Transcript 1. Large Enterprise Networks

Large Enterprise Networks
Chapter 1
Network Management, MIBs, and MPLS
Stephen B. Morris
Student: Paul L. Martin III “Tre”
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
1
Overview

Chapter 1 presents a general overview of modern
enterprise network management




Importance of Network Management
Introduction to Network Management Pyramid
Goals of Network Management Systems (NMSs)
Understanding Simple Network Management Protocol (SNMP)
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
2
Large Enterprise Networks

What is a Business ENTERPRISE?


A business enterprise is a collection of
organizations and people formed to create and
deliver products to customers
What is an NETWORK?
 A group of stations (computers, telephones, or
other devices) connected by communications
facilities for exchanging information. Connection
can be permanent, via cable, or temporary,
through telephone or other communications links.
The transmission medium can be physical (ie fiber
optic cable) or wireless (e.g. satellite).
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
3
Enterprise Networks: Defined

Two Main Categories

Enterprise Network


A networking system that allows communication and
resource sharing among all of a company's business
functions and workers. This can even include the
company's suppliers and distributors.3
Service Provider

SP’s generate contractually repeating revenues for the
services delivered to their customers over a network,
typically the Internet. These services must be
commercially available to the general population. The
price for such services must include the actual delivery of
the service including accommodations for the cost of the
infrastructure to deliver the service (e.g. hardware,
software, data center, labor, IP).4
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
4
Main Goals of the Enterprise Network

Two Main Focus Areas


Provide or improve business processes
Save the organization money rather than act as a revenue
source
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
5
Enterprise Network Characteristics






Large in Size (Equipment & Personnel)
Can be Geographically Separated
Can maintain Legacy Equipment
Generally Hard to Manage
Network Scalability affects
Network Manageability &
Network Usability
Generally owned by one organization
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
6
Enterprise Network Characteristics
(Con’t)






Many users simultaneously supported
Wide Range of Multi-vendor devices
Network Elements can contain other
intelligent devices
Individual N.Es can provide multiple
services
Specialized Servers provide advanced
services (I.e., SAN servers)
All network services are used as
essential business process components
by organizational personnel
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
7
Enterprise Network
Functional Components

Network Management
Challenges include:



Complex apps & services
Ever-changing apps & services
Geographically-dispersed
locations
and Personnel
All Bold text boxes provide some type of service
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
8
Enterprise Networks: Advantages







Centralized computing facilitates data sharing
Centralized computing facilitates data backups
Centralized management of software & resources
Network Authentication/Authorization can be
enforced
Network Administrators can perform remote
software installations
Expensive devices (laser printers, scanners, etc.,)
can be shared
Users can access their files from any workstation
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
9
Enterprise Networks: Disadvantages



Expensive to build, operate, maintain,
and upgrade
Require skilled maintenance & support personnel
Data & Voice traffic traditionally kept separate –
now merging
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
10
What is Network Management?


A set of activities (e.g. network monitoring, gathering
and analyzing the statistics, adjusting network
configuration) performed in order to increase the
network performance and availability5
The process and techniques of remotely or locally
monitoring and configuring networks. Under the OSI
model network management takes account of five key
areas: configuration management, fault management,
performance management, accounting management,
and security management6 (FCAPS)
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
11
What is Network Management?

Network management provides the means to
keep network up and running in as orderly a
fashion as possible.

Functional areas required for effective network
management include: FCAPS
 Fault – Detecting network errors/break
downs
 Configuration – The set up and fine tuning
of s/w & h/w into an existing infrastructure
 Accounting – Financial responsibilities such
as paying SPs, verifying charges
 Performance – Ensuring the network is
operating to standards/expectations;


Evaluate current and future performance
metrics
Figure: Taken from the
Telecommunications Management
Network architecture definition created by
the International Telecommunications
Union in 1988.
Source: http://www.luteus.biz/
Download/LoriotPro_Doc/V4/LoriotProV4
Doc/N14Management_Goals/FCAPS_EN.
htm#FaultManagement
Security – Protect network vs. hackers, or
malicious activity
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
12
What is Network Management?

