Content uploaded by Christos S. Tranoris
Author content
All content in this area was uploaded by Christos S. Tranoris on May 05, 2018
Content may be subject to copyright.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
1
Network Architectures and Management Group
Electrical and Computing Engineering Department,
University of Patras,
Patras, Greece
Technical Report
Open Source Software solutions implementing a
reference IoT architecture from the Things and Edge to
the Cloud
May 2018
This work is licensed under a Creative Commons Attribution 4.0 International License.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 3
Contents
1 Introduction ............................................................................................................................................ 5
1.1 The need for OSS IoT Cloud solutions ........................................................................................... 6
1.2 The research methodology adopted in this report ....................................................................... 6
2 A reference architecture and processes ................................................................................................. 7
2.1 “Things” ......................................................................................................................................... 9
2.2 IoT Edge Computing ...................................................................................................................... 9
2.2.1 The essential components of an IoT Edge .............................................................................. 10
2.3 IoT Cloud Platform ...................................................................................................................... 11
2.3.1 The essential components of an IoT Cloud platform .............................................................. 11
3 Open Source platforms for the three tiers of "Things"-Edge-Cloud ..................................................... 13
3.1 OSS for “things” .......................................................................................................................... 14
3.2 OSS for the Edge ......................................................................................................................... 16
3.2.1 Edge platforms ........................................................................................................................ 16
3.2.2 Efforts to be aware of ............................................................................................................. 17
3.2.3 OSS Smart home gateways ..................................................................................................... 18
3.2.4 Services and technologies for an Edge device ........................................................................ 18
3.3 OSS for the IoT Cloud .................................................................................................................. 21
3.3.1 Cloud IoT platform solutions .................................................................................................. 21
3.3.2 Efforts to be aware of around IoT Cloud OSS solutions .......................................................... 26
3.3.3 List of IoT OSS Add-on solutions ............................................................................................. 26
3.4 Semantic modeling for IoT .......................................................................................................... 27
3.4.1 Project Haystack ..................................................................................................................... 27
3.4.2 The SSN Ontology ................................................................................................................... 27
3.4.3 OneM2M ................................................................................................................................ 27
4 Bibliography .......................................................................................................................................... 28
C.Tranoris, S.Denazis - Network Architectures and Management Group
4
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 5
1 Introduction
According to Gartner Hype Cycle 2017 [1], IoT Platforms (among other technologies like 5G, Digital Twins,
Edge Computing) are one of the key mainstream technologies to create transformative impact in the next
2 to 5 years. Although the purpose of any IoT device, a “Thing”, is to connect with other IoT devices and
applications (cloud-based mostly), the gap between the device sensors and applications is filled by an IoT
Platform. Such a platform connects to the “Things” via a data network and provides backend services to
support applications enabled from the data generated by hundreds of sensors.
Both Cloud and IoT technologies are now mature enough and widely adopted on various vertical domains
such as Industry 4.0, Automotive, Smart Cities, eHealth etc. Edge computing lately emerged to cover near
the field problems such as real-time processing and low latency. IoT, Edge and Cloud Computing
technologies need to be delivered together in order to provide holistic solutions for these vertical
domains having “things” for the sensing/actuating part, edge for local processing, data pre-processing and
the cloud for application/business logic, data aggregation, big-data storage and analytics, managing the
infrastructure etc.
Usually an IoT platform is considered to run on the cloud, but the picture is now more complete with 5G
and Edge Computing. 5G will be a major IoT enabler given the well-publicized 5G promises of ultra-low
latencies, 100% coverage, high capacity and five nines reliability. On the other hand (Mobile) Edge
Computing arrived to raise the limitations of the centralized cloud computing architecture like ultra-short
latency. Edge Computing aims to place storage and computation resources at the network edge, while
data processing can be pushed from cloud to the edge, thus reducing the traffic bottleneck toward the
core network. Besides, it helps shorten end-to-end latency, it enables the computation offload from
power-constrained “Things” to the edge and it represents a fault resilient solution for its decentralized
architecture [2].
This report is a current snapshot of available platforms, efforts and solutions based purely on Open
Source Software that can be used to deploy an end-to-end IoT solution covering the three tiers of
“Things”, the Edge and the Cloud. While there is a wide literature, technologies and market analysis
available surveying IoT, the Edge and Cloud separately or combined, this work focuses on how OSS can be
delivered to provide complete solutions and identify missing gaps and potential opportunities. To
accomplish this, we define a high level architectural pattern and process, which tries to capture all the
core functionalities of IoT solutions. This reference architecture is defined after analyzing several
architectural IoT solutions and processes. Then we survey OSS examining features, maturity, community
support, licensing models etc. by mapping OSS platforms and components that can implement all the
parts and processes of this reference architecture.
C.Tranoris, S.Denazis - Network Architectures and Management Group
6
1.1 The need for OSS IoT Cloud solutions
Open Hardware innovations, like the Raspberry Pi, Arduino, etc, are making it easier, faster and cheaper,
even for individuals, to develop new devices for the following reasons:
- Interoperability by sharing source code
- Industry specific capabilities
- Benefit in general from OSS (user code reviews, improvements, etc)
- Based on Open Standards
- Loosely coupled and Modular via APIs
- Platform Independent: available to run in Public clouds (Amazon, Azure, etc) as well as in private
cloud or in the case of Edge able to run in multiple microprocessors
A study of [3] claims that once an AWS IoT-based system processes beyond 17-18 messages per second,
it’s time to start thinking about moving to another IoT platform to reduce operating cost. The cost per
message drops dramatically after 20 messages per second for open-source IoT platform and continues to
decrease with the growing message rate, while remaining constant for serverless IoT.
1.2 The research methodology adopted in this report
There is a plethora of open source projects regarding IoT, from the Industry to just individuals.
Contributions can be also found by many national or international funding efforts like the EU Horizon
2020. In Github for example the keyword IoT has around 39000 repositories, IoT-Edge around 200
repositories and Edge computing 145.
In this report we setup a few criteria to examine only those OSS solutions that fulfill them. These are:
• Active project: the project needs to be active, and have a clear roadmap (at least at the moment
of writing this report)
• Community support: there should be a group of contributors or at least a company or an
organization that actively supports it
• Open source code: it should be clear how to get the source code of the solution
• Clear licensing scheme: the solution should state how it can be used, even in commercial projects
• Be Interoperable and not bonded to specific close source solutions like Azure Iot Edge [4] or
Amazon Greengrass which are tightly coupled to their IoT cloud solution
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 7
2 A reference architecture and processes
The topics of this section are extensively covered by IoT and Cloud literature. However, it will help to
setup the problem area and later on to map the OSS solutions and their features. By examining a non-
exhaustive list of various reports [5, p. 12], papers [6] and reference architectures [7] [8] [9, pp. 39, 53]
[10] [11] [12] , we are able to define: i) a generic data/actions process from "things" and edge to the cloud
and ii) a typical 3-tier architectural pattern of "things"-edge-cloud for IoT solutions with the essential
elements of each tier.
In Fig 1 we provide a high level process involving the three tiers of "Things"-Edge-Cloud. This process is
very generic and try to highlight the activities involved from when data are availble from a sensor up to a
cloud application and vice-versa to sensors. Although, the defined activities are not all mandatory they
need to be considered by architects deploying an IoT solution. Also the placing of the activities either on
the edge or the cloud is based on the needs of the target vertical application. It is the authors opinion
based on existing reports and technologies the architectural placement of these activities to the edge or
to the cloud.
FIG 1 A GENERIC DATA/ACTIONS PROCESS FROM "THINGS", EDGE TO TH E CLOUD
As Fig 1 displays, when telemetry data are available from a sensor towards the edge in most cases a
needed to transform (see “Decode payload”) received data payload into typed values; e.g.
temperature/humidity values from a sensor are transmitted in 2 bytes, the 1st byte need to be decoded
into a Small Integer (representing a temperature value) and the 2nd t byte need to be decoded into a Small
Integer (representing a humidity value); for actuators on the other side there is the need to encode for
example (see “Encode payload”) a Boolean true or false value to open a valve to a bit. Of course, the
activities “Decode payload” and “Encode payload” might involve security aspects, e.g. decrypt/encrypt
values, MD5 signatures, etc.
