💾 Archived View for gmi.noulin.net › rfc › rfc7548.gmi captured on 2023-06-14 at 21:30:45. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2022-01-08)
-=-=-=-=-=-=-
Keywords: Constrained, Management, IoT, M2M
Internet Engineering Task Force (IETF) M. Ersue, Ed.
Request for Comments: 7548 Nokia Networks
Category: Informational D. Romascanu
ISSN: 2070-1721 Avaya
J. Schoenwaelder
A. Sehgal
Jacobs University Bremen
May 2015
Management of Networks with Constrained Devices: Use Cases
Abstract
This document discusses use cases concerning the management of
networks in which constrained devices are involved. A problem
statement, deployment options, and the requirements on the networks
with constrained devices can be found in the companion document on
"Management of Networks with Constrained Devices: Problem Statement
and Requirements" (RFC 7547).
Status of This Memo
This document is not an Internet Standards Track specification; it is
published for informational purposes.
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by the
Internet Engineering Steering Group (IESG). Not all documents
approved by the IESG are a candidate for any level of Internet
Standard; see Section 2 of RFC 5741.
Information about the current status of this document, any errata,
and how to provide feedback on it may be obtained at
http://www.rfc-editor.org/info/rfc7548.
Ersue, et al. Informational [Page 1]
RFC 7548 Constrained Management: Use Cases May 2015
Copyright Notice
Copyright (c) 2015 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction ....................................................3
2. Access Technologies .............................................4
2.1. Constrained Access Technologies ............................4
2.2. Cellular Access Technologies ...............................5
3. Device Life Cycle ...............................................6
3.1. Manufacturing and Initial Testing ..........................6
3.2. Installation and Configuration .............................6
3.3. Operation and Maintenance ..................................7
3.4. Recommissioning and Decommissioning ........................7
4. Use Cases .......................................................8
4.1. Environmental Monitoring ...................................8
4.2. Infrastructure Monitoring ..................................9
4.3. Industrial Applications ...................................10
4.4. Energy Management .........................................12
4.5. Medical Applications ......................................14
4.6. Building Automation .......................................15
4.7. Home Automation ...........................................17
4.8. Transport Applications ....................................18
4.9. Community Network Applications ............................20
4.10. Field Operations .........................................22
5. Security Considerations ........................................23
6. Informative References .........................................24
Acknowledgments ...................................................25
Contributors ......................................................26
Authors' Addresses ................................................26
Ersue, et al. Informational [Page 2]
RFC 7548 Constrained Management: Use Cases May 2015
1. Introduction
Constrained devices (also known as sensors, smart objects, or smart
devices) with limited CPU, memory, and power resources can be
connected to a network. Such a network of constrained devices itself
may be constrained or challenged, e.g., with unreliable or lossy
channels, wireless technologies with limited bandwidth and a dynamic
topology, needing the service of a gateway or proxy to connect to the
Internet. In other scenarios, the constrained devices can be
connected to a unconstrained network using off-the-shelf protocol
stacks. Constrained devices might be in charge of gathering
information in diverse settings including natural ecosystems,
buildings, and factories and sending the information to one or more
server stations.
Network management is characterized by monitoring network status,
detecting faults (and inferring their causes), setting network
parameters, and carrying out actions to remove faults, maintain
normal operation, and improve network efficiency and application
performance. The traditional network management application
periodically collects information from a set of managed network
elements, it processes the collected data, and it presents the
results to the network management users. Constrained devices,
however, often have limited power, have low transmission range, and
might be unreliable. Such unreliability might arise from device
itself (e.g., battery exhausted) or from the channel being
constrained (i.e., low-capacity and high-latency). They might also
need to work in hostile environments with advanced security
requirements or need to be used in harsh environments for a long time
without supervision. Due to such constraints, the management of a
network with constrained devices offers different types of challenges
compared to the management of a traditional IP network.
This document aims to understand use cases for the management of a
network in which constrained devices are involved. It lists and
discusses diverse use cases for management from the network as well
as from the application point of view. The list of discussed use
cases is not an exhaustive one since other scenarios, currently
unknown to the authors, are possible. The application scenarios
discussed aim to show where networks of constrained devices are
expected to be deployed. For each application scenario, we first
briefly describe the characteristics followed by a discussion on how
network management can be provided, who is likely going to be
responsible for it, and on which time-scale management operations are
likely to be carried out.
Ersue, et al. Informational [Page 3]
RFC 7548 Constrained Management: Use Cases May 2015
A problem statement, deployment and management topology options as
well as the requirements on the networks with constrained devices can
be found in the companion document [RFC7547].
This documents builds on the terminology defined in [RFC7228] and
[RFC7547]. [RFC7228] is a base document for the terminology
concerning constrained devices and constrained networks. Some use
cases specific to IPv6 over Low-Power Wireless Personal Area Networks
(6LoWPANs) can be found in [RFC6568].
2. Access Technologies
Besides the management requirements imposed by the different use
cases, the access technologies used by constrained devices can impose
restrictions and requirements upon the Network Management System
(NMS) and protocol of choice.
It is possible that some networks of constrained devices might
utilize traditional unconstrained access technologies for network
access, e.g., local area networks with plenty of capacity. In such
scenarios, the constrainedness of the device presents special
management restrictions and requirements rather than the access
technology utilized.
However, in other situations, constrained or cellular access
technologies might be used for network access, thereby causing
management restrictions and requirements to arise as a result of the
underlying access technologies.
A discussion regarding the impact of cellular and constrained access
technologies is provided in this section since they impose some
special requirements on the management of constrained networks. On
the other hand, fixed-line networks (e.g., power-line communications)
are not discussed here since tend to be quite static and do not
typically impose any special requirements on the management of the
network.
2.1. Constrained Access Technologies
Due to resource restrictions, embedded devices deployed as sensors
and actuators in the various use cases utilize low-power, low-data-
rate wireless access technologies such as [IEEE802.15.4], Digital
Enhanced Cordless Telecommunication (DECT) Ultra Low Energy (ULE), or
Bluetooth Low-Energy (BT-LE) for network connectivity.
In such scenarios, it is important for the NMS to be aware of the
restrictions imposed by these access technologies to efficiently
manage these constrained devices. Specifically, such low-power, low-
Ersue, et al. Informational [Page 4]
RFC 7548 Constrained Management: Use Cases May 2015
data-rate access technologies typically have small frame sizes. So
it would be important for the NMS and management protocol of choice
to craft packets in a way that avoids fragmentation and reassembly of
packets since this can use valuable memory on constrained devices.
