Overview
of the IP Quality Monitor (IQM) solution
IP
Quality Monitor (IQM) is a hardware-software system intended for measurement,
monitoring and management of the end-to-end IP quality parameters. IQM takes
into account different classes of service and zone structure of IP network. In
addition it is possible to deploy distributed monitoring with number of remote
management domains.
In automatic mode IQM system performs IP quality
parameters measurement and control. Communication service provider (CSP) can use
IQM to deploy SLA as value added service to IP VPN. SLA control could be
carried out on any directions of customer’s networks, last miles, access,
distribution and core levels. Meanwhile, any network topology could be used for
measuring. Using IQMM system CSP can provide its customers (IP VPN, L2VPN) with
reliable information about network quality parameters.
The
traditional approach is applied to the quality parameters measurement:
specialized network devices (probes) are placed on the key network nodes.
Program IQM agents (IQMA) are running on the probes.
Any
x86 or ARM- platform running Linux can be a hardware platform for
IQMA. Requirements to the hardware are minimal: memory from 512 MB,
free disk space - from 1GB. Fanless “thin client” with flash memory, Desktop
PC, plug- computers can be used as the hardware platform for IQMA.
Also you can use IQM agents deployed in NSG router.
There
are two methods available for agents connecting: T-connection, in-line
connection. It is recommended to use T-connection mode, due to its higher
reliability. Agents can initiate the test session (active mode) or accept
requests for tests (passive mode). Besides, it is
possible to use various network devices (such as routers, switches, servers)
with UDP Echo service running as the passive agents.
In
the automatic mode (scheduled) or on demand agents send a number of test
packets to peer agents and measure
parameters of their delivery: loss, delay, variations of delay (jitter). There
are special tests to measure channel capacity. The received information is
processed and stored in text files, and then passed to the higher level — to
the core system for statistics processing and analysis. This system is named
IQM Manager (IQMM).
Using
the probes at the IP network key nodes will allow
to measure not only the end-to-end parameters, but also at the certain part of
the network. It facilitate further process of a network problems
troubleshooting. Thus, for example, it is possible to carry out quality
management on one of the most problematic divisions of
a network — on the last miles.
The
IQM system could monitor the number of networks with various requirements for
quality parameters. Control policies can be applied individually to each test,
or to zone-to-zone directions.
IQM
system can be integrated with other monitoring systems by different vendors.
Managed parameters
Quality
assurance includes the analysis of following parameters: IP packet loss,
round-trip delay time, one-way delay time, jitter (packet delay variations),
available bandwidth (channel capacity) and number of remarked packets.
Measurement of these parameters can be made in different classes of service,
for example: standard, premium and real-time. It is possible to take into
account the zone structure of an IP network. This approach permits quality
management in zone-to-zone directions instead of point-to-point.
Packet loss can be influenced by a number of
factors, including signal degradation over the network medium, overloaded network
links, corrupted packets rejected in-transit, faulty networking hardware,
faulty network drivers or normal routing routines. Packet loss is calculated as
a ratio of lost packets to the total number of packets sent over the IP network
in specified direction.
Round-trip delay time (RTD — round-trip delay time, RTT —
round-trip time, RTL — round-trip latency) is the time required for a packet to
travel from a specific source to a specific destination and back again.
One-Way Delay time (OWD) – the time that a packet
spends in travelling across the IP network from source to destination. OWD (as
opposite to RTD) can assure that the go and back paths are the same (or not) in
terms of congestion, number of hops, or Quality of Service (QoS).
Jitter (Jitter, IPDV — IP Packet Delay Variation, PDV
— packet delay variation). In IP networks, jitter is the variation in the delay
of the packets. IP Packet Delay Variation is defined in RFC 3393 as a
difference of the one way delays of selected packets. Let’s designate R as time
when packet was sent, and S as delivery time. Than PDV value for i-packet and
j-packet could be calculated as:
Di,j = (Rj - Ri) - (Sj - Si) = (Rj -
Sj) - (Ri - Si).
The
figure illustrates how the PDV occures
due to irregularity of IP packets delivery:
RFC 3550 defines incremental
approach for calculation of PDV:
Ji
= Ji-1 + (|Di-1,i| - Ji-1)/16
This
method is used by IQM agents for jitter measurement.
Bandwidth is value which is calculated by IQM
as a function of packet delivery speed on the reception agent.
Number of remarked packets – a number of packets (in %) delivered with changed value of class of
service.
Functional composition
of IQM
From
the technical point of view, IP Quality Monitor consists of two basic elements:
quality parameters measurement agents (IQMA) and management system (IQMM)
The following
figure illustrates the basic components of IQM.
IQM management system
IQMM
system is program complex with sole GUI front-end, performing following
functions:
- Management
of IQM
Systems:
IP
quality control policy configuration
Policy
violation actions configuration
Management
of agents deployed in the network
Tests
configuration
On-demand
tests initiation
- Agents
state monitoring
-
Automatic
statistical data processing:
Export
of data delivered from agents to SQL database
Data
pre-processing and consolidation
Data
storage
Data
analysis
-
IP
quality policy violation alarming
-
Display
of statistics in various forms
Tabular
representation
Graphical
Representation
of agent’s state and controlled channels on the geographical map
Monitoring
results reporting
Display
alarms about the detected violations of the control policy
-
Interactive
map functions. Interactive map represents performance characteristics of
network and network devices on geographical map with network topology applied
Total
bandwidth used on network links.