The activity “Metadata enhancement”, might involve other metadata attached to the payload like
gateway information, timestamp, or GPS location of the sensor. E.g. a simple outdoor temperature sensor
does not have GPS location, so where it is located is hold as information in the metadata of the gateway
related to that sensor. “Semantic annotation” might involve transforming data to an equivalent model
C.Tranoris, S.Denazis - Network Architectures and Management Group
8
following some ontology models and standards. Semantic annotated data might be also store to reduce
latency and then pushed for further processing by the “Rules Activity” and “Local Control” for real-time
control, processing and decisions. ”Data aggregation” activity could be useful for offline-storage, local
Analytics and for forwarding aggregated sensor data in a consistent manner to/from the Cloud.
FIG 2 AN IOT 3- TIER ARCHIT ECTURAL PATTERN INVOLVING THINGS -EDGE-CLOU D
Messaging and connectivity is used to exchange data between the Edge devices and the Cloud. Cloud
maintains all the IoT infrastructure management and hosts several services that can process the data
received from a large number of Edge devices and “Things”. Again “Semantic Annotation” and Metadata
Enhancements might perform on received data. All data of the IoT infrastructure are stored to provide
global Analytics and perform the necessary Decision making. All data are available to Applications via an
Open API.
Finally, security is an aspect that should cover the data processing end-to-end. Not only the APIs and the
connectivity should be secure, but the processing of the data inside the service must be done in a secure
way. Multi-tenancy in the edge should be an aspect to be studied in future. Especially combined with the
concepts of Fog computing. There could be use-cases, for example, that it will not be possible for the Edge
to encrypt/decrypt and know the value and the type of the payload. This would be the case if an Edge is
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 9
used by multiple tenants. E.g. an Edge in a street road needs to process and send to the cloud data to two
different companies.
The described basic process that needs to be supported by an IoT solution can be implemented by a 3-tier
architectural pattern of "things"-edge-cloud. This is depicted in Fig 2 together with the essential
components of each tier which will help as later to identify OSS that implements them. Next sections
describe this 3-tier architectural pattern
2.1 “Things”
“Things” or devices measure and enable the interaction with the physical world. We distinguish two
categories, as Fig 2 displays:
- Constraint devices: Usually based on a micro-controller, contain some sensors or actuators and
implement usually a single wired or wireless protocol. Rarely can offer a remote management
system for limited configuration or firmware upgrade.
- Smart Devices: Based on micro-controllers or micro-processors typically more powerful than a
constraint device, usually have an embedded or real time OS. May include sensors and/or
actuators. May include remote management features; allow some minimal logic, device analytics
and network management and configuration especially if needed to be connected to network
types such as a cellular network.
2.2 IoT Edge Computing
Edge computing has gained a lot of attention lately due to the need of local processing or data
aggregation by many IoT data streams, or data filtering, since processing at the Edge can respond quickly
to local events, operate with intermittent connections, process locally a huge amount of data, provide
compute power and storage in the space between the device and the cloud. and transmit only data that
are worth of back to the cloud thus minimize the cost of transmitting IoT data to the cloud. Edge
computing is needed for local loop scenarios (e.g. home gateway, managing the lights of a street) and real
time processing (e.g. image processing). A forecast, cited by Huawei, reckons more than 50 billion
terminals and devices will be connected by 2020. In addition, over 50% of data is expected to be analyzed,
processed, and stored at the network edge in the future [13] . Edge Computing uses processing power of
gateway-devices to filter, pre-process and aggregate IoT data. It might be able to run complex analytics on
those data and, in a feedback loop, support decisions and actions about and on the physical world. Edge
compute devices include IoT gateways, routers, and micro data centers in mobile network base stations,
on the shop floor and in vehicles, among other places. Most applications in the areas like Industry 4.0,
Virtual Reality and Smart Cities are data intensive or time sensitive and depend on a lot of data from
sensors and devices being processed almost in real time. Furthermore, depending on the application,
information must be extracted from the data and transmitted securely. Moreover, the concept of “Edge
Intelligence” emerged in [9], a combination of Edge Computing, sophisticated Application Function and
Network Infrastructure management powered by Machine Learning. Having solutions coupled with the
new 5G network infrastructure, which provides the right software-based network for each application, in
C.Tranoris, S.Denazis - Network Architectures and Management Group
10
future, a central cloud will be obsolete for verticals with the need for robust and secure communication
and low latency. In terms of hardware there are many devices that can be used as IoT gateways
supporting many Operating Systems [14], from high-end costly solutions like the Dell Edge Gateway 5000
Series until cost-effective solutions like Raspbery-PI.
2.2.1 The essential components of an IoT Edge
Referring to Fig 2 the following components and services are identified at the Edge:
• Embedded OS/Realtime OS: the OS of the Edge device
• Remote Management: Services that allow to remotely manage the host services of the Edge, as
well as monitor the Edge, perform remote software deployment and updates
o An Identity registry of allowed tenants should be present
• Micro-services Runtime environment: an environment to run user/application code, provide
functionality isolation in the form of services possibly written in any programming language,
preferably updated in dynamic ways. Candidate implementations can be based on:
o Plugins: e.g. implementations around OSGI the Java modular framework [15], Service
Registries like consul [16]
o Containers: virtualization on the OS level with little overhead. Examples are Docker [17],
Linux Containers LXD [18] and rkt of CoreOS [19]
o Lambda functions: executable code when needed (term found e.g. in AWS Greengrass
[20]). Can be thought of as instantaneous instantiation and termination of containers in
a scalable manner
• A rules engine: needed for action management and performing local logic at the edge
• Alerts/Logging and Monitoring components responsible for:
o Continuous background Information monitoring/logging/debugging
o Custom alarms either locally or escalated to the cloud
• Data storage: a service for data storage for the generated data gathered by the Edge, given any
resource constraints (storage) of the Edge device
• Semantic annotators: implementing the activity of “Semantic annotation” as previously discussed
• Machine Learning: an optional component making the Edge intelligent by holding models that
can support decision support and local control. For example, an Edge gateway that controls 100
lamps of a street can decide about the dimming of the light of the whole street
• Real-time analytics: Services that provide real-time analytics on the data gathered by the Edge
device; given the constraint processing power of the device this can be limited for a short period
of data
• Device Registry and management: enable the Edge device to manage attached constraint/smart
devices. For example, an Edge LoRa gateway that get values from hundreds of nodes, has a
registry of their GPS location that is propagated later on to the cloud. Mobile nodes that move in
the city need to register or unregister themselves automatically from the Edge that is nearest to
them
• Data Decoding/Encoding services: as discussed in previous section implementing the
“Decode/Encode payload” activities
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 11
• Network management and communications: Edge device needs to support many connectivity
protocols especially in the South interface-towards devices (like BLE, ZigBee, OPC, etc) but also
wired and wireless network types (Ethernet, Cellular, Wi-Fi, etc)especially in the North interface-
towards the Cloud
• Security is a vertical aspect of the Edge device for the Data and the APIs. Lately there are efforts
to involve Blockchain for data like the IOTA protocol
2.3 IoT Cloud Platform
Managing an IoT Cloud platform can be a difficult process. Considerations about always-on connectivity
and latency need to be addressed, and while device to cloud data tends to be encrypted, having the data
residing outside enterprise walls might prove a problem for regulated industries. Also many organizations
deploy their private cloud or there are cases that you cannot afford any more to go with a Public IoT
solution and you just want to install your own IoT platform even on top of a cloud provider as the shirt
study in [3] explains.
