View
260
Download
3
Category
Preview:
Citation preview
1 April 11, 2023]Rushin $hah
April 11, 2023]Rushin $hah2
Localization & Management of Sensor Networks
Unit - 4
WSN
3
Localization & Management of
Sensor Networks
April 11, 2023]Rushin $hah
» Localization in Sensor Network»Network Management Requirement»Network Management Model
»Design Issue» Energy Harvesting in Sensor N/W
April 11, 2023]Rushin $hah4
Network Management for Wireless Sensor NetworksUnit 9 of book
April 11, 20235
Network Management Requirement
]Rushin $hah
A computer communication network generally consists of three
components:
Physical devices
Links (wireless or wired link),
Network nodes (hub, bridge, switch, or router), and
Terminals and Servers;
Protocol; and
Information that is being carried, including applications.
However, the physical devices and protocols are not sufficient to
support effective operation of a communications network.
Network Management (NM) tools and techniques are also
required to help provision network services & ensure co-
operation of entities in the network.
April 11, 20236
Reasons for Management Function
]Rushin $hah
The reasons for management functions are mainfold and may be summarized as follows: There are many heterogeneous devices and software
entities that comprise the network, and some may fail. It is the NM responsibility:
- To determine when, where, and why the fault had occurred
- & how to restore these entities. ‘Optimization of system performance’ as a distributed
system require NM to collaborate in the process. For example, in some networks,
Congestion Control through admission control, by changing routes or through device upgrade
Occurs/done by NM functions.
April 11, 20237
Reasons for Management Function
]Rushin $hah
Cont…
For most networks, NM functions
can be used to gather and analyze the behaviour of
user interaction during network interface,
which is very important in planning the long-term
evolution of network capacity and its performance.
April 11, 2023]Rushin $hah8
NM : Set of Functions
Network management consist of a set of
functions:
Continuous monitoring of Network status,
Detection of Network faults and abnormalities,
Manage, Control, and Help configured Network
components,
Maintain normal operation, and
Improve Network efficiency and application performance.
April 11, 2023]Rushin $hah9
Why NM, considered as an Application? To perform the previous tasks, NM needs to
collect real-time information in network devices,
analyze the information, and
apply control based on the information.
Information is often organized as a Management Information Base
(MIB) in each network device.
Usually, there is an Agent in each device
To collect the information and
Report to a network management centre.
Therefore,
“Network Management can be considered as an application.”
April 11, 202310
Simple Network Management Protocol
]Rushin $hah
For managing networks, SNMP is broadly use today.
It includes three components: Network Management System (NMS), Managed Elements, and Agents.
Role of NMS : NMS is a set of applications that monitor and/or control
managed elements. NMS can request management information/attributes
from the agent. NMS present the results to NM users in figures/tables
form. NMS can also set attributes within the agent.
SN
MP
April 11, 2023]Rushin $hah11
Simple Network Management Protocol Role of Managed Element :
“ The managed elements are the network devices that are
managed ”.
SNMP agents run on each managed element.
The managed elements:
Collect & Store management information in the MIB and
Provide access through SNMP to the MIBs.
Examples: of managed elements include: Routers, Switches, Servers, and Hosts.
SN
MP
SwRo SeHo
April 11, 2023]Rushin $hah12
Simple Network Management Protocol Role of Agents :
SNMP agents are management software modules that reside
on managed elements.
The Agents:
Collect and Store the state of the managed elements
Translate this information into a form compatible with
SNMP MIB.
Exchanges of network management information are through
messages called protocol data units (PDUs).
PDU are sent to nodes and
contain variables that have both attributes and values.
SN
MP
April 11, 2023]Rushin $hah13
Simple Network Management Protocol
The SNMP defines five types of messages or PDUs:
Two deal with the reading terminal,
Another two handle terminal configuration, and
The fifth is Trap, used to monitor events in the managed
elements.
Each PDU contains both Attributes and Values.
Importance of PDU:
NM information can be exchanged through the PDUs in
order to monitor the managed elements.
SN
MP
April 11, 2023]Rushin $hah14
Simple Network Management Protocol
Advantages of SNMP:
Its very simple and widely deployment.
