Sei sulla pagina 1di 46

Network Management and Software

Defined Networks
EEET2294 / EEET2295

Introduction to Network Management,


Maintenance and Security
References

1. Network Management Fundamentals


By Alexander Clemm
Published Nov 21, 2006 by Cisco Press.
Part of the Fundamentals series.

2. Advances in Network Management


By Jianguo Ding
Originally published: January 1, 2009

RMIT University©2017 Network Management and Software Defined Networks 2


General

• This session is all about setting the context of why network management is import
for telecommunications networks.
– You need to understand the environment that ISPs have to work in.
– What are the requirements of network management systems
– Why do we need to manage telecommunications systems using integrated,
open, distributed, object-oriented management systems
• Terminology
– You need to understand some of the key terminology used.
– Understand and be able to provide examples of
• Extensibility, Scalability
• Reliability, Availability
• Maintainability, Maturity
• and many others

RMIT University©2017 Network Management and Software Defined Networks 3


Subject main aims: Network Management

Awareness of:
• concepts & issues involved in the technical management of medium to large sized
telecommunication networks
• current technologies & techniques in addressing issues
• the context in which issues arise
Subject Philosophy:
• Emphasis is on generic concepts and techniques
– i.e. not on a particular implementation specific things come and go, but the
concepts remain the same
The context
– Political
– Social
– Technological

RMIT University©2017 Network Management and Software Defined Networks 4


The context: Political

• Global deregulation of traditional Government functions


– Telecommunications carriers no exception
– Increased competition promises:
• lower prices
• better service
• choice
• World-wide free trade
– Not a reality yet, but moving towards it, albeit slowly!
– Traditional tariff protections and bans disappearing
– Product markets are expanding towards truly global
• Doing business in many countries at once is getting easier
– International bodies such as the ITU helping to provide critical mass in
standards that are adopted by member Governments

RMIT University©2017 Network Management and Software Defined Networks 5


The context: Political

• Example: “The Great Firewall”

RMIT University©2017 Network Management and Software Defined Networks 6


The context: Social

• We are in the ‘information age’


– increasing reliance on telecommunications
systems for everyday life
– increasing expectation and demand for
access to ‘information’ from anywhere and at
any time.

• Society is becoming more educated


– Less need for manual labour
– Increasing need for skilled labour and jobs
that require skilled labour

• Shift from ‘Industrial Age’ [based on labour] to the


‘Knowledge Age’ [based on information]

RMIT University©2017 Network Management and Software Defined Networks 7


The context: Technological
• Technology changing the balance of infrastructure / labour
costs
– Electronic goods are getting less expensive
– global markets, economies of scale
– highly automated production & assembly

• Labour is becoming more expensive & more highly skilled


• The “death of distance”
– telecommunications technology is delivering cost-effective, high
quality, high bandwidth communications over any distance
 optic fibre digital transmission
 high speed switching

• Convenient, simple, access mechanisms


– wireless access modes providing high quality, low-cost,
reliable entry to the telecommunications infrastructure
 particularly for developing nations

RMIT University©2017 Network Management and Software Defined Networks 8


The context: Technological

• High-bandwidth access
– xDSL, cable modems, 3G mobile, LTE mobile (4G), GPON etc ->
5G

• Rapid change
– product life cycles getting shorter

 need for systems to cope with old and new technology

 research & development costs more significant for smaller


production runs

 increased use of “off the shelf” solutions as opposed to


custom-built

• Real-time (interactive) systems becoming


feasible
– high bandwidth, low latency, cost effective
switching systems becoming available
 Allows us to re-think the way that services can be
provided

 Allows us to introduce new services, not previously


possible

RMIT University©2017 Network Management and Software Defined Networks 9


Where have we come from? Network evolution

1. Moore’s Law: formulated by Gordon Moore of Intel in the early 1970s


– the processing power of a microchip doubles every 18 months
– computers become faster and the price of a given level of computing power
halves every 18 months.

RMIT University©2017 Network Management and Software Defined Networks 10


Where have we come from? Network evolution

2. Gilder’s Law: proposed by George Gilder in 1997, prolific author and prophet of
the new technology age
– the total bandwidth of communication systems triples every twelve months for
the next 25 years.
– Bandwidth grows at least three times faster than computer power.
– While computer power doubles every eighteen months (Moore’s law),
communications power doubles every six months.
– The cost per communication bit will begin to sink farther than it has fallen
previously.
– Eventually the cost of a telephone call, or of a bit transmitted, will be “free.”
– New developments seem to confirm that bandwidth availability will continue
to expand at a rate that supports Gilder’s Law.

