Towards an Opportunistic, Socially-driven, Self-organizing, Cloud

Networking Architecture with NovaGenesis

Antonio M. Alberti

, Waldir Moreira

, Rodrigo da Rosa Righi

, Francisco J. Pereira Neto

Ciprian Dobre

and Dhananjay Singh

ICT Laboratory, Instituto Nacional de Telecomunicac¸

oes - INATEL, Santa Rita do Sapuca

ı, Minas Gerais, Brazil

COPELABS, Univ. Lus

ofona, Lisbon, Portugal

PIPCA, Universidade do Vale do Rio dos Sinos, S

ao Leopoldo, Rio Grande do Sul, Brazil

Computer Science Dpt., Univ. Politechnica of Bucharest, Bucharest, Romania

Electronics Eng. Dpt., Hankuk Univ. of Foreign Studies,Yongin, South Korea

Keywords:

Future Internet, Opportunistic, Cloud Networking, Socially-driven, Self-organizing, Software-as-a-Service.

Abstract:

The exponential growth on the number of mobile devices and their capabilities are leveraging new possibilities

of networking architectures for processing, storing, and exchanging of information. At a glance, existing archi-

tectures take advantage of these devices, the social behavior of their users, and/or the dynamicity on resource

usage. Despite of the potential of existing initiatives, they do not interoperate which reduce their applications

and deployment. As we walk towards a very dynamic world (regarding the user needs and characteristics, the

information traversing the network, and the networking capability to adaptation at both users features and con-

tent of the demands levels), these architectures should merge into a solution that ﬁts any type of scenario. In

this paper, we specify an opportunistic, socially-driven, self-organizing, cloud networking architecture using a

future Internet proposal named NovaGenesis. We highlight the requirements and solutions that NovaGenesis

brings to accommodate the inherent challenges of today’s dynamic networking scenario. Thus, we describe

a convergent architecture, which integrates the new requirements with the already implemented NovaGenesis

features.

1 INTRODUCTION

The way that users communicate nowadays is very

different from a few years ago. The classic host-based

paradigm, in which users would request speciﬁc con-

tent stored in speciﬁc locations, is giving room to new

forms for users to exchange data. This is a prod-

uct of not only devices becoming very portable, but

also gathering the latest advancements in terms of

processing, storage, and wireless technologies. Thus,

users are able to produce and consume content any-

where and anytime, and such data exchange may take

place through spontaneously formed networks. Fur-

thermore, this content can be stored locally and/or on

the users’ personal clouds, as well as on public clouds.

Different networking approaches have emerged.

Delay/disruption-tolerant networking (DTN) (Caini

et al., 2008) deals with scenarios where intermit-

tent connectivity is rather common among network

nodes. Opportunistic Networking (ON) (Moreira

et al., 2012) exploits different contact opportunities

among users to exchange data. Cognitive Radio Net-

work (CRN) (Ahmed et al., 2010), aims at explor-

ing radio frequency spectrum holes to accommodate

communication links. Information-Centric network-

ing (ICN) (Xylomenos et al., 2014) emerged from the

idea that people care for content itself no matter where

it is, decoupling content identiﬁcation (“what”) from

its location. User-centric networking (UCN) (Soﬁa

and Mendes, 2008) includes user-provided devices

and systems to build social-driven networks. These

approaches are being used to address different chal-

lenges (e.g., intermittent connectivity, high mobility,

longer delays, expensive infrastructure and/or con-

nectivity) of emerging access networks. Also, cloud

and big data systems have been integrated to these

networking approaches as they can further increase

the capabilities of the user devices and handle the vast

amount of data that users produce.

One can observe that these networking approaches

comprise a wide variety of data exchange and pro-

cessing approaches that, despite of the already known

Alberti A., Moreira W., da Rosa Righi R., J. Pereira Neto F., Dobre C. and Singh D..

Towards an Opportunistic, Socially-driven, Self-organizing, Cloud Networking Architecture with NovaGenesis.

DOI: 10.5220/0005526300270036

In Proceedings of the 2nd International Workshop on Emerging Software as a Service and Analytics (ESaaSA-2015), pages 27-36

ISBN: 978-989-758-110-6

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

potential, still operate for speciﬁc purposes and do not

interoperate as required. Users are not so interested in

the technicalities and employed approaches that allow

them to exchange information. Instead, users expect

an integrated cloud/networking infrastructure that al-

lows them to share information directly with other

peers, without relying on infrastructure and/or expen-

sive connectivity services.

Despite the fact that existing networking and in-

formation access approaches share few concepts and

concerns, they fail to cohesively integrate the afore-

mentioned technologies (Alberti, 2012). We believe

that such approaches can neatly come together as to

form a converged architecture. Our starting point is

NovaGenesis (www.inatel.br/novagenesis) initiative,

which is a clean slate convergent information archi-

tecture (CIA) being developed by our team. By CIA,

we mean an architecture that integrates information

exchanging with processing and storage. It can be

seen as a generic architecture, where Internet is con-

verged to cloud computing and big data. NovaGene-

sis already integrates ICN, CRN, service-oriented ar-

chitecture (SOA) (Papazoglou et al., 2007), software-

deﬁned networking (SDN) (McKeown et al., 2008),

and Internet of things (IoT) (Conti, 2006). NovaGen-

esis paves the way to a CIA that advances integration

efforts to include ON, ICN, UCN, and cloud.