In SNMP version 3 it can obtain more information by a
pair of PDUs such as (GetBulkRequest and GetResponse).
SN
MP
April 11, 2023]Rushin $hah15
Simple Network Management Protocol
Disadvantages of SNMP :
It consumes considerable bandwidth since it often gets
only one piece of management information at a time:
GetRequest (GetNextRequest) and GetResponse.
Due to the usually large number of managed elements,
large bandwidth consumption still exists.
It only manages network elements; it does not support
network-level management.
SN
MP
April 11, 2023]Rushin $hah16
Telecom Operation Map It is proposed by TeleManagement Forum.
It is based on the
Service management Network management
process models.
TOM presents a model for telecommunications management
for network and service management and a view of
‘‘operations.’’
IDEA:
To introduce processes comprising operations and their
automation.
TOM only provides a framework for service management.
TO
M
April 11, 2023]Rushin $hah17
Telecom Operation Map : Levels / Layers
Vertical Layers for service management:
Network and Systems Management,
Service Development and Operations, and
Customer Care Process.
Horizontally Layers for service management
Service Fulfilment,
Service Assurance, and
Service Billing.
TO
M
April 11, 2023]Rushin $hah18
Funda
Neither SNMP nor TOM is designed particularly for
wireless sensor networks.
However, by utilizing
The Simplicity of SNMP and
The Layered Framework of TOM
Design of effective & efficient n/w management
architecture for Wireless Sensor Networks is possible.
SN
MP
& T
OM
April 11, 2023]Rushin $hah19
Network Management Design Issues
Requirement of NM in WSN:
WSN is a special type of wireless network:
possibly with ad hoc structure and
probably with limited resources.
Due to these WSN constraints: networking protocols, the
application model, middleware and sensor node OS
(operating systems) should be designed very carefully.
So here, Network management for WSNs is required to
use those limited resources effectively & efficiently.
20
Network Management Design Issues
April 11, 2023]Rushin $hah
Importance of NM in WSN: is for following reasons: (Design
Issue)
1. In order to deploy an Adaptive and Resource-Efficient algorithm in
WSNs, the current resource level needs to be gathered through
network management.
- For example: the power availability should be known before
switching a sensor node from active (or sleep) mode to sleep (or
active) mode.
- Most traditional networks do not have these requirements.
2. Collaboration and Cooperation between sensor nodes are required
to optimize system performance.
- Network management is an effective tool to provide the platform
required for this purpose.
The number are issue
April 11, 2023]Rushin $hah21
Network Management Design Issues3. Most WSN applications need to know the coverage area
so that they ensure that the entire space is being
monitored.
- Topology management can be used in case an
uncovered area is detected.
- Approaches to increasing the coverage area:
I. Increase the node’s radio power,
II. Increase the density of deployment of senor nodes,
and
III. Move the sensor nodes around to achieve equal
distribution.
4. Nodes in WSNs are usually arranged in an ad hoc
manner. - The parameters of this ad hoc network are obtained by the network management system.
The number are issue
April 11, 2023]Rushin $hah22
An issue is whether in the meanwhile, any of the existing network
management solutions (e.g., SNMP, TOM) can be used for WSNs.
SNMP is often used to manage network elements such as switches
and routers.
It uses GetResponse and GetResponse PDUs to collect information
from network elements.
In SNMP, a local management agent should run in each managed
element.
The local agent is a static and passive agent that receives
commands from a manager and returns the corresponding
response.
For Reference
April 11, 2023]Rushin $hah23
It can also issue Trap messages to the manager when the managed
element encounters a preconfigured event.
Agents in different network elements are independent, and there is
not collaboration among them.
TOM is a new operation and management model that provides a
layered architecture for management and administration.
Each layer has a different management function and set of managed
objects.
TOM can be used to manage most tasks, from the underlying physical
network element to the entire network, as well as the services
provided.
For Reference
April 11, 2023]Rushin $hah24
SNMP provides five management functions: fault management,
configuration management, accounting management,
performance management, and security management; and
TOM, the management functions are layered in network element
management, network management, and service management.