RMIT University©2017 Network Management and Software Defined Networks 11


Where have we come from? Network evolution

2. Gilder’s Law: proposed by George Gilder in 1997, prolific author and prophet of
the new technology age

RMIT University©2017 Network Management and Software Defined Networks 12


Where have we come from?

• The telephone network


– Been around for ›100 years!
– designed specifically for voice
– operates on a circuit switched basis
– spans the entire globe
– very well managed (good QoS)
– not very flexible, difficult to add new services
• Quality of service
– Traditionally, telephone networks were well managed
 guaranteed bandwidth with each call
 guaranteed signal to noise ratio
 vigilant traffic analysis & forecasting ensured that network access was
maintained at a very high level. Eg. probability of blocking ‹1%

RMIT University©2017 Network Management and Software Defined Networks 13


Where have we come from?

• Computer networks
– evolved from around 1960’s onwards
– TCP/IP and the internet started bringing together LANs, MANs and WANs
– Very little regulation
– based on packet switching - share resources equally
– Huge markets, mass production
• Poorly managed historically
– if poor service, add more bandwidth, more switches etc. despite low
utilisation
• User expectation of service different
– Little or no QoS
 difficult or impossible to reserve bandwidth
 not designed for real-time traffic, so delay is a major problem

RMIT University©2017 Network Management and Software Defined Networks 14


Where have we now?

• Where are we now?


– Internet growth explosive
– mature development, industry standardisation through consortia
– computer networking technology challenging traditional telephony markets for
cost / performance
– Resultant boom in “everything over IP”
• The complexity of, generally speaking, managing the development of services
and networks is growing rapidly when new, customised services and applications
are introduced.
• Coping with management issues is likely to be the hardest part in developing the
future information infrastructure

RMIT University©2017 Network Management and Software Defined Networks 15


Where have we now?

• IP for real-time traffic?


– IP networks are not designed for real-time traffic, such as voice
– However, given sufficient resources, it can be done
– Much research work is currently being done to adapt IP networks to the
needs of real-time traffic
• Real-time traffic
– What are these needs ?
 minimum guaranteed bandwidth or throughput
 maximum end-to-end delay
 Eg. For interactive voice, ‹ 200ms
 maximum delay variation (‘jitter’)
 maximum %loss for given quality of speech

RMIT University©2017 Network Management and Software Defined Networks 16


Where have we now?

• Current IP technology
– Current IP technology cannot guarantee ANY of these requirements in very
large (telco-style) networks
• Industry trend
– The vast majority of telecommunications carriers (‘telcos’) are currently
scrapping the circuit switched core network built up over many decades and
replacing them with IP networks
 Circuit switched systems will be gone eventually!
• Why?
– Why are telcos doing this when they know that current IP technology cannot
do what existing technology can?
– There are several reasons, but the biggest one is of course MONEY
– Most telcos are businesses whose sole purpose for existence is to return a
profit to their shareholders

RMIT University©2017 Network Management and Software Defined Networks 17


What we learn?

• The scale of communication networks expanded:


– from local telephone networks to GSM (Global System for Mobile
communications) and beyond such us Long Term Evolution (LTE) and 5G
(expected by 2020)
– from the well-known infrastructure of cellular networks to non-infrastructure
wireless ad-hoc networks and integrated with various wireless systems and
Internet.
• The scale of computer networks evolves from ARPANET to LAN, PAN, WAN,
GAN, pervasive Internet, and paralleled with the emerging of various ad hoc
networks, even to nanonetworks.
• The technology updating enables the network data traffic from a few k/s to over
10G/s,
– allows the network to transfer from plain text file to images, voice, video, and
emerging rich media.

RMIT University©2017 Network Management and Software Defined Networks 18


What we learn?

• Networks act as from simple message transformation to complex network


services:
– e-commerce,
– e-government,
– e-learning,
– e-library,
– e-laboratory,
– e-health,
– e-society,
– online virtual reality,
– 3D Internet, etc.
• Ongoing evolution of networks indicates that networks will penetrate to various
aspects of our real: research, industry, culture, military, and social life.