Devices using such access technologies might operate via a gateway
that translates between these access technologies and more
traditional Internet protocols. A hierarchical approach to device
management in such a situation might be useful, wherein the gateway
device is in-charge of devices connected to it, while the NMS
conducts management operations only to the gateway.
2.2. Cellular Access Technologies
Machine-to-machine (M2M) services are increasingly provided by mobile
service providers as numerous devices, home appliances, utility
meters, cars, video surveillance cameras, and health monitors are
connected with mobile broadband technologies. Different
applications, e.g., in a home appliance or in-car network, use
Bluetooth, Wi-Fi, or ZigBee locally and connect to a cellular module
acting as a gateway between the constrained environment and the
mobile cellular network.
Such a gateway might provide different options for the connectivity
of mobile networks and constrained devices:
o a smartphone with 3G/4G and WLAN radio might use BT-LE to connect
to the devices in a home area network,
o a femtocell might be combined with home gateway functionality
acting as a low-power cellular base station connecting smart
devices to the application server of a mobile service provider,
o an embedded cellular module with LTE radio connecting the devices
in the car network with the server running the telematics service,
o an M2M gateway connected to the mobile operator network supporting
diverse Internet of Things (IoT) connectivity technologies
including ZigBee and Constrained Application Protocol (CoAP) over
6LoWPAN over IEEE 802.15.4.
Common to all scenarios above is that they are embedded in a service
and connected to a network provided by a mobile service provider.
Usually, there is a hierarchical deployment and management topology
in place where different parts of the network are managed by
different management entities and the count of devices to manage is
high (e.g., many thousands). In general, the network is comprised of
manifold types and sizes of devices matching to different device
Ersue, et al. Informational [Page 5]
RFC 7548 Constrained Management: Use Cases May 2015
classes. As such, the managing entity needs to be prepared to manage
devices with diverse capabilities using different communication or
management protocols. In the case in which the devices are directly
connected to a gateway, they most likely are managed by a management
entity integrated with the gateway, which itself is part of the NMS
run by the mobile operator. Smartphones or embedded modules
connected to a gateway might themselves be in charge of managing the
devices on their level. The initial and subsequent configuration of
such a device is mainly based on self-configuration and is triggered
by the device itself.
The gateway might be in charge of filtering and aggregating the data
received from the device as the information sent by the device might
be mostly redundant.
3. Device Life Cycle
Since constrained devices deployed in a network might go through
multiple phases in their lifetime, it is possible for different
managers of networks and/or devices to exist during different parts
of the device lifetimes. An in-depth discussion regarding the
possible device life cycles can be found in [IOT-SEC].
3.1. Manufacturing and Initial Testing
Typically, the life cycle of a device begins at the manufacturing
stage. During this phase, the manufacturer of the device is
responsible for the management and configuration of the devices. It
is also possible that a certain use case might utilize multiple types
of constrained devices (e.g., temperature sensors, lighting
controllers, etc.) and these could be manufactured by different
entities. As such, during the manufacturing stage, different
managers can exist for different devices. Similarly, during the
initial testing phase, where device quality-assurance tasks might be
performed, the manufacturer remains responsible for the management of
devices and networks that might comprise them.
3.2. Installation and Configuration
The responsibility of managing the devices must be transferred to the
installer during the installation phase. There must exist procedures
for transferring management responsibility between the manufacturer
and installer. The installer may be the customer or an intermediary
contracted to set up the devices and their networks. It is important
that the NMS that is utilized allows devices originating at different
vendors to be managed, ensuring interoperability between them and the
configuration of trust relationships between them as well.
Ersue, et al. Informational [Page 6]
RFC 7548 Constrained Management: Use Cases May 2015
It is possible that the installation and configuration
responsibilities might lie with different entities. For example, the
installer of a device might only be responsible for cabling a
network, physically installing the devices, and ensuring initial
network connectivity between them (e.g., configuring IP addresses).
Following such an installation, the customer or a subcontractor might
actually configure the operation of the device. As such, during
installation and configuration multiple parties might be responsible
for managing a device and appropriate methods must be available to
ensure that this management responsibility is transferred suitably.
3.3. Operation and Maintenance
At the outset of the operation phase, the operational responsibility
of a device and network should be passed on to the customer. It is
possible that the customer, however, might contract the maintenance
of the devices and network to a subcontractor. In this case, the NMS
and management protocol should allow for configuring different levels
of access to the devices. Since different maintenance vendors might
be used for devices that perform different functions (e.g., HVAC,
lighting, etc.), it should also be possible to restrict management
access to devices based on the currently responsible manager.
3.4. Recommissioning and Decommissioning
The owner of a device might choose to replace, repurpose, or even
decommission it. In each of these cases, either the customer or the
contracted maintenance agency must ensure that appropriate steps are
taken to meet the end goal.
In case the devices needs to be replaced, the manager of the network
(customer or contractor responsible) must detach the device from the
network, remove all appropriate configuration, and discard the
device. A new device must then be configured to replace it. The NMS
should allow for the transferring of the configuration and replacing
an existing device. The management responsibility of the operation/
maintenance manager would end once the device is removed from the
network. During the installation of the new replacement device, the
same responsibilities would apply as those during the Installation
and Configuration phases.
The device being replaced may not have yet reached end-of-life, and
as such, instead of being discarded, it may be installed in a new
location. In this case, the management responsibilities are once
again resting in the hands of the entities responsible for the
Installation and Configuration phases at the new location.
Ersue, et al. Informational [Page 7]
RFC 7548 Constrained Management: Use Cases May 2015
If a device is repurposed, then it is possible that the management
responsibility for this device changes as well. For example, a
device might be moved from one building to another. In this case,
the managers responsible for devices and networks in each building
could be different. As such, the NMS must not only allow for
changing configuration but also the transferring of management
responsibilities.
In case a device is decommissioned, the management responsibility
typically ends at that point.
4. Use Cases
4.1. Environmental Monitoring
Environmental monitoring applications are characterized by the
deployment of a number of sensors to monitor emissions, water
quality, or even the movements and habits of wildlife. Other
applications in this category include earthquake or tsunami early-
warning systems. The sensors often span a large geographic area;
they can be mobile; and they are often difficult to replace.
