the VMOS model adopts five MOS
values to represent different video quality experience of users. The
MOS values are 5 (very good), 4 (good), 3 (average), 2 (poor), and 1
(very poor). Users can determine whether the video quality is good
according to specific MOS values.

The MOS calculation formula is as follows:

MOS = V0 + V1 x Original experience value x Network damage factor

x Application damage factor x Terminal repairing factor

•• V0 and V1 are constants. Their values are from 1 to 4.

•• The MOS algorithm adopts ACR MOS values as references. ACR

MOS values are 1, 2, 3, 4, and 5, among which 1 indicates the
poorest quality and 5 indicates the best quality.

•• The original experience value indicates the video quality experience

in the case where video files are directly played without being
transmitted on the network. This value is affected by factors such
as the coding type (for example, MPEG2 or H264), coding bit rate,
and frame rate.

•• The network damage factor indicates the experience degradation

caused by network factors such as packet loss, jitter, packet out-of-
order, and delay.

•• The application damage factor indicates the experience degradation

caused by errors of application-layer parameters, such as the error
of TS parameters.

•• The terminal repairing factor is the value that is calculated

according to network factors such as the terminal decoder type,
retransmission of lost packets, and AL-FEC.

•• The network damage factor, represented by P-net, is calculated

according to network factors such as packet loss, packet out-of-
order, delay, and burst.
28
9 Typical Applications

Access Metro Core Application

DSLAM
Internet

BTV
Headend

DSLAM

MonitorB Verifier E Verifier D VerifierC Verifier B Verifier A Monitor A

VMoS=3.2 VMoS=3.2 VMoS=3.2 VMoS=3.2 VMoS=4.8 VMoS=4.8

Figure 9-1 Typical application of the flow-by-flow IPTV service monitoring scheme

9.1 Flow-by-flow MDI and VMOS Video Monitoring

Scheme

In principle, the flow-by-flow MDI and VMOS video monitoring

scheme is used to view the quality of service flow bearing on each
node. In compliance with the RFC4445 standard, the U2520 provides
measurement indicators (MDI and VMOS) that are the same as those
used by service departments to evaluate the quality of a service
system. The U2520 also provides information about the bearing
quality comparison of all nodes on the bearer network, which helps to
locate the links whose quality degrades and to identify the root cause
of degradation. In addition, the U2520 supports the networking of
devices from different vendors.

29
 MQMC

Encoder
Central node Regional node Edge node

HMS HMS HMS

Backbone MAN
network
Router Router Router BAS LS W DS LAM STB

Detection
U2520: deployment of network nodes and fast location of faulty points point

Figure 9-2 Flow-by-flow MDI and VMOS video monitoring scheme

The U2520 detects the entire path of video flows on a flow-by-flow

basis to identify faulty points.

In cooperation with the MQMC (IPTV service platform monitoring

system), the U2520 helps service maintenance engineers to monitor
the video quality in E2E mode.

The main test indicators on the U2520 are as follows:

•• IP transport indicators
− Packet loss
− Jitter
− Discarded packet
− Packet out-of-order
− MDI: Loss rate and DF
•• Media flow transmission indicators
− PAT/PMT/NIT/CAT frequency and error
− Sync error
− PID missing error
− TS error
− PCR error
− PRS error
− CRC error
•• Video contents-related indicators
− Frame rate
− Frame loss
− ETSI TR 101-290 parameter
− PCR jitter
− Media quality index (MQI)
− Noticeble loss (IEFT RFC 3357)

30
10 Interworking
10.1 EANTC Test (Video Monitoring Test)

In January 2010, the European Advanced Networking Test Center

(EANTC) carried out a series of multi-vendor MPLS interworking tests.
Over 10 equipment vendors and testing instrument vendors, such as
Huawei, Cisco, Alcatel-Lucent, Juniper, and ZTE, took part in the tests.
Figure 10-1 shows the physical network topology used in the tests.

Figure 10-1 Physical network topology used in EANTC tests

Huawei, Cisco, and Spirent took part in the video monitoring test. To
carry out the test, Spirent provided the service head end system and
the hardware damage environment; Huawei provided the NE40E and
the U2520 to interwork with Cisco ASR900. Then, according to the
RFC4445 standard, the EANTC tested the video monitoring capabilities
of products from these vendors.

During testing, Huawei adopted the U2520 that was deployed with
the internal and external probes, becoming the sole vendor that could
provide the E2E flow-by-flow MDI value. As an upper-layer monitoring

31
 system, the U2520 supports the E2E information display in GUIs
and helps to perform SLA monitoring of services. When the damage
Spirent
environment was started by the disturber from Spirent, the external TcstCenter

probe on the U2520 and Cisco ASR9000 immediately showed that the Huawei
NE40E-X8
bearing quality of the node from Cisco had degraded. Through the
NEU100, the U2520 succeeded in displaying indicator values that are the
Spirent
same as those provided by Cisco. In addition, the U2520 supports the xGEM Huawei
U2520
abilities to display the status of each node and to compare the indicator
Cisco
ASR 9000
data of all nodes so that users can quickly identify and locate faults.
Huawei
The EANTC carries out continuous tests in compliance with the RFC4445 NEU100
standard. The test results show that Huawei U2520 supports the E2E MPEG Transport Stream
SNMP
video monitoring solution and provides the function of detecting the
quality of a third-party device. In addition, indicators can be aggregated
Figure 10-2 Topology view associated with a
and associated with SLAs. The U2520 supports data display through video monitoring test
GUIs, which helps to precisely reflect network devices' capabilities of
bearing video services.