In this context, the main aim of this work is to ful-

ﬁl the need for a solution of converged architectures

that allows for the application of different networking

paradigms in a neat way. Each of these key ingre-

dients strongly contributes to advance a speciﬁc de-

signing dimension. When two or more of these ingre-

dients are synergistically integrated, there is a “cross

fertilization”, a catalyzing effect, which favor global

architectural advances instead of local ones. NovaGe-

nesis is explored as the foundation for this architec-

ture. New services are proposed to be integrated to

NovaGenesis proposal. We contribute with a novel

approach towards user-centric architectures.

This paper is structured as follows: we start

by brieﬂy overviewing the relevant networking

paradigms to be considered by our convergent archi-

tecture in Section 2. In Section 3, we showcase No-

vaGenesis software-based CIA, the architecture we

chose as the foundation of our social-driven initia-

tive. Section 4 presents our contribution, showing a

qualitative analysis regarding how NovaGenesis ad-

dresses the challenges behind the proposed archi-

tecture, pointing perspectives, pre-requirements, and

open issues. Finally, Section 5 concludes the paper,

also highlighting some future work.

2 RELEVANT NETWORKING

PARADIGMS

This section presents the different paradigms that

shall be comprised by our convergent architecture.

Opportunistic Networking (ON) exploits the

contact opportunities taking place among users’ de-

vices to allow data exchange. Opportunistic forward-

ing can be seen from two perspectives regarding user

social behavior, namely social-oblivious and social-

aware approaches. From these approaches, the lat-

ter has gained much attention of the research com-

munity given the fact that social information is less

volatile (i.e., changes less, favoring data exchange)

than mobility (Moreira and Mendes, 2013). Our con-

vergent architecture shall take into consideration the

dynamism of user social behavior found in their daily

routine in order to properly infer the different levels

of social interactions among users and the interests

of these users (Moreira et al., 2012) (Ciobanu et al.,

2013) (Ciobanu et al., 2014b) (Moreira et al., 2014)

since the dynamics of user social behavior does have

an impact on the performance of opportunistic for-

warding (Moreira and Mendes, 2015a) (Moreira and

Mendes, 2015b). With that, our architecture is ex-

pected to provide the users with data exchange op-

portunities over only socially relevant links, thus with

improved delivery probability while reducing associ-

ated cost and experienced latency.

With Content-Centric Networking (CCN), data

traverses the network according to the match between

its name and the interests that users may have in such

contents, independently of its location, resulting in an

efﬁcient, scalable, and robust content delivery. There

are different efforts for deﬁning a CCN architecture

(DONA, NDN, NetInf), each with its own particu-

larities (e.g., employ their own naming scheme) and

looking into different CCN aspects (e.g., naming, se-

curity, routing, caching, transport) according to the

application to which they have been devised (Xy-

lomenos et al., 2014). With the advances in tech-

nology, devices have become more portable and with

increased capabilities (e.g., processing, storage). In

such dynamic networking scenarios, users are pro-

sumers (i.e., producers and consumers) of informa-

tion with a high demand to share/retrieve content

anytime and anywhere, independently of the inter-

mittency level of wireless connectivity, dynamic be-

havior of users, physical obstacles, lack of cooper-

ation, closed (i.e, secured) networks, among others.

Given its potential, the convergent architecture shall

incorporate CCN features (end-to-end path abstrac-

tion, interest-based content-driven dissemination) as

to cope with the dynamicity of today’s networking

ESaaSA2015-WorkshoponEmergingSoftwareasaServiceandAnalytics

scenario. Thus, our architecture shall provide the con-

tent that users demand based on their interests which

they propagate while carrying on their daily routines.

User-Centric Networking (UCN) focuses on the

user who is the main pillar for routing, security, and

among other networking aspects (Soﬁa et al., 2014).

This networking paradigm empowers the user that can

easily provide services (e.g., connectivity, printing)

to others. Within this context, the user besides pro-

ducing and consuming content as mentioned before,

now becomes a micro-provider (Soﬁa and Mendes,

2008) changing currently known Internet communi-

cation models (end-users comprise a user-provided

network extending services, where user willingness in

sharing resources/services allows scalable services).

There are different approaches that relate

to user-provided networking spanning a vast

range of applications: making use of proprietary

equipment to share connectivity (SparkNet at

htpp://sparknet.ﬁ/); simply turning the end-user

device into a sharing point (Whisher/WiFi.Com at

http://www.whisher.com); creating a network for

sharing resources (Wray Village’ wireless broadband

at http://www.infolab21.co.uk/livinglab). However, it

is important to note that these approaches aim solely

at sharing connectivity. This is indeed a type of

resource that users are very much interested, but there

is more to it. The ULOOP (siti.ulusofona.pt/˜uloop/)

project clearly highlights this: ULOOP users are

provided with th means to exchange resources as they

wish based on the trust levels between these users

and/or based on the exchange of virtual currency.