Furthermore, the sensors are usually not protected against tampering.
Management of environmental-monitoring applications is largely
concerned with monitoring whether the system is still functional and
the roll out of new constrained devices in case the system loses too
much of its structure. The constrained devices themselves need to be
able to establish connectivity (autoconfiguration), and they need to
be able to deal with events such as losing neighbors or being moved
to other locations.
Management responsibility typically rests with the organization
running the environmental-monitoring application. Since these
monitoring applications must be designed to tolerate a number of
failures, the time scale for detecting and recording failures is, for
some of these applications, likely measured in hours and repairs
might easily take days. In fact, in some scenarios it might be more
cost- and time-effective not to repair such devices at all. However,
for certain environmental monitoring applications, much tighter time
scales may exist and might be enforced by regulations (e.g.,
monitoring of nuclear radiation).
Since many applications of environmental-monitoring sensors are
likely to be in areas that are important to safety (flood monitoring,
nuclear radiation monitoring, etc.), it is important for management
protocols and NMSs to ensure appropriate security protections. These
protections include not only access control, integrity, and
Ersue, et al. Informational [Page 8]
RFC 7548 Constrained Management: Use Cases May 2015
availability of data, but also provide appropriate mechanisms that
can deal with situations that might be categorized as emergencies or
when tampering with sensors/data might be detected.
4.2. Infrastructure Monitoring
Infrastructure monitoring is concerned with the monitoring of
infrastructures such as bridges, railway tracks, or (offshore)
windmills. The primary goal is usually to detect any events or
changes of the structural conditions that can impact the risk and
safety of the infrastructure being monitored. Another secondary goal
is to schedule repair and maintenance activities in a cost-effective
manner.
The infrastructure to monitor might be in a factory or spread over a
wider area (but difficult to access). As such, the network in use
might be based on a combination of fixed and wireless technologies,
which use robust networking equipment and support reliable
communication via application-layer transactions. It is likely that
constrained devices in such a network are mainly C2 devices [RFC7228]
and have to be controlled centrally by an application running on a
server. In case such a distributed network is widely spread, the
wireless devices might use diverse long-distance wireless
technologies such as Worldwide Interoperability for Microwave Access
(WiMAX) or 3G/LTE. In cases, where an in-building network is
involved, the network can be based on Ethernet or wireless
technologies suitable for in-building use.
The management of infrastructure monitoring applications is primarily
concerned with the monitoring of the functioning of the system.
Infrastructure monitoring devices are typically rolled out and
installed by dedicated experts, and updates are rare since the
infrastructure itself does not change often. However, monitoring
devices are often deployed in unsupervised environments; hence,
special attention must be given to protecting the devices from being
modified.
Management responsibility typically rests with the organization
owning the infrastructure or responsible for its operation. The time
scale for detecting and recording failures is likely measured in
hours and repairs might easily take days. However, certain events
(e.g., natural disasters) may require that status information be
obtained much more quickly and that replacements of failed sensors
can be rolled out quickly (or redundant sensors are activated
quickly). In case the devices are difficult to access, a self-
healing feature on the device might become necessary. Since
infrastructure monitoring is closely related to ensuring safety,
Ersue, et al. Informational [Page 9]
RFC 7548 Constrained Management: Use Cases May 2015
management protocols and systems must provide appropriate security
protections to ensure confidentiality, integrity, and availability of
data.
4.3. Industrial Applications
Industrial Applications and smart manufacturing refer to tasks such
as networked control and monitoring of manufacturing equipment, asset
and situation management, or manufacturing process control. For the
management of a factory, it is becoming essential to implement smart
capabilities. From an engineering standpoint, industrial
applications are intelligent systems enabling rapid manufacturing of
new products, dynamic response to product demands, and real-time
optimization of manufacturing production and supply-chain networks.
Potential industrial applications (e.g., for smart factories and
smart manufacturing) are:
o Digital control systems with embedded, automated process controls;
operator tools; and service information systems optimizing plant
operations and safety.
o Asset management using predictive maintenance tools, statistical
evaluation, and measurements maximizing plant reliability.
o Smart sensors detecting anomalies to avoid abnormal or
catastrophic events.
o Smart systems integrated within the industrial energy-management
system and externally with the smart grid enabling real-time
energy optimization.
Management of Industrial Applications and smart manufacturing may, in
some situations, involve Building Automation tasks such as control of
energy, HVAC, lighting, or access control. Interacting with
management systems from other application areas might be important in
some cases (e.g., environmental monitoring for electric energy
production, energy management for dynamically scaling manufacturing,
vehicular networks for mobile asset tracking). Management of
constrained devices and networks may not only refer to the management
of their network connectivity. Since the capabilities of constrained
devices are limited, it is quite possible that a management system
would even be required to configure, monitor, and operate the primary
functions for which a constrained device is utilized, besides
managing its network connectivity.
Sensor networks are an essential technology used for smart
manufacturing. Measurements, automated controls, plant optimization,
health and safety management, and other functions are provided by a
Ersue, et al. Informational [Page 10]
RFC 7548 Constrained Management: Use Cases May 2015
large number of networked sectors. Data interoperability and
seamless exchange of product, process, and project data are enabled
through interoperable data systems used by collaborating divisions or
business systems. Intelligent automation and learning systems are
vital to smart manufacturing, but they must be effectively integrated
with the decision environment. The NMS utilized must ensure timely
delivery of sensor data to the control unit so it may take
appropriate decisions. Similarly, the relaying of commands must also
be monitored and managed to ensure optimal functioning. Wireless
sensor networks (WSNs) have been developed for machinery Condition-
based Maintenance (CBM) as they offer significant cost savings and
enable new functionalities. Inaccessible locations, rotating
machinery, hazardous areas, and mobile assets can be reached with
wireless sensors. Today, WSNs can provide wireless link reliability,
real-time capabilities, and quality-of-service and they can enable
industrial and related wireless sense and control applications.
Management of industrial and factory applications is largely focused
on monitoring whether the system is still functional, real-time
continuous performance monitoring, and optimization as necessary.
The factory network might be part of a campus network or connected to
the Internet. The constrained devices in such a network need to be
able to establish configuration themselves (autoconfiguration) and
might need to deal with error conditions as much as possible locally.