BML - Business Management Layer


SML - Service Management Layer


Manage the overall business, gaining return on
investment, market share, employee
satisfaction, community and governmental
goal.
Manage the service offered to customer or
internal users, meeting customer service level,
service quality, cost and time-to-market
objectives
NML - Network Management Layer

Manage the network and systems that deliver
those services, capacity, diversity, and
congestion
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
13
What is Network Management?

EML - Element Management Layer


Mange the elements comprising the networks
and systems
NEL - Network Element Layer

Switches, routers, transmission, distribution
systems
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
14
The Management System Pyramid

OSS – Operations Support System

System that handles workflows, management
inventory details, capacity planning,
and repair functions for SPs





OSS used by the business support system
OSS uses underlying NMS to communicate
with lower level devices
Expensive to deploy and develop
Approx. Cost: $1M+
Example: AceComm is an OSS vendor
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
15
The Management System Pyramid

NMS – Network Management System

Computer based SW application suite that
manages N.Es


Provides abstractions (signaling links,
virtual connections, etc), fault & network
configuration, retrieve performance &
billing data, execute provisioning,
security, script management, audit trails
Network-wide oversight & usage




Oversight over many N.Es (not just one)
Uses EMS to communicate with N.Es
Approx. Cost: $10K+
Example: Altiris can be used as an NMS
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
16
The Management System Pyramid

EMS – Element Management Systems



Manages one or more of a specific type of NEs
EMS allows the user to manage all the features
of each NE individually
Specific functions include:




S/W upload/download
Configuration DB backup/restore
Alarm processing & storage
Approx. Cost: $1K+
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
17
Key Difference (NMS vs EMS)


Easy to confuse NMS and EMS roles/functions
Key general difference is:



NMS operations involve more than one NE simultaneously
EMS operations focus on a single NE
The guiding principle is the same for any
OSS/NMS/EMS:

Make using the object of attention easier to use
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
18
Why is Network Management
Important?

Maintains Network Availability





Network keeps running
Law of Five Nines: The definition of
system uptime or availability as 99.999%,
or an approx. downtime of 5 min/yr
Good Network Management facilities
assist in all the lifecycle stages
Overall Operational Costs are reduced
Manage multiple incompatible management systems

SNMPv3 uses MIBs for network management of data objects
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
19
Why Use Network Management?

Comprehensive Network Oversight


Record & Audit Trail Logging


An NMS maintains useful records &
audit trails of past configuration actions
Unsupported Management Protocols


Management systems maintain entire
network oversight - N.Es typically do not
If N.Es don’t support SNMP,
then a NMS can facilitate a superior CLI
Network-wide service implementation

NMS can facilitate network wide service like
(I.e., traffic engineering, QoS, planning, modeling, & backup/restore)
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
20
Why Use Network Management?

Fast Fault Rectification



Rebalancing Facilitation


NMS enable fast access to faults.
Some network faults can only be handled
by an NMS
NMS assist in rebalancing networks
after new hardware is added
Network-wide Object Support

Management system can provide
network wide object support for service profile
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
21
Common Network Management
Challenges

Backward Compatibility





Rare to experience “forklift”
upgrades
General expectation: rich
mixture of old & new N.Es
Result: Complex set of MIBs
deployed across network
Multiplicity of Management
Systems
Lack of Standards-based
management system
consolidation
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
Older
N.Es
22
Advantages of Standards-Based
Consolidation





Fewer & simpler user
management interfaces
Reduced IT staff training time
Fast fault identification &
problem resolution
Easier integration of new
hardware/software to overall
infrastructure
Management system can
provide network wide object
support for service profile
(lessons learned; case studies
for future reference)
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
23
The Manageability Factor

For the number of reasons, not all NEs lend themselves
to flexible, integrated, centralized management.