2.3.1 The essential components of an IoT Cloud platform
Referring to Fig 2 the following components and services are identified at the IoT Cloud. The reader will
notice that there are similar components that also reside in the Edge. The difference here is that the
functionality is performed on top of an aggregation of Edge devices connected to the Cloud:
• Message Bus service, Connectivity, Network management: The Cloud platform needs to support
multiple connectivity protocols as well as a Message bus service that allows cloud services to
exchange messages
• Platform Management: Allows to manage all deployed IoT Cloud services
• Dashboards, UI: UI to Visualize and control the platform
o Data Visualization: Specific UI for visualizing data from the “things”
• Application/Services runtime environment: this is offered by almost any Cloud platform and
allows to install new Virtual Machines or Containers with new services and applications
o A useful feature here will be the presence of Lambda functions as in an Edge computing
device
• AAA of User/ Tenants registry and management and Access Control: needed for registering
access of users but also to allow IoT 3rd party applications to use the IoT Cloud API via API key
management, or define policies that allow access restrictions to resources
• Support for common APIs for 3rd part applications: External Interfaces [APIs, SDKs and gateways
that act as interfaces for 3rd party systems (e.g., ERP, CRM)]. For example , Through this interface
can be attached: billing and reporting, SDK application engine, Data export, Custom UIs, etc
• Semantic Annotations/Normalization: implementing the activity of “Semantic annotation” (e.g.
types based on ontologies and standards, geolocation, etc ) as previously discussed in Fig 1
• Rules engine: for decision making
C.Tranoris, S.Denazis - Network Architectures and Management Group
12
• Data Analytics: services for performing operations on top of data
• Data Storage: provide (big and scalable) data storage
• IoT Infrastructure Management: needed for managing the IoT infrastructure that is connected to
the IoT Cloud platform. It may offer:
o Device/Edge Management: Backend services for the management of Edge/devices,
perform remote software deployment & updates
o Device registry: keeps all the registered Edges and devices
o Monitoring: offers services for:
▪ Continuous background Information monitoring
▪ Creating and managing alarms
o Usage Analytics: Contains information about the usage and utilization of the
infrastructure
• Secure Communications: a common aspect for the whoel platform
• Device shadowing: used to store and retrieve current state information for a thing regardless of
whether the thing is connected to the Internet. Seen in commercial IoT platforms like Amazon
AWS IoT [21] a useful feature to find also in OSS.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 13
3 Open Source platforms for the three tiers of "Things"-Edge-
Cloud
In this section, we will examine OSS for each of the three tiers pf “Things”-Edge-Cloud. Although there is a
plethora of solutions found by just browsing the Internet, as explained in Section 1.2, this work will focus
on solutions that are Open sourced, with a clear licensing model, a vibrant community, large installed
base and a clear roadmap. We will present also technologies that may perform a single function (e.g. Data
Analytics) but can be easily integrated with the rest of the services existing in the Edge or the IoT Cloud, in
order to provide a more complete implementation of the reference architecture of Fig 2 .
C.Tranoris, S.Denazis - Network Architectures and Management Group
14
3.1 OSS for “things”
OS/RTOS on Embedded and Smart Devices exist for years, and many open source solutions are available.
There are useful comparisons and lists at [22] [23] Notable efforts being actively developed are: RIOT,
Zephyr from Linux foundation, Contiki and Apache Mynewt (See Table 1)
For the communication support (machine to machine) the most notable efforts are:
LWM2M client libraries:
- Eclipse Wakaama
- Leshan Client
MQTT clients:
- Paho one of the most known MQTT clients with many target programming languages, devices
like Arduino, Android and a very active community
- Mosquitto client: mainly for C,C++ clients again with an active community
Developers should be also aware of ROS. Robot Operating System (ROS) [24] is the most popular robotics
middleware and not an operating system. (Implementors call it, meta-operating system). The support for
Real-time systems is being addressed in the creation of ROS 2.0.
Finally Eclipse Edje defines a standard high-level Java API called Hardware Abstraction Layer (HAL) for
accessing hardware features delivered by microcontrollers such as GPIO, DAC, ADC, PWM, MEMS, UART,
CAN, Network, LCD, etc. that can directly connect to native libraries, drivers and board support packages
provided by silicon vendors with their evaluation kits.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
15
TABLE 1 WELL DE VELOPED O PERATING SYSTEMS FOR CONSTRAINT DEVICES
RIOT-OS
Zephyr
Apache Mynewt
Contiki
Multi-threading
yes
Partial support
yes
Partial support
Real-time
yes
yes
yes
partial
Source code URL
https://github.com/RIOT-OS/
https://github.com/zephyrproject-
rtos/zephyr
https://github.com/apache/mynewt-
core
https://github.com/contiki-
os/contiki
Support cite
http://www.riot-os.org/
www.zephyrproject.org
https://mynewt.apache.org/
http://www.contiki-os.org/
License
LGBL
Apache License, version 2.0
Apache License, version 2.0
BSD
No of Committers
181
268
70
161
Version
RIOT-2018.01
1.11.0
1.3.0
3.0
Last code update
5/5/2018
5/5/2018
5/5/2018
8/3/2018
Supported
Companies/Organiz
ations
Freie Universität Berlin, INRIA,
Hamburg University
Linux Foundation
Apache foundation
Texas Instruments, Atmel,
Cisco, ENEA, ETH Zurich,
Redwire, RWTH Aachen
University, Oxford University,
SAP, Sensinode, and many
others
Programming
languages
C, C++
C
C
C
Supported
Standards/
Communication
protocols
Partial POSIX compliance
Supports standards like 6Lowpan,
CoAP, IPv4, IPv6, and NFC.
Supports Bluetooth®, Bluetooth®
Low Energy, Wi-Fi*, 802.15.4
CoAP and 6LoWPAN . BLE, LoRa PHY
and LoRaWAN
Contiki supports fully standard
IPv6 and IPv4, along with the
recent low-power wireless
standards: 6lowpan, RPL, CoAP.
Date seen
5/5/2018
5/5/2018
5/5/2018
5/5/2018
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
16
3.2 OSS for the Edge
This section presents OSS solutions for the edge, according to the described methodology and the degree
that these solutions fulfill the presented edge architecture of Fig 2. Moreover, we present efforts by other
organizations especially from the areas of fog computing and 5G. Since most presented edge solutions do
not fully implement the proposed reference architecture we discuss complementary OSS solutions that
can be easily integrated with the target edge solutions that the underlying constraint hardware can
seamlessly execute.
3.2.1 Edge platforms
3.2.1.1 EdgeX Foundry
Hosted by The Linux Foundation, EdgeX Foundry [25] develops an open source common framework for
IoT edge computing. The framework consists of open source microservices plug-and-play components and
provides a reference software platform that claims to be hardware and OS agnostic. Dell Technologies
contributed the initial EdgeX code and now it is supported by over 60 members, Industry and SMEs from
both Europe and US. The project is at its infancy but there is already a second release called California
scheduled for June 2018. EdgeX Foundry executive director claims that “EdgeX wants to do for IoT on the
industrial side what Apache did for websites” [26]. The source code is located at [27] mostly is written in
Java and uses extensively Docker containers. Table 2 displays the architectural features that are defined
but still not all of them are implemented in the current release (in parentheses the corresponding EdgeX
Foundry architectural element)
3.2.1.2 Eclipse KURA
KURA [28] is an IoT gateway supported project by the Eclipse foundation. KURA has a lot of the
characteristic that an IoT solution needs at the Edge. It is platform independent (based on Java VM) and
contains both South Communication and Connectivity APIs and services to interface it with many device
protocols as well as a North interface for network management of various interfaces. Apart from standard
messaging solutions (MQTT) and routing it offers a remote management solution for monitoring and
controlling KURA services. Applications can be deployed remotely as OSGI bundles in similar way as other
KURA services are deployed (e.g. embedded database, GPS service etc). KURA is a quite mature platform,
initially contributed in 2014 to Eclipse by Eurotech [29]. Moreover, there is a benefit for developers that
there are already development tools around KURA . Source code is available at [30] with 31 contributors
at the day of writing and supported companies Eurotech, Redhat, IBM, Hitachi and Rapicorp
3.2.1.3 Little IoT Agent (liota)
Little IoT Agent (liota) is the answer of VMware for IoT solutions. Liota runs on any gateway that supports
Python, and has a small footprint of dependencies thus running in platforms like Raspberry Pi. Liota offers
an open source SDK for building IoT gateway applications for managing, monitoring and orchestrating
data in an IoT gateway. Thus, it does not offer a complete solution but rather the framework (liot calls
them abstractions) to develop and edge gateway solution. Liota source code is available at [31] with 11
contributors at the day of writing. In their future plans are to support more language bindings (C, C++,
Java and Lua), include CoAP transport and a blockchain mechanism for creating unique identifiers.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 17
3.2.1.4 Thingsboard IoT Gateway
This gateway solution is provided by ThnigsBoard.io as open source with the limitation that it allows to
integrate IoT devices connected to legacy and third-party systems with ThingsBoard cloud IoT. Since
ThingsBoard cloud IoT is also open source as you will see in next section we include also this solution here.