Access control has to be provided with multi-level administrative
access and security. Support and diagnostics can be provided through
remote monitoring access centralized outside of the factory.
Factory-automation tasks require that continuous monitoring be used
to optimize production. Groups of manufacturing and monitoring
devices could be defined to establish relationships between them. To
ensure timely optimization of processes, commands from the NMS must
arrive at all destination within an appropriate duration. This
duration could change based on the manufacturing task being
performed. Installation and operation of factory networks have
different requirements. During the installation phase, many
networks, usually distributed along different parts of the factory/
assembly line, coexist without a connection to a common backbone. A
specialized installation tool is typically used to configure the
functions of different types of devices, in different factory
locations, in a secure manner. At the end of the installation phase,
interoperability between these stand-alone networks and devices must
be enabled. During the operation phase, these stand-alone networks
are connected to a common backbone so that they may retrieve control
information from and send commands to appropriate devices.
Ersue, et al. Informational [Page 11]
RFC 7548 Constrained Management: Use Cases May 2015
Management responsibility is typically owned by the organization
running the industrial application. Since the monitoring
applications must handle a potentially large number of failures, the
time scale for detecting and recording failures is, for some of these
applications, likely measured in minutes. However, for certain
industrial applications, much tighter time scales may exist, e.g., in
real-time, which might be enforced by the manufacturing process or
the use of critical material. Management protocols and NMSs must
ensure appropriate access control since different users of industrial
control systems will have varying levels of permissions. For
example, while supervisors might be allowed to change production
parameters, they should not be allowed to modify the functional
configuration of devices like a technician should. It is also
important to ensure integrity and availability of data since
malfunctions can potentially become safety issues. This also implies
that management systems must be able to react to situations that may
pose dangers to worker safety.
4.4. Energy Management
The EMAN working group developed an energy-management framework
[RFC7326] for devices and device components within or connected to
communication networks. This document observes that one of the
challenges of energy management is that a power distribution network
is responsible for the supply of energy to various devices and
components, while a separate communication network is typically used
to monitor and control the power distribution network. Devices in
the context of energy management can be monitored for parameters like
power, energy, demand and power quality. If a device contains
batteries, they can be also monitored and managed.
Energy devices differ in complexity and may include basic sensors or
switches, specialized electrical meters, or power distribution units
(PDU), and subsystems inside the network devices (routers, network
switches) or home or industrial appliances. The operators of an
energy-management system are either the utility providers or
customers that aim to control and reduce the energy consumption and
the associated costs. The topology in use differs and the deployment
can cover areas from small surfaces (individual homes) to large
geographical areas. The EMAN requirements document [RFC6988]
discusses the requirements for energy management concerning
monitoring and control functions.
It is assumed that energy management will apply to a large range of
devices of all classes and networks topologies. Specific resource
monitoring, like battery utilization and availability, may be
specific to devices with lower physical resources (device classes C0
or C1 [RFC7228]).
Ersue, et al. Informational [Page 12]
RFC 7548 Constrained Management: Use Cases May 2015
Energy management is especially relevant to the Smart Grid. A Smart
Grid is an electrical grid that uses data networks to gather and act
on energy and power-related information in an automated fashion with
the goal to improve the efficiency, reliability, economics, and
sustainability of the production and distribution of electricity.
Smart Metering is a good example of an energy-management application
based on Smart Grid. Different types of possibly wireless small
meters all together produce a large amount of data, which is
collected by a central entity and processed by an application server,
which may be located within the customer's residence or off site in a
data center. The communication infrastructure can be provided by a
mobile network operator as the meters in urban areas will most likely
have a cellular or WiMAX radio. In case the application server is
located within the residence, such meters are more likely to use
Wi-Fi protocols to interconnect with an existing network.
An Advanced Metering Infrastructure (AMI) network is another example
of the Smart Grid that enables an electric utility to retrieve
frequent electric usage data from each electric meter installed at a
customer's home or business. Unlike Smart Metering, in which case
the customer or their agents install appliance-level meters, an AMI
is typically managed by the utility providers and could also include
other distribution automation devices like transformers and
reclosers. Meters in AMI networks typically contain constrained
devices that connect to mesh networks with a low-bandwidth radio.
Usage data and outage notifications can be sent by these meters to
the utility's headend systems, via aggregation points of higher-end
router devices that bridge the constrained network to a less
constrained network via cellular, WiMAX, or Ethernet. Unlike meters,
these higher-end devices might be installed on utility poles owned
and operated by a separate entity.
It thereby becomes important for a management application not only to
be able to work with diverse types of devices, but also to work over
multiple links that might be operated and managed by separate
entities, each having divergent policies for their own devices and
network segments. During management operations, like firmware
updates, it is important that the management systems perform robustly
in order to avoid accidental outages of critical power systems that
could be part of AMI networks. In fact, since AMI networks must also
report on outages, the management system might have to manage the
energy properties of battery-operated AMI devices themselves as well.
A management system for home-based Smart Metering solutions is likely
to have devices laid out in a simple topology. However, AMI network
installations could have thousands of nodes per router, i.e., higher-
end device, which organize themselves in an ad hoc manner. As such,
Ersue, et al. Informational [Page 13]
RFC 7548 Constrained Management: Use Cases May 2015
a management system for AMI networks will need to discover and
operate over complex topologies as well. In some situations, it is
possible that the management system might also have to set up and
manage the topology of nodes, especially critical routers.
Encryption-key management and sharing in both types of networks are
also likely to be important for providing confidentiality for all
data traffic. In AMI networks, the key may be obtained by a meter
only after an end-to-end authentication process based on
certificates. The Smart Metering solution could adopt a similar
approach or the security may be implied due to the encrypted Wi-Fi
networks they become part of.
The management of such a network requires end-to-end management of
and information exchange through different types of networks.
However, as of today, there is no integrated energy-management
approach and no common information model available. Specific energy-
management applications or network islands use their own management
mechanisms.
4.5. Medical Applications
Constrained devices can be seen as an enabling technology for
advanced and possibly remote health-monitoring and emergency-
notification systems, ranging from monitors for blood pressure and
heart rate to advanced devices capable of monitoring implanted
technologies, such as pacemakers or advanced hearing aids. Medical
sensors may not only be attached to human bodies, they might also
exist in the infrastructure used by humans such as bathrooms or
kitchens. Medical applications will also be used to ensure
treatments are being applied properly, and they might guide people
losing orientation. Fitness and wellness applications, such as
connected scales or wearable heart monitors, encourage consumers to
exercise and empower self-monitoring of key fitness indicators.