The project considers they dynamic behavior of

users to allow the exchange of different types of

resource beyond connectivity. With this in mind, our

convergent architecture aims at providing users with

the means of sharing the resources they have the most

and make use of resources they require at a given

moment. By combining opportunism with content

centricity, the convergent architecture is expected to

further empower the users allowing them to naturally

engage in the system y providing and consuming

resources according to their current demands.

Cloud Computing Elasticity exploits the fact

that resource allocation is a procedure that can be per-

formed dynamically according to the demand for ei-

ther the service or the user (Jamshidi et al., 2014).

Our convergent architecture is expected to increase

the number of network resources (e.g., routing ele-

ments, pre-processor nodes and gateways) in order to

provide and keep a service level agreement (SLA) be-

tween the user and the Internet architecture assembled

above the cloud. Virtual machine migration, addition

and resizing are techniques that could be combined to

offer an elasticity semantic for this novel Internet ar-

chitecture. Additionally, we envision the reduction of

subnet resources when the network demand is mod-

erated, so contributing to implement green comput-

ing with energy saving (i.e., consolidation technique,

shutting down VMs and the host node)

3 NovaGenesis (NG)

ARCHITECTURE

NovaGenesis (Alberti et al., 2014) project started in

2008 to address this question: imagine there is no

Internet architecture right now, how could we de-

sign it using the best contemporary technologies? We

selected several technologies to best implement No-

vaGenesis design principles, looking for deep syner-

gies among them. NovaGenesis can be deﬁned as a

convergent information architecture (CIA). By CIA

we mean an architecture that synergistically integrates

information processing and storage (as contended by

cloud computing), as well as information exchanging

(like the current Internet architecture or other emerg-

ing networks, e.g., software deﬁned networks or mo-

bile terminal networks). All the concepts presented in

this section are summarized in Figure 1.

3.1 Names, Identiﬁers, and Locators

The NG cornerstone is naming. A name is a set of

symbols that denote something, some existence. It

is deeply rooted in language. For example, one can

use the name “Paris” to denote a city in Europe. The

same name can be used to denote many different exis-

tences, e.g. “Paris” is also the name of a famous north

American socialite. In this context, an important de-

cision choice we did was: what existences need to be

named on a CIA? By existence we mean everything

that simply is. People love to name everything - from

cars to airplanes, applications to computers, photos to

movies, etc. Additionally, an important requirement

for future architectures is that they should be able to

better “understand” the meaning of the language used

by people - which is called semantic technology.

Many notable companies are investing on seman-

tic computing. Examples are IBM’s Watson and

Google’s Brain. With the advent of the Internet of

things (IoT) - where virtually anything can belong

to the Internet - we assumed that all possible enti-

ties could be named, bringing machines closer to peo-

ple natural language. Naming should be very ﬂexible

and broad. However, not all natural language names

(NLNs) are adequate for efﬁcient and safe naming.

Therefore, we adopted a second kind of naming in

TowardsanOpportunistic,Socially-driven,Self-organizing,CloudNetworkingArchitecturewithNovaGenesis

Figure 1: NovaGenesis concepts for convergent information architecture.

NG architecture. The so called self-certifying names

(SCNs). A SCN is typically obtained by passing a in-

put binary pattern by a hash function. The pattern can

be the entity itself, e.g., a chunk of data of a photo,

or a digital representation of some physical world at-

tributed, e.g., the digitalized patterns of a ﬁngerprint.

What is the use that NovaGenesis does for all

these names? We asked these question many times

while designing NG. The answer is: every informa-

tion processing, storage, or exchanging depends on

names and their relationships. The target of a com-

munication is an unique name in a certain scope.

The location of certain destination is also a name

that provides the relative distance among possible tar-

gets. Ownership, equivalence, “is contained”, and

many other semantic operators can be represented by

a name binding (NB). A NB can map several names

to many other names/objects. Additionally, one can

expect that people (and even machines in future) will

denote other existences by names. Therefore, name

bindings can represent the relationships (semantic op-

erators) among named-existences. In this sense, a

NB is itself another existence, a virtual/abstract one,

which can be stored as a virtual object.

NovaGenesis is generic enough to enable the cre-

ation of any naming structure. A naming structure

is an scheme to denote existences following some

planned strategy. For example, in the current Inter-

net hosts are denoted by an hierarchical name struc-

ture, where names have two portions: host and do-

main names, e.g., mycomputer.inatel.br. Using name

bindings the CIA architects can design any naming

convention. NG employs names to identify and lo-

cate communicating targets. All name bindings are

stored in a distributed software forming a giant name

bindings graph (NBG). Identiﬁcation and location is

a matter of scanning this graph to determine entities

that belong to some scope or that inhabit some space.

A communication target could be a content, a com-

puter program, a computer, or any other existence.