RMIT University©2017 Network Management and Software Defined Networks 19


Where are we going?

• Nomadic computing (Mobility)


– Mobility in both the terminals and the services will have to be taken into
consideration in future network designs.
– The number of mobile networked devices as well as nomadic users will
increase dramatically.
– Subsequently more users and devices are connected and have direct
dynamic communication link.
• Wireless high-speed networks
– The network applications are being complemented with really high-capacity
and low-cost wireless access alternatives for finest possible access
granularity and largest coverage for high speed access to networks.
– For example, next generation Mobile WiMAX network could transmit data at a
speed of up to one gigabit per second while stationary, and 100 megabits per
second in a moving vehicle.

RMIT University©2017 Network Management and Software Defined Networks 20


Where are we going?

• Wireless high-speed networks


– Current cellular technologies such as HSDPA have data speed of up to 5
megabits per second (in the downlink), and its expected to increase
dramatically from current capacities in 3G and HSDPA towards HSDPA++,
3G LTE, 4G and beyond.
• Scalability
– The increasing scale of networks brings new challenges in a number of
areas. Examples include:
 modeling, validation, and verification of business processes
 brokering, aggregation, and data management.
– Quality of software is an important factor in all of these and will become
essential to the smooth working of the “service universe.”

RMIT University©2017 Network Management and Software Defined Networks 21


Where are we going?

• Convergence
– Telcos want one network that can provide all services for all users
– virtual circuit switched systems can deliver the required services
 Connection oriented packet switching
 Packet switching networks emulate circuit switched networks
• Rapid change
– IP networks give telcos the flexibility to rapidly introduce new services and
adapt to changing market conditions
– need to stay competitive
 Eg. Give customers direct access to configuration, billing and
performance measurement information

RMIT University©2017 Network Management and Software Defined Networks 22


Where are we going?

• Security (Trust)
– Security is becoming one most important footstone for modern network
services.
– Creating trusted environments for the new service world will require:
 mechanisms to monitor, display, and analyze information flows between
nodes participating in complex collaborations in order to detect and
assess security risk;
 mechanisms to ensure trust and confidence in services created by end-
users themselves, i.e., built-in safeguards and guarantees so that others
trust the new services.
 Peer-to-peer services today are mainly associated with activities of
doubtful legality, such as illegal trading of rights-protected content.
 Technical and legal mechanisms should be found to bring about changes
in attitudes.

RMIT University©2017 Network Management and Software Defined Networks 23


Where are we going?

• Interoperability
– Service interoperability to provide the ability to integrate largely stand-alone
services with similar ones and with other services, for instance from the
business domain;
– semantic interoperability, so as to provide the (automated) understanding of
the information exchanged and ensure quality of service;
– interoperability of the service layer with network and application layers from
different providers.
• Context-awareness
– The growing importance of context-awareness, targeting enriched
experience, intuitive communications services fitting mobile lifestyle, and
mobilized workforce will in the future lead
 intelligent services that are smart but invisible to users.
– The social and economic benefits of making ICT-based services in areas as
diverse as health, sustainable environment, safety, and transportation more
intelligent and adaptive are recognized as a new driver for network services.
RMIT University©2017 Network Management and Software Defined Networks 24
Where are we going?

• Automation
– We’ve all heard about telcos slashing their work forces
– IP networks will allow a huge automation of common telco tasks such as
faults, configuration, accounting, performance and security
– The labour has shifted from technicians to IP architects
• Network Management
– Network Management is one of the biggest challenges facing telcos as they
move to IP networks
– Part of the problem is managing the huge IP networks themselves, but
another part is managing a mix of technologies effectively
• Managing Change
– As the pace of technologies increases, we’re also having to operate
networks that have many ‘generations’ of technology in place
– Thus, even though we’re moving towards convergence, the process is
resulting in networks that are increasingly heterogeneous

RMIT University©2017 Network Management and Software Defined Networks 25


Where are we going?