Different applications use Bluetooth, Wi-Fi, or ZigBee connections to
access the patient's smartphone or home cellular connection to access
the Internet.
Constrained devices that are part of medical applications are managed
either by the users of those devices or by an organization providing
medical (monitoring) services for physicians. In the first case,
management must be automatic and/or easy to install and set up by
laypeople. In the second case, it can be expected that devices will
be controlled by specially trained people. In both cases, however,
it is crucial to protect the safety and privacy of the people who use
medical devices. Security precautions to protect access
(authentication, encryption, integrity protections, etc.) to such
devices may be critical to safeguarding the individual. The level of
access granted to different users also may need to be regulated. For
Ersue, et al. Informational [Page 14]
RFC 7548 Constrained Management: Use Cases May 2015
example, an authorized surgeon or doctor must be allowed to configure
all necessary options on the devices; however, a nurse or technician
may only be allowed to retrieve data that can assist in diagnosis.
Even though the data collected by a heart monitor might be protected,
the pure fact that someone carries such a device may need protection.
As such, certain medical appliances may not want to participate in
discovery and self-configuration protocols in order to remain
invisible.
Many medical devices are likely to be used (and relied upon) to
provide data to physicians in critical situations in which the
patient might not be able to report such data themselves. Timely
delivery of data can be quite important in certain applications like
patient-mobility monitoring in nursing homes. Data must reach the
physician and/or emergency services within specified limits of time
in order to be useful. As such, fault detection of the communication
network or the constrained devices becomes a crucial function of the
management system that must be carried out with high reliability and,
depending on the medical appliance and its application, within
seconds.
4.6. Building Automation
Building automation comprises the distributed systems designed and
deployed to monitor and control the mechanical, electrical, and
electronic systems inside buildings with various destinations (e.g.,
public and private, industrial, institutions, or residential).
Advanced Building Automation Systems (BASs) may be deployed
concentrating the various functions of safety, environmental control,
occupancy, and security. Increasingly, the deployment of the various
functional systems is connected to the same communication
infrastructure (possibly IP-based), which may involve wired or
wireless communication networks inside the building.
Building automation requires the deployment of a large number (10 to
100,000) of sensors that monitor the status of devices, parameters
inside the building, and controllers with different specialized
functionality for areas within the building or the totality of the
building. Inter-node distances between neighboring nodes vary from 1
to 20 meters. The NMS must, as a result, be able to manage and
monitor a large number of devices, which may be organized in multi-
hop meshed networks. Distances between the nodes, and the use of
constrained protocols, means that networks of nodes might be
segmented. The management of such network segments and nodes in
these segments should be possible. Contrary to home automation, in
building management the devices are expected to be managed assets and
known to a set of commissioning tools and a data storage, such that
every connected device has a known origin. This requires the
Ersue, et al. Informational [Page 15]
RFC 7548 Constrained Management: Use Cases May 2015
management system to be able to discover devices on the network and
ensure that the expected list of devices is currently matched.
Management here includes verifying the presence of the expected
devices and detecting the presence of unwanted devices.
Examples of functions performed by controllers in building automation
are regulating the quality, humidity, and temperature of the air
inside the building as well as regulating the lighting. Other
systems may report the status of the machinery inside the building
like elevators or inside the rooms like projectors in meeting rooms.
Security cameras and sensors may be deployed and operated on separate
dedicated infrastructures connected to the common backbone. The
deployment area of a BAS is typically inside one building (or part of
it) or several buildings geographically grouped in a campus. A
building network can be composed of network segments, where a network
segment covers a floor, an area on the floor, or a given
functionality (e.g., security cameras). It is possible that the
management tasks of different types of some devices might be
separated from others (e.g, security cameras might operate and be
managed via a network separate from that of the HVAC in a building).
Some of the sensors in BASs (for example, fire alarms or security
systems) register, record, and transfer critical alarm information;
therefore, they must be resilient to events like loss of power or
security attacks. A management system must be able to deal with
unintentional segmentation of networks due to power loss or channel
unavailability. It must also be able to detect security events. Due
to specific operating conditions required from certain devices, there
might be a need to certify components and subsystems operating in
such constrained conditions based on specific requirements. Also, in
some environments, the malfunctioning of a control system (like
temperature control) needs to be reported in the shortest possible
time. Complex control systems can misbehave, and their critical
status reporting and safety algorithms need to be basic and robust
and perform even in critical conditions. Providing this monitoring,
configuration and notification service is an important task of the
management system used in building automation.
In some cases, building automation solutions are deployed in newly
designed buildings; in other cases, it might be over existing
infrastructures. In the first case, there is a broader range of
possible solutions, which can be planned for the infrastructure of
the building. In the second case, the solution needs to be deployed
over an existing infrastructure taking into account factors like
existing wiring, distance limitations, and the propagation of radio
signals over walls and floors, thereby making deployment difficult.
As a result, some of the existing WLAN solutions (e.g., [IEEE802.11]
or [IEEE802.15]) may be deployed. In mission-critical or security-
Ersue, et al. Informational [Page 16]
RFC 7548 Constrained Management: Use Cases May 2015
sensitive environments and in cases where link failures happen often,
topologies that allow for reconfiguration of the network and
connection continuity may be required. Some of the sensors deployed
in building automation may be very simple constrained devices for
which C0 or C1 [RFC7228] may be assumed.
For lighting applications, groups of lights must be defined and
managed. Commands to a group of light must arrive within 200 ms at
all destinations. The installation and operation of a building
network has different requirements. During the installation, many
stand-alone networks of a few to 100 nodes coexist without a
connection to the backbone. During this phase, the nodes are
identified with a network identifier related to their physical
location. Devices are accessed from an installation tool to connect
them to the network in a secure fashion. During installation, the
setting of parameters of common values to enable interoperability may
be required. During operation, the networks are connected to the
backbone while maintaining the network identifier to physical
location relation. Network parameters like address and name are
stored in the DNS. The names can assist in determining the physical
location of the device.