3.2 Substrate Resources, Services,

Contracts, Protocols, and Layers

Every software-based CIA (SB-CIA) is supported

by physical world existences called substrate re-

sources. Examples are antennas, ﬁber optics, mi-

croprocessors, memories, hard disks, etc. The ex-

ponential growth in computers capabilities is creat-

ing a phenomenon called virtualization. Maybe the

most prominent example is the so called cloud com-

puting where virtual machines (VMs) work like phys-

ical ones. More recently, virtualization on networking

technologies is being addressed under the banner of

network function virtualization (NFV) (Salsano et al.,

2014). The idea is to replace customized hardware

- many times deployed at difﬁcult access sites - by

software-implemented functionalities inhabiting VMs

in the cloud. NG assumes that computing hardware

is evolving so fast that software-based implementa-

tion of networking protocols is already possible for

the majority of the network stack.

The increasing role of software in ICT architec-

tures demand for excellence in software engineering.

A technology for this purpose is to design software-

as-a-service (SaaS). SaaS is often related to a service-

ESaaSA2015-WorkshoponEmergingSoftwareasaServiceandAnalytics

oriented architecture (SOA). We deﬁne a service as an

existence aimed at processing, exchanging, or storing

information. According to this deﬁnition, a computer

program (or a process) is a service. Any substrate

resource can be represented by named services, e.g.,

infrastructure-as-a-service (IaaS). Even protocol im-

plementations provide services.

In this paper, no distinction is done between ”pro-

tocol implementation” and a service. According to

our service deﬁnition, a protocol is implemented as

a service that processes, stores, and exchanges in-

formation in order to build networks. Thus, ser-

vices use other services indeﬁnitely, starting from the

ones required to implement a network. This paper

proposes the concept of protocol-implemented-as-a-

services (PIaaS). Observe that an interface to expose

any service to other services is required. NG enables

SCNs to be used for this purpose.

NovaGenesis envisions a service life-cycle that in-

cludes features exposition, peer discovery, negotia-

tion, contracting, monitoring, evaluation, and releas-

ing. All these steps take advantage of the NBG.

Service descriptors and contracts are named using

SCNs. A contract is deﬁned as a piece of informa-

tion that sets the limits, responsibilities, clauses to be

respected, as well as the criteria for completion and

punishment of services that were poorly executed.

The concept of a layer still prevails in NG. A def-

inition that combines the concepts of (Tanenbaum,

2003), OSI model (Standardization, 1996), (Day,

2008), and (Chaitin, 2010) is: a layer is an abstraction

for a cluster of services that is offered in a distributed

way to other services, isolating rules implementation,

following a shared language - a common syntax and

semantics - its interface. In this paper we replace the

terms: “protocol implementations that are offered in

a distributed way to other layers” by “services offered

in a distributed way to other services”. Thus, a NG

layer is composed by several services that are exposed

to other services via a predeﬁned language.

3.3 Current Proof-of-Concept

A ﬁrst implementation of the NBG and PIaaS con-

cepts, covering intra node and inter node service com-

munication was coded in 2012. We adopted a pub/sub

model, where NBs and contents are published and

subscribed by services. The NBG is implemented us-

ing distributed hash tables (DHT). NBs are published

by services and stored on DHT. Service life cycling is

build over this distributed pub/sub service. We imple-

mented some protocols for service exposition, discov-

ery, contracting, and named-content forwarding and

routing. The following services have been designed

for current version:

Hash Table Service (HTS) - It provides a domain

level hash table that is used to store published NBs.

Name bindings are categorized to improve scalability.

Generic Indirection Resolution Service (GIRS)

- It forwards name bindings (together with content)

for one or more HTS instances.

Publish/Subscribe Service (PSS) - It is the nar-

row waist for NG services. Any service will use the

publish/subscribe directives provided by the PSS hi-

erarchical service. Services can publish NBs and as-

sociated content to other services and subscribe other

name bindings of their interest.

Proxy/Gateway Service (PGS) - To facilitate mi-

gration and enable transport over other technologies,

we envisioned the PGS. The current PGS provides

software-based messages encapsulation, forwarding,

and routing. Additionally, the PGS is also a proxy for

core NG services inside an operating system (OS). It

represents these core services during bootstrapping,

forwarding public NBs to other friend PGSs, expos-

ing them to enable name-based self-organization. The

PGS can maintain inter node IPC without TCP/IP

only using Ethernet/Wi-Fi.

Application (App) - The current implementa-

tion has a generic application capable to explore the

pub/sub service (PSS) offered by the core.

4 ADAPTING NovaGenesis:

REQUIREMENTS AND

CHALLENGES

4.1 Rethinking Naming

The proposed architecture shall rely on a strong nam-

ing approach to capture relationships among entities.

Not only SCNs should be enabled for all existences,

but also natural language names (NLNs) to enable

in-architecture ontologies accommodation. NovaGe-

nesis already supports natural language and/or self-

certiﬁed name bindings to converge human and ma-

chine languages. This improves services expressive-

ness, like in current SOA, e.g., web services. People’s

names can be related to their equipment, services, and

content as illustrated on Figure 1. Distributed name

resolution enables services to explore other services,

users, and content context and scope. NovaGenesis

already implements this on PSS/GIRS/HTS services.

NG generic naming structure can support UCN, CCN,

ON, cloud requirements for naming.

TowardsanOpportunistic,Socially-driven,Self-organizing,CloudNetworkingArchitecturewithNovaGenesis

4.2 Addressing Heterogeneity

A PGS can be implemented to offer gateway func-

tionality to each technology deployed at the UCN,

ON, CCN, etc. For example, if ZigBee technology

is employed, a PGS can be implemented at the Zig-

Bee gateway to bridge frames from/to NovaGenesis.