It is written in Java, but it is based on microservices not OSGI so developers can use more languages while
integrating with the gateway API. The ThingsBoard IoT gateway offers extension mechanisms, and some
tasks like device provisioning, local data persistence, message converters. They have in their roadmap
remote configuration, edge analytics and monitoring. The source code is at [32] actively developed but
with a limited number of contributors.
3.2.1.5 Flogo
Flogo [33] is a lightweight, based on GoLang, open source framework for edge devices due to small
footprint, but can be used also to complement an IoT cloud platform. Offers connectivity to IoT
technologies like MQTT, CoAP and offers a RESTful API. It offers an engine and a flow service to execute
flows. It is quite minimal, but has a wide range of (paid) extensions and plugins to extend its capabilities.
It is quite active project and mainly supported by TIBCO software.
3.2.1.6 SDC-Edge
StreamSets Data Collector Edge (SDC Edge) [34]is a light edge solution, written in Go programming
language and it enables at-scale data ingestion and analytics for edge systems. It provides edge analytics,
data normalization, and provides pipelines for both sending and receiving data. SDC Edge is supported by
StreamSets Inc.,can be easily installed in various OSS but there is also a containerized solution based on
Docker. It seems not a complete IoT edge solution since it is only oriented for Data streaming. The source
code is available at [35] and since it is supported by a single company has a limited number of
contributors
3.2.2 Efforts to be aware of
3.2.2.1 Telecom Infra Project – Edge Computing
The Telecom Infra Project (TIP) was founded in February 2016 and is an initiative by a large number of
members to create open solutions and as many open hardware and software components as possible in
the telecom infrastructure. TIP has a specific Edge Computing group [36] which plans to deliver open
solutions for both 5G & IOT service deployments.
3.2.2.2 The Open Connectivity Foundation
The Open Connectivity Foundation [37] was founded on February 2016 and is an initiative of over 400
member organizations mainly from Industry (like Microsoft, Samsung, Cisco, etc) delivering a standard
communications platform and specification, an open source implementation and a certification program
for devices involved in the Internet of Things (IoT).
3.2.2.3 ETSI multi-access-edge-computing
The European Telecommunications Standards Institute (ETSI) has initiated the multi access edge
computing effort [38]. They have launched a set of specification for APIs and services related to Edge
computing
3.2.2.4 The Open Edge Computing Initiative
The Open Edge Computing Initiative [39] founded by Carnegie Mellon University, Intel, NOKIA, Vodafone
and others delivers an open source mobile edge solution.
C.Tranoris, S.Denazis - Network Architectures and Management Group
18
3.2.3 OSS Smart home gateways
Although not placed a IoT gateways, Smart home gateways could be used by IoT platforms that want to
extend functionality until homes. Most of these projects want to bring solutions to their users without the
need of a cloud, but still the possibility is there, if there can be a North interface towards an IoT Cloud
solution.
3.2.3.1 openHAB (for the edge, see home gateway as edge)
OpenHAB [40] is an open source software platform written in Java, it offers SDK to construct frontends
and integrates different home automation systems, devices and technologies. It has a vibrant community
and a huge list of open source add-on plugins to control numerous home devices by different
manufacturers. It has a REST interface that it can be potentially exploited by interested IoT Cloud
solutions but also the openHAB Cloud Connector [41]
3.2.3.2 Home Assistant
Home Assistant [42] is an open-source home automation platform based on Python offering a mobile
application, SDK and components for connecting home devices. It also offers a REST API.
3.2.3.3 MyController.org
MyController [43] is a Java OSS controller and a home automation server to monitor and control a home
from anywhere. It’s architecture includes an MQTT and REST API at the North side, as well as, it has
multiple features we seek in Edge computing like remote monitoring and configuration, rule engine,
dashboards, user roles, etc. It is actively maintained at [44] by 9 developers at the time of writting
3.2.4 Services and technologies for an Edge device
resinOS [45]: It’s a minimal OSS platform for running containers on constraint devices,
Ubuntu IoT Gateway [46]: Canonical’s offer for implementing IoT edge gateways based on Ubuntu Core
OpenFaas [47] , fnproject [48], Apache OpenWhisk [49] : Serverless and Function as a Service (Faas)
technologies that enable lambda functions in a constraint device
Openstack-edge computing [50]: Openstack, the well-known OSS cloud platform, can be used for Edge-
computing solutions although does not include specific IoT edge services as the one described by our
reference architecture. At [51] for example, there are solutions to operate Openstack in the Edge
IOTA technology [52]: offers a marketplace allows connected devices to securely transfer, buy and sell
fine-granular and diverse data across the global to any buyer
Apache Edgent [53]: is a programming model and micro-kernel style runtime that can be embedded in
gateways and small footprint edge devices enabling local, real-time, analytics on the continuous streams
of data coming from equipment, vehicles, systems, appliances, devices and sensors of all kinds.
See also on next section Node-RED and Apache NiFi that could also be used in Edge devices for executing
dataflows.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
19
TABLE 2 EDGE SOLU TIONS AND ARCHIT ECTURAL FEA TURES
EdgeXFoundry
KURA
LIOTA
ThingsBoard IoT Edge
Flogo
SDC Edge
Environment
Mainly Java(Python, Go, C)
Java
Python
Java
Go
Go
Remote Management
yes(System Management)
yes (Administration GUI)
no
planned
yes
yes
Monitoring
no/planned
yes(Link Monitors)
no
planned
via Extensions
yes
Remote software
deployment and
updates
yes(System Management)
yes
no
no
via Extensions
no
Tenants Identity registry
yes(Client registration)
no
no
via Extensions
no
Micro-services Runtime
environment
yes(Docker containers)
no
yes
via Extensions
no
Rules engine
yes
yes( Apache Camel)
no
no
via Extensions
no
Alerts/Logging
Yes
no
planned
via Extensions
no
Data storage
Yes(Core Data)
yes(Data Services)
no
yes
via Extensions
yes
Semantic annotators
no (perhaps Heystack)
no
no
via Extensions
yes
Machine Learning
no
no
no
via Extensions
no
Real-time analytics
no/planned (Supporting
Services (SS) Layer )
no
planned
yes
Device Registry and
management
yes(Device Profile)
yes
Entities, EdgeSystems and
Devices, Device Discovery
yes
via Extensions
Data
Decoding/Encoding
services
provided SDK for developers
(Device Services (DS))
Metrics, SI Units
yes
via Extensions
yes
Network management
and communications
yes(Device/Export Services)
yes
yes
yesy
no
North Interface
yes(Export Services)
Cloud Services
DCCComms
MQTT
MQTT,CoAP
http
South Interface
yes(Device Services (DS))
Field Protocols/Device
Abstraction
DeviceComms
Sigfox,OPC-UA
MQTT,CoAP
mqtt, http
Security
yes
yes(SSL, Firewall service)
no
via Extensions
no
Scheduling
yes
no
no
no
Committers
~30
35
12
4
17
5
Supported Companies
Linux foundation,Industry
consortium 50+ members
Eclipse foundation (Eurotech,
Redhat, IBM, Hitachi and
Rapicorp)
VMWARE
ThingsBoard.io
TIBCO
StreamSets inc.