This tends to add to the cost of ownership due to the following
range of reasons:






The NE is a legacy device with proprietary management
infrastructure
The NE implements only SNMPv1 with support for set operations
The NE implements only SNMPv1 without support for set (a set
operations is an update to a network-resident manage object
operations)
The NE supports SNMPv3, but it has been poorly implemented
The NE supports SNMP3 but has a number of low quality MIB
modules
An NE is considered to have good manageability if it supports a well
implemented SNMPv3 agent and a high-quality MIB
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
24
Operating & Managing Large Networks

Important Aspects of Network Management include:

Traffic-Management as # of services/apps increase






Bandwidth Management
Network & N.Es Availability
Network & N.E Status Monitoring
Discovery and Asset Inventory management
Network Configuration –



Resistance to attacks from both sides of the firewall
Scalability –


VLAN setup, SAN volume setup,storage
allocations, remote control software
Service level agreement (SLA) reporting, SLA verification
between an enterprise and SP
Security control


Measuring traffic levels and checking for network congestion
handling increased numbers of users, traffic, NEs,
Disaster recovery
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
25
SNMP
Layer 2, 3, and 2.5
Ports and Interfaces
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
26
Understanding OSI Layer 2 & 3

To understand Network Management, one must have a
thorough comprehension of OSI Layer 2 and Layer 3

Layer 2 – Data Link Layer

Defines rules for sending/receiving data across a physical
connection


Examples: ATM, Frame Relay, Ethernet Switch
Layer 3 – Network Layer


Ensures packets of information reach destination across multiple
point-to-point links
Interconnected networks joined by routers

Example Device: IP Router
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
27
Layers 2, 3, and 2.5



Layer 2 – Data Link Layer
Layer 3 – Network Layer
Layer 2.5 has been used to
categorize some protocols
that operate between layer
2 and layer 3.
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
28
Layer 2, 3, and 2.5

The primary protocols that SNMP implements:



User Datagram Protocol (UDP) and the
Internet Protocol (IP)
SNMP also requires Data Link Layer protocols (e.g.
Ethernet, Token Ring) to implement the communication
channel between manager and agent

Data Link Layer


Provides the means to transfer data between network entities and to
detect and possibly correct errors that may occur in the Physical layer.
The addressing scheme is physical which means that the addresses
(MAC address) are hard-coded into the network cards at the time of
manufacture
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
29
MPLS on Layer 2.5

Multiprotocol Label Switching

A method used to increase the speed of network traffic flow by
inserting information about a specific path the packet is taking en
route to its destination.




Saves the time needed for a router to look up the address for the
next receiving node
MPLS is multiprotocol in that it works with IP, ATM, and Frame Relay
communications methods
MPLS has some Quality of Service features that make it an attractive
communications technique.
Supplemental Link: Click here
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
30
How MPLS Works
Click here for full article
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
31
MPLS Advantages/Disadvantages

Advantages



Can be deployed on routers
Paths can be reserved before traffic arrives at the network
Different QoS options can be applied


Traditional IP routing protocols can be used



Ex: Higher QoS can be reserved for VoIP; lower for e-mail
Ex: OSPF, IS-IS, BGP4
Congested route problems can be reduced through dynamic
traffic monitoring and engineering
Disadvantage

All nodes in the path must run MPLS protocols – additional
burden
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
32
Ports and Interfaces

Terms “Ports” and “Interfaces” often used
interchangeably, which is not necessarily the case

Ports – underlying hardware entities


Example: ATM or Ethernet ports
Interfaces – Exist at a higher abstraction layer



Configured to run on top of ports
Referred to as “logical ports”
Interface examples include:





Routing such as OSPF, IS-IS, BGB-4
Signaling, such as RSVP-TE and LDP
MPLS
IP
General Difference: Ports work out of the box, interfaces
must be configured
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
33
SNMP
Simple Network Management Protocol
Providing Network Management Capability
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
34
The Goal of a NMS

The difference between the realtime network situation and the
NMS picture of the network
situation must be as small as
possible


Administrators constantly strive to
know and/or be able to quickly find
out what is going on with their
network
The purpose of NMSs and SNMP:

Help administrators stay on top of
network operations, events, and
faults
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
35
What is SNMP?