In each layer, different management functions are embodied.
WSNs need some of these management functions.
Therefore, WSNs need layered management architecture with
different management functions in each layer.
For Reference
April 11, 2023]Rushin $hah25
The issue of management architecture for WSNs should also be considered
carefully.
A network management platform consists of three major components:
manager, agent, and MIB.
The manager is used to manage and control the entire network and works
as an interface to other systems.
The agent is located in managed elements.
MIB is an object-oriented structured tree that informs the manager and
agent about the organization of management information.
A standardized MIB guarantees that the management products from
different vendors interconnect.
The manager receives management information and commands the
managed elements using a SNMP-like method or mobile-agent-based
entities.
For Reference
April 11, 2023]Rushin $hah26
The method of accessing management information and the
placement of the manager or agent usually determines the
management architecture.
The agent-based method can save bandwidth since it can
report only final management information.
Although WSNs have a centralized data collecting point
(sink), they are more like distributed networks.
As a result, agent-based hybrid management architectures
might be more suitable for WSNs.
For Reference
April 11, 2023]Rushin $hah27
In WSNs, management information can be used to improve
system performance.
For example, if the network management system detects a
dysfunctional sensor node, it can command another sensor
node to take over.
So the issue of integration of network management with
the functions of network protocols and algorithms becomes
critical.
For Reference
April 11, 2023]Rushin $hah28
Which factors should be consider while designing a
Network Management Protocol ?
29
Special Features of WSNs in NM function.
April 11, 2023]Rushin $hah
Management solutions should be energy efficient, using as little
wireless bandwidth as possible since communication is highly energy
demanding.
Management solutions should be scalable. This is especially important
since in future WSNs may consist of tens to thousands of nodes.
Management solutions should be simple & practical since WSNs are
resource-constrained distributed systems.
MIB for WSNs should contain a general information model for sensor
nodes, features of WSNs and WSN applications.
Management solutions for WSNs should provide a general interface
to the applications since applications can perform better when able to
access management information.
Management solutions should be implementable as middleware.
April 11, 202330
MANNA : Management Architecture
]Rushin $hah
An optimization problem for monitoring, management
provides the monitoring regions given that the ‘battery and
energy consumption rate for each sensor are known
beforehand’.
Mo
nit
ori
ng
Man
ag
em
en
t
MANNA is a management
architecture for WSNs
proposed by Ruizetal.
The architecture considers
three management
dimensions:
Management Function
Areas,
Management Levels, and
WSN Functionalities
April 11, 2023]Rushin $hah31
MANNA : Management dimensionsManagement Function Areas Management Levels
Similar to those in SNMP
Fault mng,
Configuration mng,
Performance mng,
Security mng, and
Accounting Management
NOTE: Configuration management has a notably more important
role in MANNA, where all other functions depend on it.
Similar to those in TOM
Network Element,
Network Element
Management,
Network Management,
Service Management, and
Business Management.
FCPAS
Five types of traditional management functions similar to SNMP:
April 11, 2023]Rushin $hah32
MANNA : Management dimensions
A number of other functions are proposed by MANNA:
Configuration,
Maintenance,
Sensing,
Processing, and
Communication.
With the Aim of promoting productivity and integrating the
functions of configuration, operation, administration and
maintenance of all elements and services in a WSN
- MANNA architecture includes three architectural elements:
Functional, Physical, and Informational architectures.Mo
nit
ori
ng
Man
ag
em
en
t
ProCoMaSeComm
PLZ!
April 11, 2023]Rushin $hah33
MANNA : Architectural Element
Functional architecture:
- It provides functions executed in
the management entities (manager, agent, and MIB) and
the location scheme for managers and agents.
Physical architecture:
- It is where functional architecture is implemented.
- MANNA uses a lightweight protocol as a communication
interface between management entities.
Information architecture:
- The Architecture element provides an object-oriented model
for mapping manageable resources and supporting object
classes.
