A Social Inspired Broker for M2M Protocols

Vincenza Carchiolo

1 a

, Alessandro Longheu

2 b

, Michele Malgeri

2 c

and Giuseppe Mangioni

2 d

Dipartimento di Matematica ed Informatica, Università degli studi di Catania, Catania, Italy

DIEEI, Università degli studi di Catania, Catania, Italy

Keywords: IoT, Social Networks, P2P Overlay Networks, M2M Protocols.

Abstract: Internet of things can be viewed as the shifting from a network of computers to a network of things.To support

M2M communication, several protocols have been developed; many of them are endorsed by client-broker

model with a publish-subscribe interaction mechanism. In this paper we introduce a multi broker solution

where the network of brokers is inspired by social relationships. This allow data sharing among several IoT

systems, leads to a reliable and effective query forwarding algorithm and the small world effect coming from

mimic humans relations guarantees fast responses and good query recall.

1 INTRODUCTION

Internet of Things (IoT) can be considered the natural

evolution of the Internet and it is expected to modify

our habits in a most shocking way rather than the

Internet itself. In the IoT context, several network-

enabled devices provide connectivity for a set of

(possibly many) objects, places and/or environments

to Internet, shifting from a network of computers to a

network of things (Giri et al., 2017).

While IoT represents a convergence point among

different existing technologies as Radio Frequency

IDentification -RFID- tags (Marquardt et al., 2010)

and sensor/actuator networks (Sgroi et al., 2005,

Akyildiz et al., 2002), it actually poses many

challenges. In particular, to leverage smarter

intelligent devices to really improve everyday life,

they should interact in a fast, automatic and seamless

fashion. Moreover, devices often work in a

geographically distributed area subject to dynamic

changes, with heterogeneity in data type, format,

availability and granularity.

The resulting need for an effective Machine to

machine (M2M) communication lead to several

different protocols developed during last years, as

CoAP, MQTT, AMQP and many others (Hunkeler et

al. 2008, Vinoski 2006). Most of them are based on a

https://orcid.org/0000-0002-1671-840X

https://orcid.org/0000-0002-9279-3129

https://orcid.org/0000-0002-9898-8808

https://orcid.org/0000-0001-6910-0112

client-broker model with a publish-subscribe

interaction mechanism, where each client publishes

messages to a broker, and brokers receive

subscription requests from other clients. Every

message is published to an address, known as topic.

Clients can subscribe to multiple topics and receive

from bokers every message published for those

topics.

In this work we propose a multi broker solution

for M2M protocols where each broker is a node of a

social based peer-to-peer network that operates as

PROSA (Carchiolo et al., 2006), (Carchiolo et al.,

2008), a semantic social inspired overlay network

whose query forwarding algorithm is reliable and

effective and whose small world structure guarantees

fast responses and good query recall.

In the proposed architecture, brokers interact not

only with devices to endorse publish-subscribe

mechanism for the specific M2M protocol, but also

each other, providing information sharing for a

distributed broker solution. Brokers communication

does not depend on the actual M2M protocol being

used, rather only publish-subscribe mechanism is

required.

To share information among brokers, the idea we

adopt is that semantic proximity of resources is

mapped onto topological proximity of node, whereas

Carchiolo, V., Longheu, A., Malgeri, M. and Mangioni, G.

A Social Inspired Broker for M2M Protocols.

DOI: 10.5220/0007765101010105

In Proceedings of the 4th International Conference on Complexity, Future Information Systems and Risk (COMPLEXIS 2019), pages 101-105

ISBN: 978-989-758-366-7

101

query forwarding/answering effectiveness and

efficiency come from the social nature of PROSA

network here exploited.

The paper is organized as follows. In Section 2 we

consider the impact of social-based inspiration within

IoT scenario, whereas in Section 3 we present our

proposal in more detail, finally showing our

concluding remarks and a plan for future works in

Section 4.

2 SOCIAL INSPIRATION IN IoT

The influence that social-based paradigma can have

in the IoT context is manifold, depending on the

vision we consider. The simple “Social IoT”

classification is therefore incomplete, and currenlty

three types of social inspiration can be discussed,

given that in IoT system is built over three concepts

(thing oriented, Internet oriented and semantic

oriented):

 Thing-to-Thing Social IoT (TTS-IoT)

 Thing-to-Human Social IoT (THS-IoT)

 Application-to-Application Social IoT

(AAS-IoT)

Two examples of TTS-IoT are described in

(Holmquist et al, 2001) and (Mendes, 2011). In the

first (quite dated) work, authors exploits wireless

sensor nodes to build temporary social relationships,

considering how sensor owners drive the building of

such relationships. In (Mendes, 2011) the idea is to

provide objects with global information exchange so

they become context-aware and can be involved in

“conversation” like humans.