Licensing
Apache 2
Eclipse Public License
BSD 2
Apache 2
BSD
Apache 2.0
C.Tranoris, S.Denazis - Network Architectures and Management Group
20
EdgeXFoundry
KURA
LIOTA
ThingsBoard IoT Edge
Flogo
SDC Edge
Introduced
4/2017 Dell
6/2014 Eurotech
6/2016 VMWare
1/2017 ThingsBoard.io
TIBCO 10/2016
StreamSets 2017
Documentation
https://www.edgexfoundry.org/
https://www.eclipse.org/kura/
https://github.com/vmware/liota
https://thingsboard.io/d
ocs/iot-gateway/
https://tibcosoft
ware.github.io/fl
ogo/
https://streamset
s.com/products/s
dc-edge
Source code
https://github.com/edgexfound
ry
https://github.com/eclipse/ku
ra
https://github.com/vmware/liota
https://github.com/thing
sboard/thingsboard
https://github.co
m/TIBCOSoftwar
e/flogo
https://github.co
m/streamsets/dat
acollector-edge
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
21
3.3 OSS for the IoT Cloud
Similarly, as in previous section for the edge, this section presents IoT Cloud OSS solutions, according to
the described methodology and the architecture of Fig 2.
3.3.1 Cloud IoT platform solutions
3.3.1.1 KAA project
Kaa [54] is an open-source IoT platform, mainly written in Java, supported by a company called KaaIoT
[55] which actively maintains it and support is. The source code of the platform is available at [56] and
seems to be maintained by an adequate number of contributors. It can be deployed as a single node or in
a cluster deployment for high availability and has multi-tenancy support, that is a single Kaa instance can
support multiple independent applications. Overall Kaa, offers a good solution, however in order for
developers to connect a device to the platform, they need to embed Kaa SDK types. Still there is a RESTful
API for accessing Kaa’s services.
3.3.1.2 Eclipse Kapua
Kapua [57] is a platform for managing IoT edge gateways backed by Eclipse foundation. Kapua offers
device, data and registry management as well as messaging services. It allows remote administration and
easy integration with other applications via RESTful API. It also includes data storage for analytics as well
as dashboards for visualizations. However, it seems that it is not quite generic but more oriented for
specific solutions integrated with Eclipse KURA. The source code is available at [58], it is actively
maintained with primary backer companies Red Hat and Eurotech.
3.3.1.3 TTN
The Things Network [59] was created in 2015 and it is based on LoRaWAN technology. It is a free and
open IoT network around the world which consists of hundreds of gateways. It is included in this report
since TTN offers as OSS all its software stack at [60] and one can either use it for a private deployment or
contribute resource to the TTN IoT cloud. TTN IoT Cloud platform offers most of the envisaged services of
an IoT Cloud stack: connectivity with edge LoRaWAN gateways, multitenancy, device management, and an
MQTT as well as a RESTful API for end user applications.
3.3.1.4 FIWARE
FIWARE [61] is the outcome of the Future Internet Public-Private Partnership Programme (FI-PPP), the
European Commission launched in 2011. Today FIWARE has formed an open community supporting from
platform APIs to FIWARE labs and startups. FIWARE software components and services are deployed as
FIWARE calls the Generic Enablers (GEs). In this context FIWARE offers an IoT software stack [62]. The core
component is FIWARE flagship GE Orion Context Broker (also for northbound connectivity) while IoT
Agents offer connectivity (i.e. MQTT, http, etc) in south interface to gateways and devices. Most of the
FIWARE components are supported by Telefonica and the source code is available at [63]. Although there
is adequate documentation for the offered components, there is no simple installation and integrating
them all together is not a trivial process.
C.Tranoris, S.Denazis - Network Architectures and Management Group
22
3.3.1.5 ThingsBoard
ThingsBoard [64] is an OSS IoT platform offering from core IoT cloud services to Dashboards via widgets.
Standards like MQTT, CoAP and HTTP are supported as well as storage, device management, logging, etc.
It claims to support seamless horizontal scalability and multitenancy, meaning that one can deploy it even
for more complex scenarios of supporting multiple customers and applications. There are two ways to
integrate devices, either by implementing within your device their MQTT, CoAP, HTTP APIs or by
integrating the ThingsBoard Gateway as discussed in section 3.2.1.4. Source code is available at [65] and it
is actively supported by an adequate number of contributors.
3.3.1.6 OpenRemote
OpenRemote [66] is an open source project offering three main elements around cloud configuration
tools (OpenRemote Manager), a controller for integrating protocols acting as a gateway (OpenRemote
Controller) and managing rules and UI control panel for designers (OpenRemote Designer). OpenRemote
Manager is the IoT cloud part of the whole solution and offers APIs for 3rd party services. It is actively
maintained by the Openremote Inc company and the source code is available at [67].
3.3.1.7 IoTivity
IoTivity [68] a Linux Foundation hosted project is actively developed and maintained and is highly
sponsored by the Open Connectivity Foundation (OCF).
3.3.1.8 Devicehive
Devicehive [69] is another OSS IoT platform which distributed under Apache 2.0 license. It can be easily
deployed via Docker and Kubernetes. It can connect to any device via REST API, WebSockets or MQTT.
3.3.1.9 SiteWhere
SiteWhere [70] is an OSS solution that provides the ingestion, storage, processing, and integration of
device data. It runs on top of Apache Tomcat coupled with MongoDB and HBase. It delivers a complete
device management solution, allows connecting devices with MQTT, AMQP, Stomp, and other protocols
and adding devices through self-registration, REST services, or in batches
3.3.1.10 Thinger.io
Thinger.io [71] is an OSS platform that claims easy scalability. It offers admin consoles and REST API for
Application integration and easy integration with Docker and IFTT [72]. It offers also libraries for Arduino
and Android. The source code is not maintained the last year
3.3.1.11 ThingSpeak
ThingSpeak [73] is an OSS platform that is tightly coupled with MATLAB allowing users to analyze and
visualize uploaded data using MATLAB without requiring the purchase of a MATLAB license. Although the
commercial solution is active the source code is not maintained the last 3 years
3.3.1.12 Zetta
Zetta [74] is an open source platform built on Node.js for creating Internet of Things servers that run
across geo-distributed computers and the cloud and build APIs for device interaction. Although Zetta has
a small number of contributors is actively developed.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 23
3.3.1.13 Distributed Services Architecture (DSA)
Distributed Services Architecture (DSA) [75] is an active OSS IoT platform that facilitates device inter-
communication, logic and applications at every layer of the Internet of Things infrastructure.
3.3.1.14 WSo2
WSo2 [76] seems a mature solution and active OSS IoT platform, which also claims that can distribute and
manage applications/firmware of devices as well as Share device operations/data with other users. It also
has mechanisms to Deploy lightweight streaming analytics on the edge or a local gateway to optimize
network traffic. Also support real-time analysis of IoT data streams with WSO2 streaming analytics,
identifying complex patterns and kicking off alerts and initiating responsive actions.