• Autonomic computing
– Besides enhanced user experience for human-to-human or human-to-
machine interactions, autonomous machine-to-machine communication has
gained significant importance.
– More and more business transactions and processes will be automated and
will take place based on autonomous decisions without any human
intervention.
– These will be often based on or influenced by context information obtained
from the physical world, without the requirement of human input to describe
the situation.
– Enabling environment where real-world physical phenomena are
electronically sampled and influenced by heterogeneous sensors and
sensor/actuator islands and are at the fingertips of applications and
humans alike, thus linking the physical world with the future networks.
– Consequently, human environment can be adjusted to human needs, or we
can adjust our behavior following environmental changes.

RMIT University©2017 Network Management and Software Defined Networks 26


Where are we going?

• Integration
– The network will become increasingly integrated with phones, televisions
sets, home appliances, portable digital assistants, and a range of other small
hardware devices, providing an unprecedented, nearly uniform level of
integrated data communications.
– Users will be able to access, status, and control this connected infrastructure
from anywhere on the network.
• Expanded Services
– Services (not only those for the end Users but also network services) are
likely to be comprised of a variety of components, provided by a variety of
Players (e.g., ASP, Prosumers) and running over a decentralized hosting
(low-cost) infrastructure (including end-user devices, PC, servers, storage,
computing and networking/forwarding resources, etc.).

RMIT University©2017 Network Management and Software Defined Networks 27


Where are we going?

• Expanded Services
– This vision is expected to pave the way for a deep integration of service and
network frameworks for network convergence thus allowing broad federations
of Players (e.g., Network and Service Providers and Application Service
Providers) according to new business models.
– Openness, broad federations of Players, and do-it-yourself innovative
services and knowledge management will allow people to be the true center
of Information Society.

RMIT University©2017 Network Management and Software Defined Networks 28


Where are we going?

• Co-operate with competitors?


– Customers want end-to-end services
– Individual telcos cannot necessarily supply these with their own networks
– Need for telco management systems to interact seamlessly
– Need to be compatible!
• Stay Competitive
– New services need to be introduced rapidly, reliably and seamlessly and then
managed effectively
– This can only be achieved with modular, flexible automated network
management systems
• Follow market demand
– If lots of customers sign up for a service, we need to be able to easily and
rapidly scale the network infrastructure and the management system to cope
with the increased demand
 i.e. the system must be scalable
RMIT University©2017 Network Management and Software Defined Networks 29
Telecommunications Network Management

• What is it?
– The network operation must also ensure that
network resources are used effectively under
normal as well as under problem conditions.
– Traffic controls are necessary to ensure the
smooth flow of information through the
network.
– Network management functions includes:
 Monitoring the performance of the
network,
 Detecting and recovering from faults,
 Configuring the network resources,
 Maintaining accounting information for
cost and billing purposes,
 Providing security by controlling access
to the information flows in the network.

RMIT University©2017 Network Management and Software Defined Networks 30


Telecommunications Network Management

• What is it?
– For users:
 Ensuring that users of the network get access to a variety of services:
 when they want (on-demand)
 how they want (specified quality of service, reliability) & pay
accordingly
– For network operators:
 Ensuring that services can be reliably provided in a timely and
professional manner
 Ability to rapidly change services or characteristics
 Minimise cost / maximise profit
 Remain competitive in the marketplace

RMIT University©2017 Network Management and Software Defined Networks 31


Telecommunications Network Management

• What is it important?
– For telecommunications vendors:
 A way to make a lot of money, as operators will pay huge dollars for
effective network management tools
 A headache, because they can’t find enough people with the right skills
to employ for development and implementation of network management
tools

RMIT University©2017 Network Management and Software Defined Networks 32


Telecommunications Network Management

• Why is it important ?
– To remain competitive in the marketplace, operators must:
 provide services in a professional manner (certain quality of service)
 be able to deploy new services and manage these alongside their
existing infrastructure & within their operating environment
 minimise operating costs (labour!)
– Scale:
 Networks are getting bigger
 Need an efficient way of managing very large and complex,
heterogeneous networks
 Automated or semi-automated network management is one possibility

RMIT University©2017 Network Management and Software Defined Networks 33


Telecommunications Network Management

• Why is it important ?
– Complexity:
 Desire to make network management tools intelligent enough to allow
non-technical people to manage networks
 Also desire to implement certain business tools within a network
management framework
 Eg. Give all residential customers a 50% discount for calls longer
than 20 minutes for the next 4 Sundays