Errors
occurred on network links.
Network
devices performance parameters: CPU load, memory usage etc.
Colored
indication of network links utilization.
-
Administration:
User
authorization and system access control
User
and role profiles management
Subordinate
IQMM (distributed setup) systems access control
Customer’s
accounts management (Requires additional software module “IQM: Users Personal Area”)
In the presence of
the
additional software module IQMM-CISCO-SAA, IQMM supports functional work with
SAA (Service Assurance Agent) – built-in network devices manufactured by Cisco
Systems:
-
Managing
Cisco
SAA,
-
Data
collection from Cisco SAA,
-
Analysis
and presentation of statistical information is fully realized in IQMM system
functional, described in this document
Statistics of measurements
Statistics can be represented
in the graphical form as well as in tabular form. To obtain the necessary
reports a user can use a filter to restrict the query by a group of criteria: source
zone, target zone, active (source) agent, passive (destination) agent, test ID,
customer ID, class of service, type of controlled parameter.
Average data is also
available: system provides hourly and daily aggregated historical reports of
controlled parameters. The averaging period can be changed. The illustrations
below are reports with hourly and daily aggregation.
Violations control
There are several
means for operator notification:
- Representation of agent’s state and
controlled channels on the geographical map
- Alarms list represents violations occurred,
- Dashboards – table-view of alarms,
- Additional notification means are: e-mail,
syslog, SNMP Trap.
Map
Map
displays state of controlled agents and channels. The map contains controls
that allows: scaling, moving to the area of interest, get a link, switch
on/off layers. Agents of different types (IQM, UDP, SAA) and tests are
displayed on the map. Different types of agents have different
icons. The operational status of agents is coded by color.
Violations are displayed by colors according to
their importance. There is the possibility of obtaining additional information
about the status of objects by using pop-ups. A content of the map is updated
periodically.
Alarms list
Alarms
list provides the operator with access to a list of violation signals, allows
to analyze the problem, change the status of the signal, add a comment, remove
the signal. The importance of the signal is determined by the color.
Dashboards
The dashboard presents signals in
the summary table “source-zone – destination-zone”. This approach to the visualization allows locating
the problem areas in the zonal level.
Notifications
When
violation is discovered IQM can use the following means for operator
notification:
-
Sending
e-mail with the appropriate report,
-
Sending
syslog,
-
Sending
SNMP-Trap,
-
Logging
to the database.
The e-mail notification contains a
summary about the violation, graph and table.
Reports
SLA
- reports
SLA reports provide information about controlled
parameters availability according to used policy. The reports are substituted
in the form of slices for each parameter for each
test. The shares corresponds to the proportion of time that controlled
parameter was within the threshold of certain control policy.
QoS reports
The
QoS reports provides information about the minimum, average and maximum values
of monitored parameters over a specified period.
TopX QoS reports
TopX
QoS reports provides information about TopX worst directions from the quality
parameters monitoring point of view.
Interactive network map
Interactive map
functionality designed for network links and devices utilization monitoring.
Interactive map represents performance characteristics of network and network
devices on geographical map with network topology depicted:
Total bandwidth used on network
links.
Errors occurred on network links.
Network devices performance
parameters: CPU load, memory usage etc.
Colored indication of network links
utilization.
Distributed monitoring
Distributed
monitoring is possible by means of managing remote monitoring domains.
In
case of monitored network has regional-distributed structure with number of
traffic concentration points and regional networks it make sense to deploy
distributed multilayer monitoring system by dividing the whole network into a
number of monitoring domains. Each monitoring domain will consist of number of
IQM agents and local IQM manager. Local IQM manager will gather intra-domain
quality measurements from its own agents, perform analysis of data received,
and provide reports and alarms for regional operators. Central
IQM manager will manage all local IQM managers and all agents. If
needed, it could be possible to access the local IQM manager data from central
IQM manager.
The
figure depicted an example of deploying of distributed SLA
monitoring system. Regional operators use their own IQM management systems for
the local goals of monitoring and agents management. Intra-domain measurements
will be gathered by local IQM management systems. If needed, the rights for
management could be delegated to the center leaving view-only rights to
regional operators.
Personal Area
IQM
Personal Area - additional software that implements the function of limited customer’s
access to IQMM reports and maps. Using a Personal Area is
recommended to provide end users of a network with
information of the quality parameters status. It is also possible to use the
Personal Area by the territorial divisions of large telecom operators in the
case of delimitation of responsibility areas.
Personal
area’s users have access to the following IQM-information depending on user’s
role access profile:
- Graphical and table reports with
quality parameters measured on customer’s networks.
- SLA reports based on quality parameters
measured on customer's networks.
- Monitoring map with customer’s
network.
|