April 11, 2023]Rushin $hah34
MANNA : Managed Object Classes
Network (information on network behaviour and features
such as data delivery model, network structure, and
mobility),
Managed elements (such as sensor nodes),
Equipment (the physical components of sensor nodes),
System (information on operating system),
Environment (the environment the WSN is running),
Phenomenon, and
Connection.
April 11, 2023]Rushin $hah35
MANNA : Common Management Functions
An Environment monitoring functions,
A coverage area supervision function,
A topology map discovery function,
An energy-level discovery function,
An energy map generation function, and several others.
April 11, 2023]Rushin $hah36
MANNA : dynamic MIB model
MANNA also provides a dynamic MIB model for WSNs:
A sensing coverage area map,
A communication coverage area map,
A WSN behaviour model,
A node dependence model,
network topology,
residual energy, and so on.
In MANNA, the management functions have the lowest
granularity and can be combined into management
services.
-1st sense then communication coverage area map.-Then create two model : behaviour & node dependence-Now define network topology and residual energy
April 11, 2023]Rushin $hah37
Issues Related to Network Management The most important issues in Network
management areI. Naming,II. Localization,III. Maintenance, andIV. Fault tolerance.
April 11, 2023]Rushin $hah38
Meaning of Each IssueД Naming is the scheme used to identify a sensor node.
-An efficient naming scheme can lower computation overhead and
make routing protocol energy efficient.
Д Localization schemes determine the location of sensor nodes since
such information is important for some sensor applications.
Д The maintenance issue may involve actions such as replacing
batteries, keeping connectivity, and configuring sensor nodes.
-The maintenance activity is used to maintain normal operation of
the entire network for as long as possible.
Д Several factors can cause faults in network operation, including
hardware and software error. Therefore, different schemes must be
implemented to provide fault tolerance.
April 11, 202339
Naming
]Rushin $hah
A node in a networked system is identified through
Naming.
This identifier is then used for communication between
nodes.
Approaches to Naming: Low-level naming such as node addresses is typically
application independent but topology and location dependent.
High-level naming is usually application dependent and location independent.
High-level Naming is built on the top of Low-level
Naming.
Communication between applications uses high-level
naming only, whereas physical communication relies on
low-level naming.
Therefore, a binding mechanism is required to realize
mapping bbetween high- and low-level naming.
April 11, 202340
Naming
]Rushin $hah
For example: the domain name system (DNS) in the
Internet uses two types of naming: a domain name and an
IP address.
The domain name is used by applications such as
Internet browsers.
The IP address is used by routing protocol to guarantee
packet forwarding.
Domain name and IP addresses are often directed to the
same host. The DNS servers map between domain
names and IP addresses.
When a Web site is accessed using a domain name, the
application program requests a corresponding IP address
from the DNS so as to set up low-level communications.
April 11, 202341
Naming : traditional hierarchical naming advantages
]Rushin $hah
Although the traditional hierarchical naming approaches
can be used for WSN, but not efficient compared with
application oriented low-level naming, which has the
following advantages: It avoids the overhead resulting from mapping between
high- and low-level naming. This feature is attractive for a sensor since it has limited resources.
Location-dependent addressing is not required. Since the topology of WSNs is highly variable due to node mobility, node life span, and wireless channel quality, a location-dependent address would cause additional problems.
It enables application-specific processing in the network, such as data compression and data fusion, which in turn reduces data transmission.
April 11, 2023]Rushin $hah42
Sensor nodes are usually classified by the type of data they
gather.
For sensor nodes that gather only one type of data or can
have differing personalities and gather multiple types of
data, one name as their identifier would be sufficient.
The objective of low-level application naming is to realize
energy efficiency and fault tolerance in a variety of
environments.
April 11, 202343
Localization
]Rushin $hah
Sensor nodes are distributed all over the place
for sensing and data collection.
It is usually helpful if the locations of sensor
nodes are also known.
April 11, 2023]Rushin $hah44
Localization : Advantages of Location Knowledge
Some applications, such as those for tracking of objects, are
highly location dependent;
Location-based Routing, which may also result in energy
conservation is enabled;
Knowledge of location usually enhances security;
Locations are helpful for sensor network management and
monitoring;
Locations stimulate the creation of new applications;
Sensor nodes that move can be controlled through knowledge of
their location; and
For applications with low-level naming and/or data-centric WSNs,
knowledge of location information is absolutely necessary.