It is also important for a building automation NMS to take safety and
security into account. Ensuring privacy and confidentiality of data,
such that unauthorized parties do not get access to it, is likely to
be important since users' individual behaviors could be potentially
understood via their settings. Appropriate security considerations
for authorization and access control to the NMS is also important
since different users are likely to have varied levels of operational
permissions in the system. For example, while end users should be
able to control lighting systems, HVAC systems, etc., only qualified
technicians should be able to configure parameters that change the
fundamental operation of a device. It is also important for devices
and the NMS to be able to detect and report any tampering they might
find, since these could lead to potential user safety concerns, e.g.,
if sensors controlling air quality are tampered with such that the
levels of carbon monoxide become life threatening. This implies that
an NMS should also be able to deal with and appropriately prioritize
situations that might potentially lead to safety concerns.
4.7. Home Automation
Home automation includes the control of lighting, heating,
ventilation, air conditioning, appliances, entertainment and home
security devices to improve convenience, comfort, energy efficiency,
and safety. It can be seen as a residential extension of building
automation. However, unlike a BAS, the infrastructure in a home is
operated in a considerably more ad hoc manner. While in some
Ersue, et al. Informational [Page 17]
RFC 7548 Constrained Management: Use Cases May 2015
installations it is likely that there is no centralized management
system akin to a BAS available, in other situations outsourced and
cloud-based systems responsible for managing devices in the home
might be used.
Home-automation networks need a certain amount of configuration
(associating switches or sensors to actuators) that is either
provided by electricians deploying home-automation solutions, by
third-party home-automation service providers (e.g., small
specialized companies or home-automation device manufacturers) or by
residents by using the application user interface provided by home-
automation devices to configure (parts of) the home-automation
solution. Similarly, failures may be reported via suitable
interfaces to residents or they might be recorded and made available
to services providers in charge of the maintenance of the home-
automation infrastructure.
The management responsibility either lies with the residents or is
outsourced to electricians and/or third parties providing management
of home-automation solutions as a service. A varying combination of
electricians, service providers, or the residents may be responsible
for different aspects of managing the infrastructure. The time scale
for failure detection and resolution is, in many cases, likely
counted in hours to days.
4.8. Transport Applications
"Transport application" is a generic term for the integrated
application of communications, control, and information processing in
a transportation system. "Transport telematics" and "vehicle
telematics" are both used as terms for the group of technologies that
support transportation systems. Transport applications running on
such a transportation system cover all modes of the transport and
consider all elements of the transportation system, i.e. the vehicle,
the infrastructure, and the driver or user, interacting together
dynamically. Examples for transport applications are inter- and
intra-vehicular communication, smart traffic control, smart parking,
electronic toll-collection systems, logistic and fleet management,
vehicle control, and safety and roadside assistance.
As a distributed system, transport applications require an end-to-end
management of different types of networks. It is likely that
constrained devices in a network (e.g., a moving in-car network) have
to be controlled by an application running on an application server
in the network of a service provider. Such a highly distributed
network including cellular devices on vehicles is assumed to include
a wireless access network using diverse long-distance wireless
technologies such as WiMAX, 3G/LTE, or satellite communication, e.g.,
Ersue, et al. Informational [Page 18]
RFC 7548 Constrained Management: Use Cases May 2015
based on an embedded hardware module. As a result, the management of
constrained devices in the transport system might be necessary to
plan top-down and might need to use data models obliged from and
defined on the application layer. The assumed device classes in use
are mainly C2 [RFC7228] devices. In cases, where an in-vehicle
network is involved, C1 devices [RFC7228] with limited capabilities
and a short-distance constrained radio network, e.g., IEEE 802.15.4
might be used additionally.
All Transport Applications will require an IT infrastructure to run
on top of, e.g., in public-transport scenarios like trains, buses, or
metro networks infrastructure might be provided, maintained, and
operated by third parties like mobile-network or satellite-network
operators. However, the management responsibility of the transport
application typically rests within the organization running the
transport application (in the public-transport scenario, this would
typically be the public-transport operator). Different aspects of
the infrastructure might also be managed by different entities. For
example, the in-car devices are likely to be installed and managed by
the manufacturer, while the local government or transportation
authority might be responsible for the on-road vehicular
communication infrastructure used by these devices. The backend
infrastructure is also likely to be maintained by third-party
operators. As such, the NMS must be able to deal with different
network segments (each being operated and controlled by separate
entities) and enable appropriate access control and security.
Depending on the type of application domain (vehicular or stationary)
and service being provided, it is important for the NMS to be able to
function with different architectures, since different manufacturers
might have their own proprietary systems relying on a specific
management topology option, as described in [RFC7547]. Moreover,
constituents of the network can either be private, belong to
individuals or private companies, or be owned by public institutions
leading to different legal and organization requirements. Across the
entire infrastructure, a variety of constrained devices is likely to
be used, and they must be individually managed. The NMS must be able
to either work directly with different types of devices or have the
ability to interoperate with multiple different systems.
The challenges in the management of vehicles in a mobile-transport
application are manifold. The up-to-date position of each node in
the network should be reported to the corresponding management
entities, since the nodes could be moving within or roaming between
different networks. Secondly, a variety of troubleshooting
information, including sensitive location information, needs to be
reported to the management system in order to provide accurate
service to the customer. Management systems dealing with mobile
Ersue, et al. Informational [Page 19]
RFC 7548 Constrained Management: Use Cases May 2015
nodes could possibly exploit specific patterns in the mobility of the
nodes. These patterns emerge due to repetitive vehicular usage in
scenarios like people commuting to work and supply vehicles
transporting shipments between warehouses, etc. The NMS must also be
able to handle partitioned networks, which would arise due to the
dynamic nature of traffic resulting in large inter-vehicle gaps in
sparsely populated scenarios. Since mobile nodes might roam in
remote networks, the NMS should be able to provide operating
configuration updates regardless of node location.
The constrained devices in a moving transport network might be
initially configured in a factory, and a reconfiguration might be
needed only rarely. New devices might be integrated in an ad hoc
manner based on self-management and self-configuration capabilities.
Monitoring and data exchange might be necessary via a gateway entity
connected to the backend transport infrastructure. The devices and
entities in the transport infrastructure need to be monitored more
frequently and may be able to communicate with a higher data rate.