A protocol used by network hosts to:






Exchange information,
Monitor and control network devices, and to
Manage configurations, statistics collection, performance, and
security used in the management of networks.
SNMP network management is based on the client and
server model
Used almost exclusively in TCP/IP networks, but not
limited to them
SNMP facilitates communication between network
devices
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
36
What is SNMP?
An SNMP-Managed
Network Consists of
Managed Devices,
Agents, and NMSs
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
37
What does SNMP do?

Notify network administrators about network
status through “network notifications”

Events

An indication from the network to the NMS of some item
of interest


Faults

An indication of a service-affecting network problem


EX: User logging onto a NE via Command Line Interface
(CLI)
EX: Communication line link failure
Alarms

An indication that a potentially service-affecting problem
is about to occur

EX: Congestion threshold being exceeded
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
38
Main SNMP Components

Principal Components of SNMP are:

1. Manager


An application that performs the operational roles of generating
requests to modify and retrieve management information, and
receiving the requested information and trap-event reports that are
generated by the SNMP agent
2. Agent

An application that performs the operational role of receiving and
processing requests, sending responses to the manager, and
sending traps when an event occurs
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
39
Main SNMP Components

Four Principal Components of SNMP are:

3. Management Information Base (MIB)



The set of parameters (database) that an SNMP management
station can query or set in the SNMP agent of a networked device
(e.g, router).
The unique identifier of each managed object includes the type
(such as counter, string, gauge, or address), access level (such as
read/write), size restrictions, and range information of the object.
4. Protocol Data Units (PDUs)

A data object exchanged by protocol machines (e.g. SNMP agents)
& consisting of both protocol control information and user data
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
40
SNMP Manager Role (Explained)

SNMP managers are the
entities that interact with
the agent



Establishing & obtaining
the values of MBI objects
instances on agent
Receiving notifications from
agents
Exchanging messages with
other managers
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
41
SNMP Manager Role (Explained)

Facilities offered by management systems are:






FCAP
A centralized database
Reporting Capabilities
Support for many simultaneous client users
Topology discovery
Full featured, multi-level Graphical User Interface (GUI)
representing the managed network
Click here to see Fujitsu’s NETSMART 500
Network Element Manager
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
42
SNMP Agent Role (Explained)

SNMP agent are the entities that
reside on manage devices.

Agent are the workhorses of
management & provide the
following functionality





Implementing and maintaining MIB
objects
Responding to management
operations such as requests
Generating trap & inform
notifications
Security Implementation
Set Access Policy for External
Managers
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
43
SNMP Agent Role (Explained)
The SNMP agent listens on UDP port 161
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
44
SNMP MIBs (Explained)

Each management station or agent in an SNMPmanaged network maintains a local database of
information relevant to network management, known
as the management information base (MIB)

An SNMP-compliant MIB


Managed objects/Management variables


Contains definitions and information about the
properties of managed resources and the services that
the agents support.
The manageable features of resources
A management station gets and sets objects in the
MIB, and an agent notifies the management station
of significant but unsolicited events called traps
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
45
SNMP MIBs (Explained)

MIBs






The most crucial/important NMS component
MIBs contain data definitions for managed objects
SNMP managers & agents exchange object instances using
SNMP protocol
Are Plain-text files
MIBs are compiled into agent source code -> executable file
Textual Conventions

MIB refinements (similar to programming language data types or
classes in Java or C++)
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
46
SNMP MIB (Explained)

The following keywords are used to define a MIB object:

Syntax


Access


Defines the abstract data structure corresponding to the object type
Defines whether the object value may only be retrieved but not
modified (read-only) or whether it may also be modified (readwrite)
Description