April 11, 2023]Rushin $hah45
Localization : Classification
Localization
Algorithm
Centralized
Schemes
Distributed
Schemes
Range-based
Scheme
Range-free
Schemes
April 11, 2023]Rushin $hah46
Localization : Classification
Centralized Scheme
In this scheme Sensor nodes send control messages to a
central node whose location is known.
The central node then computes the location of every sensor
node and informs the nodes of their locations.
Distributed Scheme
Each sensor node determines its own location independently.
The distributed localization can be further grouped into:
Range-based schemes and
Range-free schemes.
April 11, 2023]Rushin $hah47
Localization : Classification
In the range-based approach, some range information, such
as time of arrival, angle of arrival, or time difference of
arrival is required.
The range-free algorithms works as follows:
Several seed nodes are distributed in WSNs.
Seed nodes know their own locations, and they
periodically broadcast a control message with their
location information.
Sensor nodes that receive these control messages can
then estimate their own locations.
April 11, 2023]Rushin $hah48
Last PPT of S
IR
Lecture -38
Localizatio
n
April 11, 2023]Rushin $hah49
Introduction
The development of large scale distributed sensor
system is a significant scientific and engineering
challenge.
By placing sensors close to each other can increase
signal quality at reduce cost.
April 11, 2023]Rushin $hah50
Taxonomy of localization
Commonly localization is divided in two major
categories
1. Active Localization
2. Passive Localization
April 11, 2023]Rushin $hah51
Active Localization
Active localization technique emit signals into the
environment that are used to measure range to the
target.
These signals may be emitted by infrastructure
components or by targets.
Within this category of active localization there are
three subcategories.
1. Non-cooperative target
2. Co-operative target
3. Co-operative infrastructure
April 11, 2023]Rushin $hah52
Non Co-Operative Target
In active Non Co-operative system, system elements
(Sensors) emit ranging signals, which are distorted or
reflected in flight by passive elements.
The system element then receive the signals and
analyze them to deduce their location relative to
passive elements of the environment.
Example includes RADAR System and SONAR systems
in robots.
April 11, 2023]Rushin $hah53
Co-Operative Target
In co-operative target system, targets emit signal with
known characteristics and other element of the
system detects the signal and use information about
signal arrival to deduce the target’s location.
April 11, 2023]Rushin $hah54
Co-Operative Infrastructure
In co-operative infrastructure system, elements of the
infrastructure emit signals that target can receive.
The infrastructure itself is assumed to be carefully
configured & synchronized to simplify the processing
done by the target.
April 11, 2023]Rushin $hah55
Passive Localization
Passive localization techniques differ from active ones in
that they discover ranges and locations by passively
monitoring existing signals in a particular channel.
The term passive does not imply that they emit no signals,
instead of that the signals they emit are out side the
channel that is primarily analyzed for time-of-flight
measurement.
1. Blind source localization
2. Passive Target Localization
3. Passive Self-Localization
April 11, 2023]Rushin $hah56
Blind source Localization
In a blind source localization system, a signal source
is localized without any priori knowledge of the type
of signal emitted.
Typically this is done by “blind beam forming”,
which effectively cross-correlates the signals from
different receivers.
April 11, 2023]Rushin $hah57
Passive Target Localization
Passive target localization system is usually based on
coherent combination of signals, with the added
assumption of some knowledge of the source.
By assuming model for the signals generated by the
source, filtering can be applied to improve the
performance of the algorithm.
April 11, 2023]Rushin $hah58
Passive Self Localization
In passive self-localization , existing beacon signals
from known infrastructure elements are used by
Target to passively deduce its own location.
April 11, 2023]Rushin $hah59
Localization challenges in Multi hop ad-hoc sensor Networks
Physical layer challenges
Algorithm Design Challenge
- Noisy measurement
- Computation and Communication trade-offs
- Problem setup
April 11, 2023]Rushin $hah60
Recommended