RMIT University©2017 Network Management and Software Defined Networks 34


Elements of Network Management

• Fault
• Accounting
• Configuration
• Performance
• Security
Fault Management Overview
• When fault occurs
 Determine where
 Isolate rest of network so it can continue to function
 Reconfigure or modify network to minimize impact of operation without
failed components
 Repair or replace failed components
• Fault is abnormal condition that requires action to repair
 Errors (e.g. single bit error on line) occur and are not faults
Fault Management User Requirements

• Fast and reliable problem resolution


• Immediate notification
• Impact and duration minimized by
redundant components and routes
 Fault tolerance
• Fault management capability should be
redundant
• After correction, fault management must
ensure problem is truly resolved and no
new problems introduced
 Problem tracking and control
• Minimal effect on network performance
Accounting Management
Overview
• Individual cost centers or projects charged for use of network services
• Network manager needs to be able to track use of network
 User or group may be abusing access
 Users may be making inefficient use of network
 Need to plan for network growth
• User requirements:
 Needs to specify sort of accounting information recorded at various
nodes
 Interval between successive sendings of information
 Algorithms used in calculating charges
 Accounting reports under network manager control
 Verify users' authorization to access and manipulate accounting
information
Configuration and Name Management Overview
• Choose appropriate software and attributes and values (e.g., a transport layer
retransmission timer) for device depending on function(s)
• Initializing network and gracefully shutting down
• Maintaining, adding, and updating relationships among components
• Status of components during network operation
• User Requirements
 Startup and shutdown
 Unattended
 Need to identify components of network and connectivity
 Define and modify default attributes
 Load predefined attributes
 Change connectivity
 Reconfiguration
 Status information and notification of changes
 Routine or request driven reports
 Authorized users (operators) only to manage and control operation (e.g.,
software distribution and updating)
Performance Management Overview

• Monitoring
 Tracks activities
 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?
• Controlling
 Make adjustments to improve
performance
• Identify resources to be monitored
 Metrics and values for resources
Performance Management User Requirements

• Must be known in sufficient detail to assess user queries


• Applications need consistently good response time
• Performance statistics help planning, management and maintenance
• Recognize potential bottlenecks before they cause problems
• Capacity planning based on performance information
• User Requirements
 Must be known in sufficient detail to assess user queries
 Applications need consistently good response time
 Performance statistics help planning, management and maintenance
 Recognize potential bottlenecks before they cause problems
 Capacity planning based on performance information
Security Management Overview

• Generating, distributing, and storing encryption keys


• Passwords and access control information maintained and distributed
• Monitoring and controlling access
• Logging
 Collection, storage, and examination of audit records and security logs
General Requirements of NMS

• Scalable
– must work well for managing 10 devices, or 10 million devices
• Inter-operable
– Must inter-work with devices and management systems from multiple
vendors - conform to standards
• Inter-technology capable
– Must work for new and old devices
• Flexible
– Able to be adapted to provide for future needs
• Cost-effective
– Meet certain performance requirements
• Simple, configurable interface
– Usually a GUI

RMIT University©2017 Network Management and Software Defined Networks 43


General Requirements of NMS

• Complete Solutions
– Operators want to be able to buy a complete solution to their management
needs in the one product
– Must be flexible to adapt to changing, future needs

RMIT University©2017 Network Management and Software Defined Networks 44


Where is network management going ?

• Network evolution have lead to the big trend in network management:


– The move to integrated, distributed, open, object-oriented management
solutions
• Why integrated ?
– The network is integrated in terms of services
– The network is heterogeneous in terms of telecommunications technologies
– The management system has to work with the whole network to provide
integrated services
• Why distributed ?
– Telcos need to inter-work with other telcos up the road or around the corner
– Distributed systems can be designed for scalability more easily than
centralised systems

RMIT University©2017 Network Management and Software Defined Networks 45


Where is network management going ?

• Why open?
– No single supplier can give all the telcos worldwide what they need in the
next decade
– Telcos have often been ‘burnt’ in the past by proprietary systems
– We don’t want another Microsoft!
– Telcos systems are going to need to interact with each other

• Why object oriented ?


– Allows for re-use of basic functions
– Prevents today’s systems from becoming tomorrow’s legacy systems
– Better to build functional blocks than isolated systems
– Object-oriented analysis helps keep focus on the core business

RMIT University©2017 Network Management and Software Defined Networks 46