The THS-IoT approach is certainly the most

natural and followed by many researchers. For

instance, "Socialite" (Kim et al., 2017) is a tool that

uses semantic technologies for SIoT end-user

programming. In particular, authors extracted some

rules from online survey, exploiting them in

automatic runtime decisions in IoT scenarios. In this

work, knowledge representation is semantic based

and rules support information sharing to facilitate

social relationships among IoT users.

Figure 1: Publish-subscriber basic model (B=broker,

D=device, S=subscriber).

Other works, as (Kranz et al., 2010) and (Guinard,

2010) highlight the integration of IoT with social

networks, where social networks is used as a base for

resources discovery by IoT. Even if some

applications are shown (Kranz et al., 2010), these

works though do not address the quiestion of building

social relationship. Other approaches use cloud-based

platform, as "Lysis" (Girau et al., 2017); in this work,

objects act as agents with social relationships,

increasing both network scalability and information

discovery.

For what concerns the third approach (AAS-IoT),

in (Saleema et al., 2018) a first attempt was proposed;

the work promotes data exchange and reuse among

IoT applications, so they can use mutual social

relationships to leverage their services.

Our proposal actually can be thinked about a

fourth social inspired model in IoT, i.e. the Broker-

to-broker Social IoT (BBS-IoT); in other proposals,

e.g. (D’Elia et al., 2018), the idea of distributed multi

broker overlay platform is proposed, though no social

inspiration is considered in query forwarding and

answering.

Figure 2: Multiple brokers and their IoT networks.

3 BBS-IoT WITH PROSA

Publish-subscriber protocols adopt a communication

mechanism that relies on message brokers for

multicast messages exchange. In particular, instead of

having the sender forwarding messages to the set of

receivers, the sender publish its messages (concerning

a given topic) to a broker. Receivers that are

interested in (and were previously subscribed for) a

specific topic will receive related messages from that

broker. A single IoT network can be viewed as the

broker with its set of subscribers and devices

(publishers), as shown in Fig. 1.

COMPLEXIS 2019 - 4th International Conference on Complexity, Future Information Systems and Risk

102

Multiple-broker approach is largely used within

IoT context; each broker can share its information

with others to implement a distributed broker. The

general architecture is shown in Fig. 2, where

multiple brokers with their IoT networks are

represented.

The question is which subset of brokers should be

selected to guarantee effectiveness and efficiency of

the whole architecture, at the same time avoiding to

spread information to brokers that are not interested

in the same topic. For instance, if a broker B1 knows

that B2 is the reference broker for a given topic T

(nodes interested in T will subscribe to B2), a

semantic link from B1 to B2 arise; whenever B1

receives a message concerning T, it sends the

message not only to T’s subscribers but also to B2. In

general, if more brokers are involved in T, B1

forwards the message to the broker peer-to-peer (P2P)

network to reach such brokers, while avoiding to

spread the message to any broker.

Currently, no definitive standardization exist for

this mechanism, and the overall performance in

searching and retrieving resources heavily depends

on the organization of broker P2P network.

The solution we propose here is to build the

broker network according to the PROSA model. In

particular, PROSA leverage social relatioships to

exploit the small-world emerging property (Watts and

Strogatz, 1998), thus resulting in efficient message

forwarding.

In PROSA, two kinds of social links are

considered: acquaintances and semantic link; the

former models social relationships raising from

interactions in everyday life (e.g. those concerning

colleagues in the same office at work), whereas a

semantic link models those acquaintances with which

a stronger relation exists, for instance if I need IT

support for my laptop, I will search not any of my

colleagues, but the IT specialist. Note that a

semantic–link is not symmetric.

Actually, in PROSA semantic links are split in

two subcategories, i.e. temporary and full semantic

links. To describe their difference, consider an

example: if a friend asked us something about "golf"

and we were not able to answer him, we will anyway

remember that he is involved with golf. This results

into a link stronger than simple acquaintances (AL),

thanks to past queries, and it is called Temporary

Semantic Link (TSL). Whenever an answer to a query

is provided, this lead to a stronger and stronger link

named Full Semantic Link (FSL).

To promote an acquaintance link to a semantic

one, some additional semantic information (e.g. about

interest, culture, abilities) are required. In real life

semantic links building simply comes from sharing a

knowledge field or a passion or simply an interest

with a person and interact with him in some

circumstances. Once such a semantic link is

established, as soon as a need concerning that field

occurs, you’re ready to use that link to get assistance

or collaboration. In real life semantic links are widely

used to speed up information retrieval.