Table 3 displays some of the core features of the most active solutions.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
24
TABLE 3 IOT CLOUD OSS SOLUTIONS
Kapua
KAA
TTN
FIWARE
ThingsBoard
OpenRemote
IoTivity
DeviceHive
SiteWhere
Thinger.io
ThingSpeak
Zetta
DSA
WSo2
Environment
Java
Java
Go/Python/
Javascript
~
Java
Java
C,C++
Java,Go
Java,Pythion,
Javascript
C++, Java
Ruby
Node.js
Java,
Dart,Python
Java,Javascr
ipt
IoT Platform
Management
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Monitoring
Yes
Yes
Yes
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Remote software
deployment and
updates
Yes
Yes
No
No
No
Yes(for
controllers)
No
No
No
No
No
No
No
Yes
Tenants Identity
registry
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
Yes
No
No
No
Yes
Micro-services
Runtime
environment
no
No
No
No
No
No
No
No
No
No
No
No
No
No
Rules engine
No
No
No
No
Yes
Yes
No
No
No
No
No
No
No
Alerts/Logging
Yes
Yes
Yes
Yes
Yes
No
Via plugins
Yes
No
Yes
No
No
Yes
Data storage
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Semantic
annotators
Yes
No
No
Yes via
Attributes
No
No
via
extensions
No
No
No
No
No
No
Machine
Learning
no
No
No
No
No
No
No
via
extensions
No
No
No
No
No
No
Real-time
analytics
Yes
Yes
No
Yes
Yes
No
No
Yes
No(Free)
Yes
No
Yes
(Stream)
Yes
Yes
Device Registry
and
management
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
Yes
Data
Decoding/Encodi
ng services
no
via SDK
Yes
Yes
via extensions
Yes
Yes
via
extensions
No
No
No
via
extensions
No
No
Network
management
and
communications
Yes
Yes
No
Yes
No
No
Yes
Yes
Yes
Yes
No
Via Scouts
Yes
No
North Interface
REST
REST
MQTT,
REST
Yes
REST
REST API
REST API
REST API,
MQTT
REST API
REST API
REST API,
MQTT
REST API
REST API
REST API
South Interface
MQTT,
CoAP
via SDK
LoRaWAN
MQTT,Co
AP
MQTT,CoAP,H
TTP
proprietary
CoAP,BLE,
etc
WebSocket
s or MQTT
MQTT,
AMQP,
Stomp,
WebSockets
MQTT,
CoAP
REST API
REST API
REST API
MQTT,
WSO2
Security
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Committers
17
56
20-113
~40
~25
~8
~60
33
13
-
8
11
~12
~80
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 25
Kapua
KAA
TTN
FIWARE
ThingsBoard
OpenRemote
IoTivity
DeviceHive
SiteWhere
Thinger.io
ThingSpeak
Zetta
DSA
WSo2
Supported
Companies
RedHat,Eur
otech,Bosc
h
KaaIoT
Thethingsin
dustries
Telefonic
a
ThingsBoard
OpenRemote
Inc
Open
Connectiv
ity
Foundatio
n
DeviceHive
SiteWhere
Thinger.io
ThingSpeak
Apigee,
Zetta
Community
DGLogik
WSO2
Licensing
EPL 1.0
Apache 2.0
MIT
AGPL3.0,
Apache2.
0,…
Apache2.0
GNU AGPL
3,0
Apache
2.0
Apache 2.0
CPAL-1.0
MIT
GPL3
MIT
Apache 2.0
Apache 2.0
Introduced
6/2016
Eurotech
~2015
~2015
~2015
~2016
~2013
~12/2015
~2012
~2013
2015
2010
2014
2016
2015
Documentation
https://ww
w.eclipse.or
g/kapua/
http://kaap
roject.githu
b.io/kaa/do
cs/v0.10.0/
Welcome/
http://thet
hingsnetwo
rk.org/
https://fi
ware-iot-
stack.rea
dthedocs.
io/en/lat
est/
https://things
board.io/docs
/
http://www.
openremote.
com/commu
nity/
https://w
ww.iotivit
y.org/doc
umentati
on
https://devi
cehive.com
http://www.
sitewhere.or
g/
http://doc
s.thinger.i
o/console/
https://thin
gspeak.com
/
http://www
.zettajs.com
/
https://gith
ub.com/IOT
-
DSA/docs/
wiki
https://doc
s.wso2.com
/display/IoT
S310/
Source code
https://gith
ub.com/ecli
pse/kapua
https://gith
ub.com/kaa
project/kaa
https://gith
ub.com/Th
eThingsNet
work/ttn
https://gi
thub.com
/Fiware
https://github
.com/thingsb
oard/thingsbo
ard
https://githu
b.com/openr
emote
https://git
hub.com/i
otivity/iot
ivity
https://gith
ub.com/dev
icehive
https://githu
b.com/sitew
here
https://git
hub.com/t
hinger-io
https://gith
ub.com/iob
ridge/things
peak
https://gith
ub.com/zet
tajs/zetta
https://gith
ub.com/IOT
-DSA
https://gith
ub.com/ws
o2
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
26
3.3.2 Efforts to be aware of around IoT Cloud OSS solutions
3.3.2.1 IoT-EPI - European Initiative for IoT platform development
The IoT-European Platforms Initiative - IoT-EPI [77] is s a European Initiative addressing EU-funded H2020
programs in the domain of IoT. Currently under its umbrella are the following IoT research projects: Inter-
IoT [78], BIG IoT [79], AGILE [80], symbIoTe [81], TagItSmart! [82], VICINITY [83] and bIoTope [84]. Most
European funded projects are publishing their results as open source. From the above EU funded projects,
related with this work are INTER-IoT, Big IoT and AGILE (especially for gateways).
3.3.3 List of IoT OSS Add-on solutions
Useful solutions for IoT Cloud developers to be aware of are the following:
3.3.3.1 IOTivity
IOTivity [85] is a Linux Foundation OSS project sponsored by the Open Connectivity Foundation (OCF) [86]
with a single goal to provide a framework for device-to-device connectivity. It offers a cloud interface at
the northbound as well as discovery and messaging services at the southbound interface. It is not an IoT
Cloud solution, but by integrating IoTivity components with other 3rd part components one can have a
good solution. The IoTivity model Things as resources while its offered cloud stack is composed of 4
subsystems, Interface, Account Server, Resource Directory and Message Queue. The project offers also
software components for IoT device side for handshaking, resource registration/discovery, etc. There are
some docker containers to ease the setup process. AllJoyn, another OCF project, is also merged as work
now under IoTivity . IoTivity ‘s source code is available at [87] while it is a very active project.
3.3.3.2 Node-RED
Node-RED [88] is an IBM OSS contributed development tool for creating flows. The interesting part is the
runtime engine in Node.js that can run your flows in the background, thus it can be very easily integrated
with your IoT solution to ease the process of deploying flows and defining new logic for your IoT
applications. The source code is available at [89] and is currently a very active project with more than 60
contributors to the project.
3.3.3.3 Apache NiFi
A similar solution to Node-RED, Apache NiFi [90] is a light-weight dataflow execution engine, to compose
and execute generic IoT applications.
3.3.3.4 Analytics
There are no specific solution for IoT Analytics, but in most platforms well known Stream and Big data
solutions such as ElasticSearch, Apache Spark, Cassandra and Kafka for real-time and batch processing can
be easily integrated.
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 27
3.3.3.5 For Lambda functionality support
Serverless approaches are a feature of your cloud hosting. Apche OpenWhisk (as presented on edge
section) is a serverless, open source cloud platform that executes functions, but still in incubation phase.
Also if you have your own Cloud infrastructure for example based on OpenStack, there is no equivalent of
e.g. Amazon Lambda, but there is something to expect from Openstack projects like Zun [91]
3.3.3.6 Eclipse Hono
Eclipse Hono [92] is an IoT connector which provides remote service interfaces for connecting large
numbers of IoT devices to a back end and interacting with them in a uniform way regardless of the device
communication protocol. Born out of the collaboration of large companies, including Red Hat and Bosch
both members of the Eclipse Foundation, Hono aims to provide an open source framework that is
completely open source, through which to build an end-to-end IoT solution that supports the main
patterns: telemetry, events and command/control.