The same can be done for every technology in the net-

work. We are expanding PGS to include software-

deﬁned control functionalities. This new service will

include proxy/gateway/controller (PGC) functionali-

ties. It can expose non NovaGenesis devices fea-

tures (and available conﬁgurations) to other NG ser-

vices. Hence, after contract establishment, other ser-

vices can publish conﬁguration change requests that

are translated to other technology devices, like Zig-

Bee, Bluetooth, etc. This allows NG services to con-

ﬁgure the network directly, creating what we are call-

ing service-deﬁned architecture (SDA).

4.3 Encouraging Collaboration

UCN provides the means for cloud and big data sys-

tems to be accessible to users, increasing their pro-

cessing and storing capabilities. Users’ devices can

be enhanced by being given the chance of exploiting

other users’ devices in the vicinity, and being able to

access the content that is made pervasively available

at the user’s current location. NovaGenesis SOA fa-

vors paid collaboration via dynamic SLA negotiation

and establishment. Free models of collaboration are

also possible. NG approach should be merged with

our previous work towards a convergent approach.

Previous work like SENSE (Ciobanu et al., 2014a)

can help clarify which are the requirements for en-

hancing collaboration. SENSE is a collaborative self-

ish node detection and incentive mechanism for mo-

bile networks where collaboration among users is a

must. Since information collected locally by each

node may not be sufﬁcient to reach an informed deci-

sion, nodes running SENSE collaborate through gos-

siping. After informing each other of their obser-

vations, nodes reach decisions individually based on

their local and received information. New NG ser-

vices will implement a range of collaboration models,

i.e. selﬁsh, paid, and free. Spontaneity cluster forma-

tion based on user labeling can also be implemented.

Services to determine node popularity are welcome.

4.4 Supporting Broad Opportunism

Our convergent architecture will require contex-

tualized opportunities detection, which can be

implemented as NG opportunity detection services

(ODSs). These services can be fully integrated

with NG PGCs to enable self-orchestration of

exposed substrate resources. The PGC services

expose physical resources (hardware) for software

orchestration. Opportunities can include proximity,

battery, ofﬂoading, spectrum, etc. All the required

information to expose and explore opportunities is

available for authorized services via PSS/GIRS/HTS.

The architecture needs novel solutions for data

aggregation. Since nodes running PGC services are

generally small hand held devices such as smart-

phones, their memory is limited. Moreover, the

access speed is also very important when there is

a contact between nodes, since the duration of an

encounter between two nodes (i.e. the time window

when they can exchange data) is relatively short, due

to the high degree of node mobility. Opportunities

can be detected and shared using ODS. NG approach

should be merged with previous work. Examples are

ULOOP (http://copelabs.ulusofona.pt/˜uloop/), UCR

(http://copelabs.ulusofona.pt/index.php/research/-

projects/past-projects/151-ucr). In other words,

a convergent name-based opportunistic rout-

ing/forwarding approach will be required. Regarding

proximity, not only physical world contacts can be

explored, but also service contracts. Opportunity

notiﬁcation can be done via PSS.

4.5 Seeing Everything-as-a-Service

NovaGenesis PGC services expose hardware re-

sources to software allowing all physical resources

to be seen as a service. Controllers, proxies, and

gateways can be exposed as a services. All the re-

quired orchestration for our convergent architecture

will be service-based. Service life-cycling is intrin-

sic, covering all aspects from exposition, discovery,

negotiation, contracting, content exchanging, quality

monitoring, and releasing. The goal is to accommo-

date protocol implementation as services, enabling

them to dynamic establish SLAs, giving rise to net-

working self-organization based on detected opportu-

nities. Flexible smart network services can discover

each other, prepare SLAs proposals, negotiate with

possible peers, establish SLAs, work together to ex-

plore social- and context-aware opportunities, evalu-

ate partners, and ﬁnish SLAs.

4.6 Security, Privacy and Trust

The distributed scenario behind the proposed archi-

tecture poses several challenges regarding security,

privacy, and trust. Networking cache, trafﬁc ofﬂoad-

ing, opportunistic collaborations, and cloud ofﬂoad-

ESaaSA2015-WorkshoponEmergingSoftwareasaServiceandAnalytics

ing are examples of user-centric approaches that will

require new security models. NG employs a pub/sub

communication model that favors the rendezvous

among authenticated and authorized services. Con-

tent exchanging only happens after SLA establish-

ment. Hence, it is secured by asymmetric cryptog-

raphy. Publishers maintain a secure association with

the PSS, which stores SCNs of authorized subscrib-

ing entities. Subscribers also have a secure associ-

ation with PSS. The PSS only delivers the content

after proper authentication and authorization. Addi-

tionally, the PSS provides revoking of published bind-

ings, data, permissions, etc. The SLA-based self-

organization enables the establishment of a trust net-

work among peer services - which is ideal for UCN,

ON, CCN, and clouds. All messages are conﬁden-

tial and have SCN-based integrity. We envision that

NG will need new services for trust network forma-

tion, assertion, and management, as well as services

for unbiased contract, reputation, and trust evaluation.