Our goal here is to build such a broker network,

exploiting both acquaintance and semantic links; a

broker joins the network to achieve links to others

according to the social model described above, i.e. by

linking (semantically) with broker with similar

interests, culture, hobbies, works and so on, and

keeping a certain number of “random” acquaintances.

If the network of brokers catches the dynamics of the

social model, the resulting network should be a small-

world. To achieve this, we need i) a system to model

knowledge, culture, interests, and ii) a network

management algorithm implementing the social

model.

For a given broker B1, we could assume that the

set of other brokers in the broker network is viewed

as a Virtual Smart Device (VSD) capable of

providing information concerning various topics (Fig.

3); this way, social behaviour is encapsulated into the

VSD, and this allows to not affect existing IoT

networks. A better approach is tough to view the rest

of broker network as a Virtual Subscriber to which B1

sends its information; this is discussed in the next

paragraph.

Figure 3: Modelling the broker network as a VSD.

3.1 Broker Net as Virtual Subscriber

As shown in Figure 4, the broker network is modeled

as a Virtual Subscriber. In particular, a generic broker

B1 behaves as a server and each one uses data

variables named “Topics” and routes all the messages

among connected subscriber. Topics are represented

in a hierarchical form, e.g.:

/root/level1/../leveln/Measure

A Social Inspired Broker for M2M Protocols

103

Where the so-called Localization is:

/root/level1/../leveln/

And Measure is the topic itself (named “Measure”

according to typical IoT values provided by devices).

Messages are sent to all subscriber interested in

the topic. Here we consider two set of subscribers:

 Local subscribers, directly connected to B1

 Remote subscribers, i.e. those reachable via the

broker network

To model the attractiveness (interest) for a given

topic, for Local Subscriber we exploit the MQTT

standard, whereas for the Remote, hidden inside the

Virtual Subscriber, the PROSA approach is adopted.

In particular, a Topic Vector (TV) is introduced,

with three fields:

 Localization, that allows to identify where the

topic can be extracted

 Measure, that allows to identify what topic is

obtained

 Authorization, that allows to identify the level of

security to be compliant with when spreading the

information; this field is defined by the broker

during the network joining process

The Localization is derived from the topic

provided by each publisher; he/she can insert

different information, as absolute, relative or

descriptive spatial coordinates (examples in Table 1).

Table 1: Localization examples.

USA/California/SanFrancisco/Silicon

Valley/temp

myhome/groundfloor/livingroom/temp

Absolute spatial information can be used as they

are, whereas for relative information this field is

turned into absolute position by fetching the position

of the broker the publisher is connected to.

Finally, in the case of descriptive information, a

semantic analysis is exploited to assess absolute

spatial information.

In PROSA, a peer (node in the network) receiving

a query forwarded by an unknown peer can extract

some information about source peer knowledge from

the query itself, and this can be used to establish a new

link with the source peer. Whenever a broker sends a

topic, in PROSA this is interpreted as a peer search

request, i.e. a query to find a broker interested to that

topic; in (Carchiolo et al., 2007) and (Carchiolo et al.,

2010) search query management is described in

detail.

During query processing, physical distance are

evaluated to deal with Localization field, whereas a

semantic distance (e.g. ontology based) is exploited

to cope with Measure field; finally, a filtering

function (ACL based) is considered to implement the

Authorization field.

According to PROSA, in the broker network links

building follows the way people “link” to others in

social networks, i.e. relationships among people are

usually based on similarities in interests, culture,

hobbies, knowledge and evolve from simple

acquaintance links to semantic links (TSL or FSL).

Since relationships (links) are not symmetric, it is

necessary to distinguish the source broker (SB) from

the destination broker (DB) in a link.

Each Broker maintains a list of known brokers,

(Broker List, or BL), built on the query strategy

previously described. If the link is a simple AL, the

broker doesn’t know the corresponding TV: in this

case an empty TV is placed into the vector field.

When TSL is considered, the broker doesn’t know the

TV of the linked Broker, but a Temporary Broker

Vector (TBV) is built based on the topic information

received in the past from that broker. Finally, if the

link is a FSL, the TV is put in the vector field.

A new Broker that wants to join broker network,

just searches other Brokers and adds some of them in

his BL as ALs.

In PROSA links dynamics are strictly related to

clients connected with a Broker and with the topic

requested from this client. When a new client of a

Broker requires information about a topic, he modify

the TV of the Broker.