3.4 Semantic modeling for IoT
IoT testbed providers have recently started to add semantics to their frameworks allowing the creation of
the semantic sensor web (SSW), which is an extension of the current Web in which information is given
well-defined meaning, enabling M2M communications and interactions between objects, devices and
people [93]. There is academic literature around the proper semantic modeling in IoT data and
applications, but here are some efforts that have wide adoption
3.4.1 Project Haystack
Project Haystack [94] is an open source initiative to develop naming conventions and taxonomies for
building equipment and operational data by fostering the common association and interests of software
and technology companies. It is focused on developing semantic modeling solutions for data related to
smart devices including: building equipment systems, automation and control devices, sensors and
sensing devices, promotion and education with respect to the semantic data modeling industry for
building automation systems, and to engage in educational activities directed towards the improvement
of business conditions of the semantic data modeling industry for smart device data, all on a not-for-profit
basis. All work developed by the project-haystack.org community is provided for use as open source
software under the Academic Free License 3.0.
3.4.2 The SSN Ontology
The SSN ontology [95] is one of the most significant and widespread models to describe sensors and IoT-
related concepts. The SSN ontology provides concepts describing sensors, such as outputs, observation
value, feature observed, observation time, accuracy, precision, deployment configuration, method of
sensing, system structure, sensing platforms and feature of interest.
3.4.3 OneM2M
OneM2M has published a report for home automation and describes concepts and relationships [96]. It
collects and studies the state ‐of‐the‐art technologies that may be leveraged by oneM2M to enable its
abstraction & semantics capability. This includes a collection of terminology and use cases considered by
other standardization or industrial for a working on ontologies, semantics and abstraction.
C.Tranoris, S.Denazis - Network Architectures and Management Group
28
4 Bibliography
[1]
"Top Trends in the Gartner Hype Cycle for Emerging Technologies, 2017," [Online]. Available:
https://www.gartner.com/smarterwithgartner/top-trends-in-the-gartner-hype-cycle-for-
emerging-technologies-2017/. [Accessed 12 12 2017].
[2]
H. C. e. al., "Bringing the Cloud to the Edge," in Proc. IEEE INFOCOM Workshop, Toronto, Ontario,
Canada, May 2014.
[3]
H. o. I. a. D. Igor Ilunin, "How to Choose Your IoT Platform – Should You Go Open-Source?," 10 10
2017. [Online]. Available: https://www.iotforall.com/iot-platform-open-source-vs-serverless.
[Accessed 18 10 2017].
[4]
"Azure iot-edge," [Online]. Available: https://github.com/Azure/iot-edge.
[5]
SDxCentral, "2017 SDxCentral IoT Infrastructure Report - Online Edition," 2017. [Online].
Available: https://www.sdxcentral.com/reports/2017/internet-of-things/. [Accessed 20 10 2017].
[6]
"The Art of Low - Cost IoT Solutions," June 2017. [Online]. Available:
http://www.dataart.com/downloads/DataArt_White_Paper_Art_of_Low_Cost_IoT_Solution.pdf.
[Accessed 18 10 2017].
[7]
E. I. W. Group, "The Three Software Stacks Required for IoT Architectures," 2016. [Online].
Available: https://iot.eclipse.org/resources/white-papers/Eclipse%20IoT%20White%20Paper%20-
%20The%20Three%20Software%20Stacks%20Required%20for%20IoT%20Architectures.pdf.
[Accessed 20 10 2017].
[8]
IoT-Analytics, "White paper – Guide to IoT Solution Development," 2016. [Online]. Available:
https://iot-analytics.com/product/guide-to-iot-solution-development/. [Accessed 10 10 2017].
[9]
IEC, "White Paper - Edge Intelligence," 2017. [Online]. Available:
http://www.iec.ch/whitepaper/pdf/IEC_WP_Edge_Intelligence.pdf.
[10]
I. I. Consortium, "Industrial Internet Reference Architecture," June 2015. [Online]. Available:
https://www.iiconsortium.org/IIRA-1-7-ajs.pdf. [Accessed 20 10 2017].
[11]
ETSI, "GS MEC 003 V1.1.1 Mobile Edge Computing (MEC); Framework and Reference
Architecture," March 2016. [Online]. Available:
http://www.etsi.org/deliver/etsi_gs/MEC/001_099/003/01.01.01_60/gs_MEC003v010101p.pdf.
[Accessed 20 10 2017].
[12]
e. a. Yun Chao Hu, "ETSI White Paper No. 11 - Mobile Edge Computing A key technology towards
5G," September 2015. [Online]. Available:
http://www.etsi.org/images/files/ETSIWhitePapers/etsi_wp11_mec_a_key_technology_towards_
5g.pdf. [Accessed 20 10 2017].
[13]
Huawei, "Huawei Launched Edge-Computing-IoT Solution, Enabling Industry Digital
Transformation," 2017. [Online]. Available:
http://www.huawei.com/en/events/mwc/2017/b2b/iot/huawei-edge-computing-iot-solution.
[Accessed 2017].
[14]
postscapes, "IoT Hardware Guide," [Online]. Available: https://www.postscapes.com/internet-of-
things-hardware/. [Accessed 19 10 2017].
[15]
O. Alliance, "OSGI," [Online]. Available: https://www.osgi.org/. [Accessed 20 10 2017].
[16]
HashiCorp, "Consul," [Online]. Available: https://www.consul.io/. [Accessed 20 10 2017].
[17]
Docker, "Docker," [Online]. Available: https://www.docker.com/. [Accessed 20 10 2017].
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 29
[18]
L. Containers, "LXD," [Online]. Available: https://linuxcontainers.org/lxd/. [Accessed 20 10 2017].
[19]
"CoreOS," [Online]. Available: https://coreos.com/rkt/. [Accessed 20 10 2017].
[20]
"Greengrass," Amazon, 10 2017. [Online]. Available: https://aws.amazon.com/greengrass/.
[Accessed 20 10 2017].
[21]
A. A. IoT, "iot-thing-shadows," [Online]. Available:
http://docs.aws.amazon.com/iot/latest/developerguide/iot-thing-shadows.html. [Accessed 22 10
2017].
[22]
Wikipedia, "Comparison of real-time operating systems," [Online]. Available:
https://en.wikipedia.org/wiki/Comparison_of_real-time_operating_systems . [Accessed 05 05
2018].
[23]
"List of open source real-time operating systems," [Online]. Available: https://www.osrtos.com/ .
[Accessed 05 05 2018].
[24]
R. O. S. (ROS), "Robot Operating System (ROS)," [Online]. Available: http://www.ros.org/.
[Accessed 5 5 2018].
[25]
"EdgeX Foundry," [Online]. Available: https://www.edgexfoundry.org/.
[26]
"Edgexfoundry move forward its push harmonize iot edge computing," [Online]. Available:
http://www.ioti.com/infrastructure/edgex-foundry-moves-forward-its-push-harmonize-iot-edge-
computing.
[27]
"EdgeXFoundry source code," [Online]. Available: https://github.com/edgexfoundry.
[28]
"Eclipse KURA," [Online]. Available: https://www.eclipse.org/kura/.
[29]
Eurotech, https://www.eurotech.com/en/press+room/news/?672.
[30]
"KURA source code," [Online]. Available: https://github.com/eclipse/kura. [Accessed 3 12 2017].
[31]
VMWARE, "LIOTA," [Online]. Available: https://github.com/vmware/liota. [Accessed 28 11 2017].
[32]
Thingsboard.io, "Thingsboard-gateway," [Online]. Available:
https://github.com/thingsboard/thingsboard-gateway. [Accessed 3 12 2017].
[33]
Flogo. [Online]. Available: http://www.flogo.io/. [Accessed 12 12 2012].
[34]
"StreamSets Data Collector Edge," [Online]. Available: https://streamsets.com/products/sdc-edge.
[Accessed 14 12 2017].
[35]
"SDC Edge source code," [Online]. Available: https://github.com/streamsets/datacollector-edge.
[Accessed 14 12 2017].
[36]
T. I. Project. [Online]. Available: http://telecominfraproject.com/project/access-projects/edge-
computing/. [Accessed 3 12 2017].
[37]
O. C. F. (OCF), "OPEN CONNECTIVITY FOUNDATION (OCF)," [Online]. Available:
https://openconnectivity.org/. [Accessed 5 5 2018].