For descriptions of EANTC test instances, see the EANTC-MPLSEWC2010-

WhitePaper.

10.2 Compliant Standards and Protocols

Table 10-1 and Table 10-2 list the standards and protocols that the
U2520 complies with.

Standard/Protocol
Standard/Protocol Title
Number
Triple-play Services Quality of Experience (QoE)
DSL Forum TR-126
Requirements
Digital Video Broadcasting (DVB)
ETSI TR 101 290
Measurement guidelines for DVB systems
Information technology — Generic coding
ISO/IEC 13818-1 of moving pictures and associated audio
information: Systems
The E-model, a computational model for use
ITU-T G.107
in transmission planning
Pulse code modulation (PCM) of voice
ITU-T G.711
frequencies
Dual rate speech coder for multimedia
ITU-T G.723 communications transmitting at 5.3 & 6.3
kbit/s

32
 Standard/Protocol
Standard/Protocol Title
Number
40, 32, 24, 16 kbit/s Adaptive Differential
ITU-T G.726
Pulse Code Modulation (ADPCM)
Coding of speech at 8 kbit/s using conjugate-
ITU-T G.729 structure algebraic-code-excited linear-
prediction (CS-ACELP)
Quality of experience requirements for IPTV
ITU-T G.1080
services
ITU-T G.1081 Performance monitoring points for IPTV
ITU-T H.248 Gateway control protocol
Packet-based multimedia communications
ITU-T H.323
systems
RFC 0768 User Datagram Protocol
RFC 0791 Internet Protocol
RFC 0792 Internet Control Message Protocol
RFC 0793 Transmission Control Protocol
RFC 0854 Telnet Protocol Specification
RFC 0959 File Transfer Protocol
RFC 1034 Domain names - concepts and facilities
Domain names - implementation and
RFC 1035
specification
RFC 1305 Network Time Protocol (Version 3)
RFC 1350 The TFTP Protocol (Revision 2)
RFC 2131 Dynamic Host Configuration Protocol
Internet Group Management Protocol,
RFC 2236
Version 2
RFC 2326 Real Time Streaming Protocol (RTSP)
RFC 2327 SDP: Session Description Protocol
RFC 2330 Framework for IP Performance Metrics
A Method for Transmitting PPP Over Ethernet
RFC 2516
(PPPoE)
RFC 2543 SIP: Session Initiation Protocol
RFC 2616 Hypertext Transfer Protocol -- HTTP/1.1
RFC 2679 A One-way Delay Metric for IPPM
RFC 2680 A One-way Packet Loss Metric for IPPM
IP Packet Delay Variation Metric for IP
RFC 3393
Performance Metrics (IPPM)

33
 Standard/Protocol
Standard/Protocol Title
Number
Management Information Base (MIB) for
RFC 3418 the Simple Network Management Protocol
(SNMP)
RFC 3433 Entity Sensor Management Information Base
Media Gateway Control Protocol (MGCP)
RFC 3435
Version 1.0
RTP: A Transport Protocol for Real-Time
RFC 3550
Applications
Dynamic Authorization Extensions to Remote
RFC 3576
Authentication Dial In User Service (RADIUS)
RFC 3611 RTP Control Protocol Extended Reports (RTCP XR)
RFC 4445 A Proposed Media Delivery Index (MDI)
Table 10-1 Network standards and protocols

Standard/Protocol
Standard/Protocol Title
Number
IEEE802.1ad Provider bridges
IEEE802.1ag Connectivity fault management
IEEE802.1D Media access control (MAC) bridges
IEEE802.1Q Virtual bridged local area networks
OAM functions and mechanisms for Ethernet
ITU-T G.1731
based networks
ITU-T G.8010 Architecture of Ethernet layer networks
Ethernet over Transport - Ethernet services
ITU-T G.8011
framework
ITU-T G.8012 Ethernet UNI and Ethernet NNI
Characteristics of Ethernet transport network
ITU-T G.8021
equipment functional blocks
ITU-T G.8031 Ethernet protection switching
Requirements and framework for Ethernet
MEF MEF2
service protection in metro Ethernet networks
Metro Ethernet network architecture
MEF MEF4
framework - Part 1: generic framework
Table 10-2 Ethernet service standards and protocols

34
11 Acronyms and Abbreviations
Acronym/Abbreviation Full Spelling

VoIP Voice over IP

SAP Service Access Point

SE Service Element

NQA Network Quality Analyse

SLA Service Level Agreement

KPI Key Performance Indicator

KQI Key Quality Indicator

DSQ Degraded Service Quality Event

35
36