Contains a textual definition of the object type. The definition
provides all semantic definitions necessary for interpretation; it
typically contains information of the sort that would be
communicated in any ASN.1 commentary annotations associated
with the object
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
47
SNMP MIB (Explained)

Object Identifer (OID) – used
by the management station to
request the object's value from
the agent


OID - a sequence of integers that
uniquely identifies a managed
object by defining a path to that
object through a tree-like
structure called the OID tree or
registration tree
When an SNMP agent needs to
access a specific managed object,
it traverses the OID tree to find
the object.
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
Lexicographic Ordering
48
SNMP MIB & Lexicographic Ordering
The OID serves as a unique
name that represents a node
in the tree-based structure
All objects can be traced from
the root in a process called
“walking the MIB.”
During a walk, each branch of
the MIB is traversed from left
to right, starting at the root
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
49
SNMP PDUs (Explained)

SNMP uses “very simple messaging protocol”

Three basic commands




Fetch (GET)
Store (SET)
Notification/Inform Message
Each SNMP message has the format



Version Number
Community Name - kind of a password
One or more SNMP PDUs - assuming trivial authentication
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
50
SNMP PDUs (Explained)

SNMPv1 originally defined six PDUs


These PDUs have been redefined over the years
The current SNMP Framework categorizes the PDUs
into different classes.
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
51
SNMP “Get” PDU
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
52
SNMP “Get-Next” PDU
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
53
SNMP “Set” Protocol
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
54
SNMP “Trap” PDU
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
55
SNMPv1

SNMPv1 is probably best known for its relative simplicity


Much more complicated than following versions
SNMPv1 message format is simple & straight-forward
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
56
SNMPv2

Certain issues with SNMPv1 were noticed & areas for
improvement identified




MIB object definitions
Protocol operations
security
Several versions of SNMPv2 & several message formats


PDU format is the same for all the SNMPv2 types,
The overall message format differs for each variant.
Click here to see SNMPv2 Message Format
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
57
SNMPv3



Created in the late 1990s
SNMP version 3 resolved the problems that occurred
with the many different variations of SNMPv2
Adopts many components created in SNMPv2 including:




SNMPv2 protocol operations,
PDU types and
PDU format.
Significant changes made in SNMPv3 include

more flexible way of defining security methods and parameters –
allowing the coexistence of multiple security techniques
Click here to see SNMPv3 Message Format
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
58
SNMP Management Systems & Agents

SNMP manager



Any computer that sends queries for
IP-related information to a managed
computer
SNMP manager can send a request to
an SNMP agent to change a
configuration value
SNMP agent


Any computer or other network device
that monitors and responds to queries
from SNMP managers
Can send a trap message to the
manager when specified events (I.e.,
system reboots, illegal access
notification)
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
59
SNMP Manager/Agent Communication
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
60
Chapter 1 Summary


Large Enterprise Networks require good enterprise
management from both people/administrators and
systematic/technical components
The goals of enterprise networks:



Make organizational personnel more productive
Save the organization money (not act as a revenue producer)
The goal of network management:

Maintain network availability

“You can’t achieve the goals of enterprise networks if you can’t use
them.”
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
61
Chapter 1 Summary

The Simple Network Management Protocol has become the de
facto standard for internetwork management





It is a simple solution,
Requires little code to implement,
Vendors can easily build SNMP agents to their products
SNMP is extensible - allowing vendors to easily add network
management functions to their existing products
SNMP separates the management architecture from the architecture of
the hardware devices, which

Broadens the base of multi-vendor support
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
62
Supplemental Materials

General Network Management



MPLS


Article: “Network Management as Core Competency”
Computerworld, 20 March 2006
Article: “Cisco Moves to Support Network Management”
Computerworld, 12 December 2005
Article: “Building Large Metro Ethernets requires MPLS”
Converge! Network Digest, 24 October 2004
SNMP

Article: “Is it time to re-engineer SNMP”
Networkworld, 22 March 2004
Revised Spring 2006
Chap 1 - Large Enterprise
Networks
63