The connectivity of such entities does not necessarily need to be
wireless. The time scale for detecting and recording failures in a
moving transport network is likely measured in hours, and repairs
might easily take days. It is likely that a self-healing feature
would be used locally. On the other hand, failures in fixed
transport-application infrastructure (e.g., traffic lights, digital-
signage displays) are likely to be measured in minutes so as to avoid
untoward traffic incidents. As such, the NMS must be able to deal
with differing timeliness requirements based on the type of devices.
Since transport applications of the constrained devices and networks
deal with automotive vehicles, malfunctions and misuse can
potentially lead to safety concerns as well. As such, besides access
control, privacy of user data, and timeliness, management systems
should also be able to detect situations that are potentially
hazardous to safety. Some of these situations could be automatically
mitigated, e.g., traffic lights with incorrect timing, but others
might require human intervention, e.g., failed traffic lights. The
management system should take appropriate actions in these
situations. Maintaining data confidentiality and integrity is also
an important security aspect of a management system since tampering
(or malfunction) can also lead to potentially dangerous situations.
4.9. Community Network Applications
Community networks are comprised of constrained routers in a multi-
hop mesh topology, communicating over lossy, and often wireless,
channels. While the routers are mostly non-mobile, the topology may
be very dynamic because of fluctuations in link quality of the
(wireless) channel caused by, e.g., obstacles, or other nearby radio
Ersue, et al. Informational [Page 20]
RFC 7548 Constrained Management: Use Cases May 2015
transmissions. Depending on the routers that are used in the
community network, the resources of the routers (memory, CPU) may be
more or less constrained -- available resources may range from only a
few kilobytes of RAM to several megabytes or more, and CPUs may be
small and embedded, or more powerful general-purpose processors.
Examples of such community networks are the FunkFeuer network
(Vienna, Austria), FreiFunk (Berlin, Germany), Seattle Wireless
(Seattle, USA), and AWMN (Athens, Greece). These community networks
are public and non-regulated, allowing their users to connect to each
other and -- through an uplink to an ISP -- to the Internet. No fee,
other than the initial purchase of a wireless router, is charged for
these services. Applications of these community networks can be
diverse, e.g., location-based services, free Internet access, file
sharing between users, distributed chat services, social networking,
video sharing, etc.
As an example of a community network, the FunkFeuer network comprises
several hundred routers, many of which have several radio interfaces
(with omnidirectional and some directed antennas). The routers of
the network are small-sized wireless routers, such as the Linksys
WRT54GL, available in 2011 for less than 50 euros. Each router, with
16 MB of RAM and 264 MHz of CPU power, is mounted on the rooftop of a
user. When a new user wants to connect to the network, they acquire
a wireless router, install the appropriate firmware and routing
protocol, and mount the router on the rooftop. IP addresses for the
router are assigned manually from a list of addresses (because of the
lack of autoconfiguration standards for mesh networks in the IETF).
While the routers are non-mobile, fluctuations in link quality
require an ad hoc routing protocol that allows for quick convergence
to reflect the effective topology of the network (such as
Neighborhood Discovery Protocol (NHDP) [RFC6130] and Optimized Link
State Routing version 2 (OLSRv2) [RFC7181] developed in the MANET
WG). Usually, no human interaction is required for these protocols,
as all variable parameters required by the routing protocol are
either negotiated in the control traffic exchange or are only of
local importance to each router (i.e. do not influence
interoperability). However, external management and monitoring of an
ad hoc routing protocol may be desirable to optimize parameters of
the routing protocol. Such an optimization may lead to a topology
that is perceived to be more stable and to a lower control traffic
overhead (and therefore to a higher delivery success ratio of data
packets, a lower end-to-end delay, and less unnecessary bandwidth and
energy use).
Ersue, et al. Informational [Page 21]
RFC 7548 Constrained Management: Use Cases May 2015
Different use cases for the management of community networks are
possible:
o A single NMS, e.g., a border gateway providing connectivity to the
Internet, requires managing or monitoring routers in the community
network, in order to investigate problems (monitoring) or to
improve performance by changing parameters (managing). As the
topology of the network is dynamic, constant connectivity of each
router towards the management station cannot be guaranteed.
Current network management protocols, such as SNMP and Network
Configuration Protocol (NETCONF), may be used (e.g., use of
interfaces such as the NHDP-MIB [RFC6779]). However, when routers
in the community network are constrained, existing protocols may
require too many resources in terms of memory and CPU; and more
importantly, the bandwidth requirements may exceed the available
channel capacity in wireless mesh networks. Moreover, management
and monitoring may be unfeasible if the connection between the NMS
and the routers is frequently interrupted.
o Distributed network monitoring, in which more than one management
station monitors or manages other routers. Because connectivity
to a server cannot be guaranteed at all times, a distributed
approach may provide a higher reliability, at the cost of
increased complexity. Currently, no IETF standard exists for
distributed monitoring and management.
o Monitoring and management of a whole network or a group of
routers. Monitoring the performance of a community network may
require more information than what can be acquired from a single
router using a network management protocol. Statistics, such as
topology changes over time, data throughput along certain routing
paths, congestion, etc., are of interest for a group of routers
(or the routing domain) as a whole. As of 2014, no IETF standard
allows for monitoring or managing whole networks instead of single
routers.
4.10. Field Operations
The challenges of configuring and monitoring networks operated in the
field by rescue and security agencies can be different from the other
use cases since the requirements and operating conditions of such
networks are quite different.
With technology advancements, field networks operated nowadays are
becoming large and can consist of a variety of different types of
equipment that run different protocols and tools that obviously
increase complexity of these mission-critical networks. In many
scenarios, configurations are, most likely, manually performed.
Ersue, et al. Informational [Page 22]
RFC 7548 Constrained Management: Use Cases May 2015
Furthermore, some legacy and even modern devices do not even support
IP networking. A majority of protocols and tools developed by
vendors that are being used are proprietary, which makes integration
more difficult.
The main reason for this disjoint operation scenario is that most
equipment is developed with specific task requirements in mind,
rather than interoperability of the varied equipment types. For
example, the operating conditions experienced by high altitude
security equipment is significantly different from that used in
desert conditions. Similarly, equipment used in fire rescue has
different requirements than flood-relief equipment. Furthermore,
interoperation of equipment with telecommunication equipment was not
an expected outcome or (in some scenarios) may not even be desirable.