Figure 4: Modeling the broker network as a Virtual

Subscriber.

4 CONCLUSIONS

In this position paper, a multi broker solution to

provide support within IoT context were presented.

The proposal relies on social relationships, in

COMPLEXIS 2019 - 4th International Conference on Complexity, Future Information Systems and Risk

104

particular on the PROSA network whose query

forwarding algorithm is effective and small world

structure assure fast responses and good query recall;

our proposal can be adopted to share information

even with heterogenous IoT system.

Several questions though call for further

investigation, in particular (1) how different brokers

could share topic semantics and (2) how privacy and

trust (Carchiolo et al., 2015) can be effectively

managed.

ACKNOWLEDGEMENTS

This work has been supported by the Universita' degli

Studi di Catania, “Piano della Ricerca 2016/2018

Linea di intervento 2”.

REFERENCES

Giri et al., 2017. Internet of things (IoT): a survey on

architecture, enabling technologies, applications and

challenges. In Proc. of the 1st International Conference

on Internet of Things and Machine Learning (IML '17).

ACM press.

Marquardt et al., 2010. Rethinking RFID: Awareness and

Control for Interaction with RFID Systems. In Proc.

CHI 2010. ACM Press.

Sgroi et al., 2005. A service-based universal application

interface for ad hoc wireless sensor and actuator

networks. In Ambient Intelligence. Springer Verlag.

Akyildiz et al., 2002. Wireless sensor networks: A survey.

In Computer Networks 38, 4. Elsevier Press.

Hunkeler et al., 2008. MQTT-SA publish/subscribe protocol

for wireless sensor networks. In Proceedings of the 3rd

Intl Conf. on Communication System Software and

Middleware (COMSWARE’08). IEEE Press.

Vinoski, 2006. Advanced message queuing protocol. In

IEEE Internet Computing, vol 6. IEEE Press.

Carchiolo et al., 2008. PROSA: P2P Resource Organisation

by Social Acquaintances. In Agents and Peer-to-Peer

Computing. AP2PC 2006. LNCS, vol 4461. Springer,

Berlin, Heidelberg.

Carchiolo et al., 2006. Social Behaviours in P2P Systems:

An Efficient Algorithm for Resource Organisation. In

15th IEEE International Workshops on Enabling

Technologies: Infrastructure for Collaborative

Enterprises (WETICE'06). IEEE Press.

Holmquist et al. 2001. Smart-its friends: A technique for

users to easily establish connections between smart

artefacts. In Int. Conf. on Ubiquitous Computing.

Springer Berlin Heidelberg.

Mendes, 2011. Social-driven Internet of connected objects.

In Proc. of the Interconnected Smart Objects with the

Internet Workshop. IEEE Press.

Kim et al., 2017. Empowering End Users for Social Internet

of Things. In Proc. of the Second Int. Conf. on Internet-

of-Things Design and Implementation. ACM Press.

Guinard et al., 2010. Sharing using social networks in a

composable web of things. In 8th IEEE Int. Conf. on

Pervasive Computing and Communications Workshops

(PERCOM Workshops). IEEE Press.

Kranz et al., 2010. Things that twitter: social networks and

the Internet of things. In What can the Internet of Things

do for the Citizen (CIoT) Workshop at 8th Int. Conf. on

Pervasive Computing.

Girau et al., 2017. Lysis: A Platform for IoT Distributed

Applications over Socially Connected Objects. In IEEE

Internet of Things Journal. IEEE Press.

Saleema et al., 2018. SCDIoT: Social Cross-Domain IoT

Enabling Application-to-Appl. Communications. In

IEEE International Conference on Cloud Engineering

(IC2E). IEEE Press.

D’Elia et al., 2018. A Multi-Broker Platform for the Internet

of Things. In Internet of Things, Smart Spaces, and Next

Generation Networks and Systems. Springer.

Watts and Strogatz, 1998. Collective dynamics of ’small-

world’ networks. In Nature, 393:440–442.

Carchiolo et al., 2007. Efficient Searching and Retrieval of

Documents in PROSA. In: Databases, Information

Systems, and Peer-to-Peer Computing. DBISP2P.

LNCS vol 4125. Springer, Berlin, Heidelberg.

Carchiolo et al., 2010. An adaptive overlay network

inspired by social behavior. Journal of Parallel and

Distributed Computing, Vol.70-3. Elsevier.

Carchiolo et al., 2015. The Cost of Trust in the Dynamics of

Best Attachment. Computing and Informatics journal,

vol 34. S. Academy of Sciences Press.

A Social Inspired Broker for M2M Protocols

105