[38]
ETSI, "multi access edge computing," [Online]. Available: http://www.etsi.org/technologies-
clusters/technologies/multi-access-edge-computing. [Accessed 3 12 2017].
[39]
o. e. computing. [Online]. Available: http://openedgecomputing.org.
[40]
OpenHAB. [Online]. Available: http://www.openHAB.org. [Accessed 3 12 2017].
[41]
"Connector, Openhab cloud," [Online]. Available:
http://docs.openhab.org/addons/ios/openhabcloud/readme.html. [Accessed 3 12 2017].
[42]
"Home-assistant," [Online]. Available: https://home-assistant.io/.
[43]
M. Controller. [Online]. Available: https://mycontroller.org. [Accessed 3 12 2017].
[44]
"Mycontroller-org sourcecode," [Online]. Available: https://github.com/mycontroller-
org/mycontroller. [Accessed 3 12 2017].
[45]
ResinOS. [Online]. Available: https://resinos.io/. [Accessed 6 12 2017].
C.Tranoris, S.Denazis - Network Architectures and Management Group
30
[46]
I. g. o. Ubuntu, "Ubuntu Core," [Online]. Available: https://www.ubuntu.com/internet-of-
things/gateways. [Accessed 6 12 2017].
[47]
"OpenFaaS," [Online]. Available: https://www.openfaas.com/. [Accessed 6 12 2017].
[48]
fnproject. [Online]. Available: http://fnproject.io/. [Accessed 6 12 2017].
[49]
A. Openwhisk. [Online]. Available: https://openwhisk.apache.org/. [Accessed 12 12 2012].
[50]
O. edge-computing. [Online]. Available: https://www.openstack.org/edge-computing/. [Accessed
6 12 2017].
[51]
A. L. e. al, "Revising OpenStack to Operate Fog/Edge Computing Infrastructures," in IEEE
International Conference on Cloud Engineering (IC2E), Vancouver, BC, Canada, 2017.
[52]
IOTA, "The Next Generation of Distributed Ledger Technology | IOTA," [Online]. Available:
https://www.iota.org/. [Accessed 5 5 2018].
[53]
A. Edgent, "Apache Edgent," [Online]. Available: http://edgent.apache.org/. [Accessed 05 05
2015].
[54]
K. project. [Online]. Available: https://www.kaaproject.org/. [Accessed 12 12 2017].
[55]
KaaIoT. [Online]. Available: https://www.kaaiot.io/company/. [Accessed 12 12 2017].
[56]
K. s. code. [Online]. Available: https://github.com/kaaproject/kaa. [Accessed 12 12 2017].
[57]
E. Kapua, "Eclipse Kapua," [Online]. Available: Eclipse Kapua. [Accessed 6 12 2017].
[58]
"Kapua source code," [Online]. Available: https://github.com/eclipse/kapua.
[59]
"The Things Network," [Online]. Available: https://www.thethingsnetwork.org/. [Accessed 12 12
2017].
[60]
"The Things Networks source code," [Online]. Available: https://github.com/thethingsnetwork.
[Accessed 12 12 2012].
[61]
FIWARE. [Online]. Available: https://www.fiware.org/. [Accessed 14 12 2017].
[62]
"FIWARE IoT Stack," [Online]. Available: https://catalogue.fiware.org/iot_stack. [Accessed 14 12
2017].
[63]
"FIWARE source code," [Online]. Available: https://github.com/telefonicaid. [Accessed 14 12
2017].
[64]
"ThingsBoard," [Online]. Available: https://thingsboard.io. [Accessed 14 12 2017].
[65]
"ThingsBoard source code," [Online]. Available: https://github.com/thingsboard/thingsboard.
[66]
"Openremote Community edition," [Online]. Available:
http://www.openremote.com/community/. [Accessed 14 12 2017].
[67]
"Openremote source code," [Online]. Available: https://github.com/openremote/. [Accessed 14
12 2017].
[68]
IoTivity, "IoTivity," [Online]. Available: https://iotivity.org/. [Accessed 5 5 2018].
[69]
Devicehive, "Devicehive," [Online]. Available: https://devicehive.com. [Accessed 5 5 2018].
[70]
sitewhere, "sitewhere," [Online]. Available: http://www.sitewhere.org/. [Accessed 5 5 2018].
[71]
thinger.io, "thinger.io," [Online]. Available: https://thinger.io/. [Accessed 5 5 2018].
[72]
I. T. T. That, "IFTT," [Online]. Available: http://ifttt.com/. [Accessed 5 5 2018].
[73]
ThingSpeak, "ThingSpeak," [Online]. Available: https://thingspeak.com/. [Accessed 5 5 2018].
[74]
zetta, "zetta," [Online]. Available: http://www.zettajs.com/. [Accessed 5 5 2018].
[75]
D. S. A. (DSA), "Distributed Services Architecture (DSA)," [Online]. Available: http://iot-dsa.org/.
[Accessed 05 05 2018].
[76]
wso2, "wso2," [Online]. Available: https://wso2.com/iot. [Accessed 5 5 2018].
[77]
"IoT-EPI - European Initiative for IoT platform development," [Online]. Available: http://iot-
epi.eu/. [Accessed 14 12 2017].
Technical Report | OSS solutions implementing a reference IoT architecture from the Things and Edge to the Cloud
This work is licensed under a Creative Commons Attribution 4.0 International License 31
[78]
"INTER-IoT project," [Online]. Available: http://www.inter-iot-project.eu/. [Accessed 14 12 2017].
[79]
"Big IoT," [Online]. Available: http://big-iot.eu/. [Accessed 14 12 2017].
[80]
"AGILE," [Online]. Available: http://agile-iot.eu/. [Accessed 14 12 2017].
[81]
"symbIoTe," [Online]. Available: https://www.symbiote-h2020.eu/. [Accessed 14 12 2017].
[82]
"TagItSmart!," [Online]. Available: http://www.tagitsmart.eu/. [Accessed 14 12 2017].
[83]
"VICINITY," [Online]. Available: http://vicinity2020.eu/. [Accessed 14 12 2017].
[84]
"bIoTope," [Online]. Available: http://www.biotope-project.eu/. [Accessed 14 12 2017].
[85]
"IoTivity," [Online]. Available: https://www.iotivity.org/. [Accessed 14 12 2017].
[86]
"Open Connectivity Foundation (OCF)," [Online]. Available: http://www.openconnectivity.org/.
[Accessed 14 12 2017].
[87]
"IoTivity source code," [Online]. Available: https://github.com/iotivity. [Accessed 14 12 2017].
[88]
"Node-RED," [Online]. Available: http://nodered.org/. [Accessed 20 12 2017].
[89]
"Node-RED source code," [Online]. Available: https://github.com/node-red/node-red. [Accessed
20 12 2017].
[90]
A. NiFi, "Apache NiFi," [Online]. Available: https://nifi.apache.org/. [Accessed 5 5 2018].
[91]
O. Zun, "OpenStack Zun," [Online]. Available: https://wiki.openstack.org/wiki/Zun. [Accessed 5 5
2018].
[92]
E. HONO, "Eclipse HONO," [Online]. Available: https://www.eclipse.org/hono/. [Accessed 5 5
2018].
[93]
A. Sheth, C. Henson and S. Sahoo, "Semantic sensor web.," IEEE Internet Computing, vol. 12, no. 4,
p. 78–83, 2008.
[94]
P. Haystack, "Project Haystack," [Online]. Available: https://project-haystack.org/. [Accessed 5 5
2018].
[95]
M. Compton, P. Barnaghi and P. Bermudez , "The SSN ontology of the W3C semantic sensor
network incubator group," Web Semantics: Science, Services and Agents on the World Wide Web,
vol. 17, pp. 25-32, 2012.
[96]
OneM2M, "Study of abstraction and semantics enablement v.0.7.0. study of existing abstraction
and semantic capability enablement technologies for consideration by OneM2M," Technical
Report OneM2M (TR 0007), 2014.