Currently, field networks operate with a fixed Network Operations
Center (NOC) that physically manages the configuration and evaluation
of all field devices. Once configured, the devices might be deployed
in fixed or mobile scenarios. Any configuration changes required
would need to be appropriately encrypted and authenticated to prevent
unauthorized access.
Hierarchical management of devices is a common requirement in such
scenarios since local managers or operators may need to respond to
changing conditions within their purview. The level of configuration
management available at each hierarchy must also be closely governed.
Since many field operation devices are used in hostile environments,
a high failure and disconnection rate should be tolerated by the NMS,
which must also be able to deal with multiple gateways and disjoint
management protocols.
Multi-national field operations involving search, rescue, and
security are becoming increasingly common, requiring interoperation
of a diverse set of equipment designed with different operating
conditions in mind. Furthermore, different intra- and inter-
governmental agencies are likely to have a different set of
standards, best practices, rules and regulations, and implementation
approaches that may contradict or conflict with each other. The NMS
should be able to detect these and handle them in an acceptable
manner, which may require human intervention.
5. Security Considerations
This document discusses use cases for management of networks with
constrained devices. The security considerations described
throughout the companion document [RFC7547] apply here as well.
Ersue, et al. Informational [Page 23]
RFC 7548 Constrained Management: Use Cases May 2015
6. Informative References
[RFC6130] Clausen, T., Dearlove, C., and J. Dean, "Mobile Ad Hoc
Network (MANET) Neighborhood Discovery Protocol (NHDP)",
RFC 6130, DOI 10.17487/RFC6130, April 2011,
<http://www.rfc-editor.org/info/rfc6130>.
[RFC6568] Kim, E., Kaspar, D., and JP. Vasseur, "Design and
Application Spaces for IPv6 over Low-Power Wireless
Personal Area Networks (6LoWPANs)", RFC 6568,
DOI 10.17487/RFC6568, April 2012,
<http://www.rfc-editor.org/info/rfc6568>.
[RFC6779] Herberg, U., Cole, R., and I. Chakeres, "Definition of
Managed Objects for the Neighborhood Discovery Protocol",
RFC 6779, DOI 10.17487/RFC6779, October 2012,
<http://www.rfc-editor.org/info/rfc6779>.
[RFC6988] Quittek, J., Ed., Chandramouli, M., Winter, R., Dietz, T.,
and B. Claise, "Requirements for Energy Management",
RFC 6988, DOI 10.17487/RFC6988, September 2013,
<http://www.rfc-editor.org/info/rfc6988>.
[RFC7181] Clausen, T., Dearlove, C., Jacquet, P., and U. Herberg,
"The Optimized Link State Routing Protocol Version 2",
RFC 7181, DOI 10.17487/RFC7181, April 2014,
<http://www.rfc-editor.org/info/rfc7181>.
[RFC7228] Bormann, C., Ersue, M., and A. Keranen, "Terminology for
Constrained-Node Networks", RFC 7228,
DOI 10.17487/RFC7228, May 2014,
<http://www.rfc-editor.org/info/rfc7228>.
[RFC7326] Parello, J., Claise, B., Schoening, B., and J. Quittek,
"Energy Management Framework", RFC 7326,
DOI 10.17487/RFC7326, September 2014,
<http://www.rfc-editor.org/info/rfc7326>.
[RFC7547] Ersue, M., Ed., Romascanu, D., Schoenwaelder, J., and U.
Herberg, "Management of Networks with Constrained Devices:
Problem Statement and Requirements", RFC 7547,
DOI 10.17487/RFC7547, May 2015,
<http://www.rfc-editor.org/info/rfc7547>.
[IOT-SEC] Garcia-Morchon, O., Kumar, S., Keoh, S., Hummen, R., and
R. Struik, "Security Considerations in the IP-based
Internet of Things", Work in Progress, draft-garcia-core-
security-06, September 2013.
Ersue, et al. Informational [Page 24]
RFC 7548 Constrained Management: Use Cases May 2015
[IEEE802.11]
IEEE, "Part 11: Wireless LAN Medium Access Control (MAC)
and Physical Layer (PHY) Specifications", IEEE Standard
802.11, March 2012,
<http://standards.ieee.org/about/get/802/802.11.html>.
[IEEE802.15]
IEEE, "WIRELESS PERSONAL AREA NETWORKS (PANs)", IEEE
Standard 802.15, 2003-2014,
<https://standards.ieee.org/about/get/802/802.15.html>.
[IEEE802.15.4]
IEEE, "Part 15.4: Low-Rate Wireless Personal Area Networks
(LR-WPANs)", IEEE Standard 802.15.4, September 2011,
<https://standards.ieee.org/about/get/802/802.15.html>.
Acknowledgments
The following persons reviewed and provided valuable comments during
the creation of this document:
Dominique Barthel, Carsten Bormann, Zhen Cao, Benoit Claise, Bert
Greevenbosch, Ulrich Herberg, Ted Lemon, Kathleen Moriarty, James
Nguyen, Zach Shelby, Peter van der Stok, and Martin Thomson.
The authors would like to thank the reviewers and the participants on
the Coman mailing list for their valuable contributions and comments.
Juergen Schoenwaelder and Anuj Sehgal were partly funded by Flamingo,
a Network of Excellence project (ICT-318488) supported by the
European Commission under its Seventh Framework Programme.
Ersue, et al. Informational [Page 25]
RFC 7548 Constrained Management: Use Cases May 2015
Contributors
The following persons made significant contributions to and reviewed
this document:
o Ulrich Herberg contributed Section 4.9, "Community Network
Applications".
o Peter van der Stok contributed to Section 4.6, "Building
Automation".
o Zhen Cao contributed to Section 2.2, "Cellular Access
Technologies".
o Gilman Tolle contributed Section 4.4 "Energy Management".
o James Nguyen and Ulrich Herberg contributed to Section 4.10 "Field
Operations".
Authors' Addresses
Mehmet Ersue (editor)
Nokia Networks
EMail: mehmet.ersue@nokia.com
Dan Romascanu
Avaya
EMail: dromasca@avaya.com
Juergen Schoenwaelder
Jacobs University Bremen
EMail: j.schoenwaelder@jacobs-university.de
Anuj Sehgal
Jacobs University Bremen
EMail: s.anuj@jacobs-university.de
Ersue, et al. Informational [Page 26]