NG name- and contract-based ”social security” goes

beyond traditional mechanisms.

4.7 User-Centric Life-Cycling

User-awareness and social-behavior awareness need

to be estimated properly to drive NG ecosystem. The

aim is to adapt protocol implementations according

to user/social data. Hence, new services to estimate

social trends and achieve context-awareness will be

necessary. NG enables users to deﬁne high level poli-

cies that can be published to other services by a pol-

icy deﬁnition service (PDS). Published policies can

be subscribed by peer services and used in their de-

cision cycles. Big data and cloud information can be

shared together with policies to make all the environ-

ment user-aware/socially-aware. For example, imag-

ine a user agrees on selling part of this bandwidth to

other users’ radios when its device battery is charged

more than 50%. This policy can feed network level

protocols, in order to establish opportunistic routing

among users devices. Use cases and proper policies

should be designed and new NG services created to

implement UCN and content life-cycling. Social en-

gagement can be derived from available public in-

formation and explored by these emerging services.

Interest-driven is favored published ontologies, help-

ing on establishing successful partnerships.

4.8 Tolerating Delays and Disruption

Long delays and disruptive communication represent

a challenge for current Internet stack. TCP does not

ﬁt well on long delays scenarios and UDP requires

excessive application level programming. The asyn-

chronous communication model provided by the PSS

together with the HTS networking cache enable NG

services to change information in different time mo-

ments. Connectivity disruption does not impact on

sockets since a new naming structure is provided,

turning names perennial, independent of connectivity.

Protocols implemented as a service (PIaaS) offer the

required ﬂexibility to deal with intermittent connec-

tivity and long delays. These PIaaSs need to be de-

signed and implemented using NG software. A hard-

ware implementation is also possible, but will require

complete new designs using FPGA.

4.9 Supporting Mobility of Everything

What are the entities one expect to move in future

user-centric architectures? The answer we propose is

everything, from terminals, people, services, up to en-

tire opportunistic networks. NG naming structure en-

ables any name that satisﬁes some requirements to be-

come an identiﬁer or a locator. This decouples ”what

do you want” from ”where it is”. Mobility of every-

thing is supported by rebinding names during move-

ment. The identiﬁer of what is moving remains the

same - the only thing that changes are the locators.

This solution relays on NG distributed NBs pub/sub

and storage. UCN, CCN, NFV, and ON need mobil-

ity of everything support. NG can address this.

4.10 Integrating Cloud and Networking

Another important requirement is to alternate the use

of computing and communication resources inter-

changeably. In other words, the lack of computing

resources due to energy shortage or high load can

be compensated by communication resources, which

help on migrating the tasks to other machines. In con-

trary, when energy is limited (like in mobile devices),

functions can be virtualized in the cloud (providing

cloud ofﬂoading). Cloud networking is naturally sup-

ported. NG design should be merged with our previ-

ous work in cloud. Two approaches to be considered

are: context-aware cloud computing infrastructure for

taking good budgets. It is a SaaS that streamlines the

interaction between for customers and sellers (sales-

man); and resource provisioning on cloud computing

environments, which is an IaaS approach aimed to of-

fer load balancing for a service that runs parallel ap-

plications. It consists in task migration considering

the current pool of allocated processors. After that, if

the SLA is not satisﬁed, our second approach will be

to allocate new resources in order to run the applica-

tion with the previous established requirements. Also,

TowardsanOpportunistic,Socially-driven,Self-organizing,CloudNetworkingArchitecturewithNovaGenesis

it comprises the deallocation of resources if they are

super estimated for running the application. Finally,

the resources used by services can vary along their

lifetime in accordance with the application behavior.

4.11 Providing Self-Organization

The scenario we are imagining requires new ap-

proaches for management and control. One can not

expect people will manage or control dozens of de-

vices connected to others, manually. The current

management model depends on frequent human in-

terference, having poor scalability when considering

the swarms of devices we expect on next years. Em-

bedded control functions (usually, at equipment con-

trol plane) create a complex distributed states solu-

tion, which challenges operators to keep a coherent

and efﬁcient network conﬁguration. Future architec-

tures need to self-organize according to user needs

and detected opportunities. We envision a hierarchy

of control loops that follow user-awareness and so-

cial behavior awareness to autonomically conﬁgure

and manage ICT architectures. NG enables services

to self-organize using NBs and content pub/sub. Self-

management is fundamental in the proposed architec-

ture, since no one will be responsible alone to manage

a socially driven, possibly infrastructure-less architec-

ture. Or previous work on UCN, ON, and CCN can

help on designing these hierarchical control loops.

4.12 Control and Management

To address the requirements of effective utilization

and optimization of heterogeneous resources (stor-

age, processing, and networking), the architecture

requires an innovative software-deﬁned everything

(SDE) paradigm. NG addresses this challenge by

means of PGCs. A PGC represents some hard-

ware resource in the software layer and can estab-

lish dynamic contracts in the name of them. It also

controls the hardware devices, e.g., software-deﬁned

systems and/or radio, to perform the changes re-

quired. It can even expose hardware status to other

NG services, enabling them to proactively prepare

cloud/network solutions in advance to user require-

ments. It is a paradigm shift towards service-based

network control and management. NovaGenesis’s

PGC services will represent all physical world re-

sources used. Controllers- and managers-as-a-service

will establish contracts with PGCs to create a new

control/management model. Decision loops can be

formed by establishing chains of control/management

services linked via NG service contracts.

