Network Management Functions Performance
Network Management
1
Lectures Schedule
Week
Week 1
Week 2
Week 3
Week 4
Week 5
Week 6
Week 7
Topic
Computer Networks - Network Management Architectures & Applications
Network Management Standards Architectures & Applications
Simple Network Management Protocol - SNMP v1, ASN, MIB, BER
Network Management Functions - Fault
Simple Network Management Protocol - SNMP v2 - Configuration
Network Management Functions - Accounting
Midterm
Week 8
Week 9
Simple Network Management Protocol - SNMP v3 - Performance
Week 10
Week 11
Week 12
Remote Network Monitoring RMON 1, SLA
Week 13
Week 14
NM Project Presentations
NM Project Presentations
Week 15
NM Project Presentations
Network Management Functions – Security
Remote Network Monitoring RMON 2
Management Tools, Systems and Applications
2
OSI NM Architecture and Model
Manager
Network Management
Organization Model
Agent
Object
Structure of Management Information (SMI)
Information Model
Management Information Base (MIB)
Communication Model
Protocol Data Unit (PDU)
Configuration Management (CM)
Fault Management (FM)
Functional Model
Performance Management (PM)
Security Management (SM)a
Accounting Management (AM)
3
Network Management Functions
Corporate
Network
Management
Administration
Mgmt
Capacity
Mgmt
Performance
Mgmt
Configuration
Mgmt
Accounting
Mgmt



Provisioning
Mgmt
Traffic monitoring
Throughput
Utilisation
Fault
Mgmt
Security
Mgmt
4
SNMP Milestone
SGMP SNMP
draft
SNMP
SNMPv2
Security SMP (parties)
full
standard
SNMPv2
(community)
SNMPv3
implementation
experience
proposed
draft
standard
standard
draft
standard
proposed
standard
5
Standard
• SMI (Structure of Management Information)
Define how management information may look like
– SMIv1: RFC 1155
– SMIv2: RFC2578 (currently used)
• MIBs (Management Information Bases)
Tell what management information exists
– MIB-I: RFC1156
– MIB-II: RFC1213
• SNMP (Simple Network Management Protocol)
Define how information is exchanged
– SNMPv1: RFC1157
– SNMPv2: RFC1901, 1905, 1906
– SNMPv3: RFC2571-2575
6
Structure
7
Performance Management
• Issues of concern to the network manager
include:
–
–
–
–
–
What is the level of capacity utilization?
Is there excessive traffic?
Has throughput been reduced to unacceptable levels?
Are there bottlenecks?
Is response time increasing?
• Network managers need performance statistics
to help them plan, manage, and maintain large
networks
8
Performance Monitoring
• After availability, the second most important
characteristic of a network is performance
• Performance data must be available in
sufficient detail to be able to assess the
bottleneck that is causing the problem and to
plan appropriate remedial action
9
Performance Management: Definition
The Utilization of
• Network Resources
with
The Ability to meet
• User Service Level Objectives
10
Performance Management: Activities
• Information Gathering
– Current System (Traffic / Resources)
– System History (Logs)
• Performance Evaluation (Analysis)
For Different States / Conditions
• Understanding / Reporting
11
Network: Performance
Traffic
Process
Arrival
(Rate)
Service
(Time)
Load
Busy Period
Measures
Errors
Channel
Quality
Monitor
Load /
Capacity
Delay
Measures
Utilization
Throughput
12
Traffic
Load
Limits
It's Not The Volume, It's The Time It
Takes To Double...
• “It’s not the heat, it’s the humidity…”
• Example: Daily Usenet News volume (e.g.,
~200GB/day now, doubling every 6 mos.)
13
Performance Management Sub-Categories
and Related Activities
Collecting Baseline
Utilization Data
·
·
·
·
·
·
Collecting a History of
Utilization Data
· Measuring utilization and response times at different times of
the day
· Measuring utilization and response times on different days
over an extended period
Capacity Planning
· Manually graphing or using a network management tool to
graph utilization as a function of time to detect trends
· Preparing trend reports to document projected need for and
the cost of network expansion.
Measuring link utilization using a probe
Counting packets received/transmitted by a specific device
Measuring device processor usage
Monitoring device queue lengths
Monitoring device memory utilization
Measuring total response times
Performance Management Sub-Categories and
Related Activities (cont.)
Setting Notification
Thresholds
· Having a network management tool poll devices for values of
critical parameters and graphing these values as a function of
time
· Setting polling intervals
· Setting alarms/alerts on those parameters when the
threshold is reached or a percentage of it is reached
· Initiating an action when the threshold is reached such a
sending a message to the network manager.
Building Databases
· Having the network management tool create a database of
records containing device name, parameter, threshold and time
for off-line analysis.
· Using the database to extract time dependence of utilization
· Using the time dependence of parameters to decide when
network upgrades will be necessary to maintain performance
Running Network
Simulations
· Using a simulation tool to develop a model of the network
· Using the model’s parameters and utilization data to
optimize network performance
Latency
·
Query/Response time interval
Chapter 12
MRTG
• Multi Router Traffic Grouper (Oeticker and Rand)
• www.ee.ethz.ch/stats/mrtg/
• Generates graphic presentation of traffic on Web
• Daily view
• Weekly view
• Monthly view
• Yearly view
Example of Performance Management
Obtaining Baseline Measurements
• Baseline: report of network’s current state of
operation
– Baseline measurements allow comparison of future
performance increases or decreases caused by
network changes with past network performance
• The more data gathered while establishing the
baseline, the more accurate predictions will be
• Several software applications can perform
baselining
18
Performance and Fault Management
• Performance management: monitoring how well
links and devices are keeping up with demands
• Fault management: detection and signaling of
device, link, or component faults
• Organizations often use enterprise-wide network
management software
– At least one network management console collects
data from multiple networked devices at regular
intervals
• Polling
19
Performance and Fault Management
(continued)
• Each managed device runs a network
management agent
– Collects information about device’s operation and
provides it to network management application
• Definition of managed devices and data collected
in a Management Information Base (MIB)
• Simple Network Management Protocol (SNMP):
TCP/IP protocol used by agents to communicate
20
Performance and Fault Management
(continued)
21
Performance and Fault Management
(continued)
• Network management application can present
an administrator with several ways to view
and analyze data
• Network management applications are
challenging to configure and fine-tune
• Multi Router Traffic Grapher (MRTG):
command-line utility that uses SNMP to poll
devices, collects data in a log file, and
generates HTML-based views of data
22
Performance and Fault Management
(continued)
23
Network Statistics
•
•
•
•
Baseline
Trouble shooting
Capacity planning for the future
Reports
Traffic Load:
Source
HostTopN
Host 1
Host 2
Host 3
Host 4
Host 5
Host 6
Host 7
Host 8
Host 9
Host 10
0
100
200
Giga Octets
300
400
Traffic Load: Source/Destination
Protocol Distribution
Chapter 12
Traffic Load: Source
Figure 12.16 Load Statistics: Monitoring of Sources
Chapter 12
Traffic Load: Destination
Chapter 12
Traffic Load: Conversation
Chapter 12
Protocol Distribution
Figure 12.19 Protocol Distribution (NetMetrix)
Conclusions
•SLA
•Measure Performance
•Meet requirements
•Evaluations of the Behaviour &
Cost-Effectiveness of Network
Resources