4.13 Supporting Context-awareness

User-, regulation-, situation-awareness, and many

other context-awareness features are required to make

sound decisions - decisions that consider the relevant

contexts to every situation. NG enables services to

securely and privately subscribe the relevant contexts

following their trust network. In other words, the PSS

provides a distributed networking cache from where

services can subscribe the relevant contexts for deci-

sion making. For example, a radio resource manager

could subscribe spectrum usage data from a spectrum

analyzer service. The manager can form a logically

centralized view of radio frequency spectrum usage

at some location, the so called situation-awareness.

Another example is related to social-awareness.The

dynamism of social behavior can be derived from big

data services implemented at NG or at any other soft-

ware. In the latter case, a NG service will be required

to bridge legacy software to NG cloud. Well estab-

lished social trends can be published by a big data

service (BDS) in order to feed other services.

4.14 Implementing Decision Cycles

The architecture shall adapt its behavior (cloud and

networking aspects) according to users needs and de-

tected opportunities, changing protocol implemen-

tations, data paths, parameters, etc. The architec-

ture should explore dynamically the available con-

nectivity, frequencies, bandwidths, technologies, and

nearby friends capabilities. It must conﬁgure itself to

take advantage of perceived opportunities in a reason-

able time. In long term, autonomic and cognitive de-

cision cycling will be required for self-management,

auto-piloting, and opportunistic networking. Special-

ized services could implement required decision cy-

cles. The cycle starts with user generated objectives,

policies, rules, and regulations. An existing plan is se-

lected. The plan is executed. The obtained results are

collected and analyzed to measure the degree of suc-

cess. If success was achieved, the objective is consid-

ered as met. Else, decision making can select chang-

ing the plan (or adapting it). This is an aim for future.

4.15 Addressing Society Challenges

Digital inclusion is one of the main aims of user-

centric, opportunistic, infrastructure-less (or public

infrastructures) architectures. It could be an impor-

tant driver for developing countries like Brazil and In-

dia. Also, social-driven proposal have the potential to

change our society towards ”smart solutions”, where

every resource is better used, optimized for inclusion,

ESaaSA2015-WorkshoponEmergingSoftwareasaServiceandAnalytics

and green technologies requirements. We believe ar-

chitectures like the one we are proposing on this paper

have the important role of helping us to solve impor-

tant social, economical, and environmental problems.

5 CONCLUSION

This paper proposed a convergent architecture that

integrates emerging socially-driven, opportunistic,

user-centric networking with NovaGenesis name-

based, software-deﬁned, information-centric, service-

centric, self-organizing cloud networking proposal.

Pre-requirements and open challenges regarding sev-

eral topics have been discussed. NovaGenesis prin-

ciples and current implementation provide a satisfac-

tory substrate to implement the proposed architecture.

NG joint orchestration of named-services and con-

tents provides an appropriated environment to imple-

ment socially-driven/opportunistic/cloud/networking

approaches as services. Even protocols are imple-

mented as services, enabling the resultant architecture

to react according to user-deﬁned policies, rules, reg-

ulations, environment situations. Context-awareness

can be included on decision making, changing proto-

col implementations according to social trends. The

paper is a ﬁrst step of an ongoing work that con-

tributes to the community by discussing how to inte-

grate so many relevant issues in only one architecture.

We envision that the proposed architecture can be

implemented by: (i) specifying new NG services that

meet the raised pre-requirements; (ii) adapting previ-

ous work techniques as new NG services; (iii) modi-

fying NG core services accordingly; or (iv) integrat-

ing already existing software (without any modiﬁca-

tion) with NG via proxy/gateway/controller. This ef-

fort is expected to result into a convergent solution

comprising the best of the considered architectures

(ICN, DTN, UCN) and that allows users to seamlessly

access content, and share resources anytime and any-

where in today’s dynamic scenario over their power-

ful personal devices. Future work include NG perfor-

mance, portability, and embedding on mobile devices.

ACKNOWLEDGEMENT

This work was partially supported by Finep/Funttel

Grant No. 01.14.0231.00, under the Radiocommu-

nication Reference Center (Centro de Refer

encia em

Radicomunicac¸

oes- CRR) project of the National In-

stitute of Telecommunications (Instituto Nacional de

Telecomunicac¸

oes - Inatel), Brazil.

REFERENCES

Ahmed, S., Raza, A., Asghar, H., and Ghazia, U. (2010).

Implementation of dynamic spectrum access using en-

hanced carrier sense multiple access in cognitive radio

networks. In Wireless Comm. Networking and Mob.

Comp. (WiCOM), 2010 6th Int. Conf. on, pages 1 –4.

Alberti, A., de O Fernandes, V., Casaroli, M., de Oliveira,

L., Pedroso, F., and Singh, D. (2014). A novagen-

esis proxy/gateway/controller for openﬂow software

deﬁned networks. In Network and Service Manage-

ment (CNSM), 2014 10th Conf. on, pages 394–399.

Alberti, A. M. (2012). Searching for synergies among fu-

ture internet ingredients. In Lee, G., Howard, D.,

Slezak, D., and Hong, Y., editors, Convergence and

Hybrid Info. Technology, v. 310 of Comm. in Comp.

and Information Science, pages 61–68. Springer.

Caini, C., Cornice, P., Firrincieli, R., and Lacamera, D.

(2008). A dtn approach to satellite communica-

tions. Selected Areas in Comm., IEEE Journal on,

26(5):820–827.

Chaitin, G. (2010). Mathematics, Complexity and Philoso-

phy. Midas.

Ciobanu, R.-I., Dobre, C., and Cristea, V. (2013). Sprint:

social prediction-based opportunistic routing. In

World of Wireless, Mobile and Multimedia Networks

(WoWMoM), 14th Int. Symp. on a, pages 1–7. IEEE.

Ciobanu, R.-I., Dobre, C., Dasc

alu, M., Tr

aus¸an-Matu, S¸.,

and Cristea, V. (2014a). Sense: A collaborative self-

ish node detection and incentive mechanism for op-

portunistic networks. Journal of Network and Comp.

App., 41:240–249.

Ciobanu, R.-I., Marin, R.-C., Dobre, C., Cristea, V., and

Mavromoustakis, C. X. (2014b). Onside: Socially-

aware and interest-based dissemination in opportunis-

tic networks. In Network Operations and Manage-

ment Symp. (NOMS), 2014 IEEE, pages 1–6. IEEE.

Conti, J. P. (2006). The internet of things. Communications

Engineer, Vol 4, 2006.

Day, J. (2008). Patterns in Network Architecture: A Return

to Fundamentals. Prentice Hall.

Jamshidi, P., Ahmad, A., and Pahl, C. (2014). Auto-

nomic resource provisioning for cloud-based soft-

ware. In Symp. on Software Eng. for Adaptive and

Self-Managing Systems, SEAMS 2014, pages 95–104,

ACM.

McKeown, N., Anderson, T., Balakrishnan, H., Parulkar,

G., Peterson, L., Rexford, J., Shenker, S., and Turner,

J. (2008). Openﬂow: enabling innovation in cam-

pus networks. SIGCOMM Comput. Commun. Rev.,

38(2):69–74.

Moreira, W. and Mendes, P. (2013). Social-aware oppor-

tunistic routing: The new trend. In Woungang, I.,

Dhurandher, S. K., Anpalagan, A., and Vasilakos,

A. V., editors, Routing in Opportunistic Networks,

pages 27–68. Springer.

Moreira, W. and Mendes, P. (2015a). Dynamics of social-

aware pervasive networks. In Int. Workshop on the Im-

pact of Human Mobility in Pervasive Syst. and App.,

2015 (PerMoby’15).

TowardsanOpportunistic,Socially-driven,Self-organizing,CloudNetworkingArchitecturewithNovaGenesis

Moreira, W. and Mendes, P. (2015b). Impact of hu-

man behavior on social opportunistic forwarding. Ad

Hoc Networks, 25, Part B(0):293 – 302. New

Research Challenges in Mobile, Opportunistic and

Delay-Tolerant Networks Energy-Aware Data Cen-

ters: Arch., Infrast., and Comm..

Moreira, W., Mendes, P., and Sargento, S. (2012). Op-

portunistic routing based on daily routines. In World

of Wireless, Mobile and Multimedia Networks (WoW-

MoM), 2012 IEEE Int. Symp. on a, pages 1–6.

Moreira, W., Mendes, P., and Sargento, S. (2014). Social-

aware opportunistic routing protocol based on users

interactions and interests. In Sherif, M. H., Mellouk,

A., Li, J., and Bellavista, P., editors, Ad Hoc Networks,

v. 129 of Lecture Notes of the Institute for Comp. Sci-

ences and Telecomm. Eng., pages 100–115. Springer.

Papazoglou, M., Traverso, P., Dustdar, S., and Leymann, F.

(2007). Service-oriented computing: State of the art

and research challenges. Computer, 40(11):38 –45.

Salsano, S., Blefari-Melazzi, N., Presti, F., Siracusano, G.,

and Ventre, P. (2014). Generalized virtual networking:

An enabler for service centric networking and network

function virtualization. In Telecommunications Net-

work Strategy and Planning Symp. (Networks), 2014

16th Int., pages 1–7.

Soﬁa, R. and Mendes, P. (2008). User-provided networks:

consumer as provider. Communications Magazine,

IEEE, 46(12):86–91.

Soﬁa, R., Mendes, P., and Moreira, Waldir, J. (2014). User-

centric networking: Living-examples and challenges

ahead. In Aldini, A. and Bogliolo, A., editors, User-

Centric Networking, Lecture Notes in Social Net-

works, pages 25–51. Springer.

Standardization, I. (1996). Iso/iec 7498-1: 1994 informa-

tion technology-OSI basic reference model. Int. Stan-

dard ISOIEC, 74981:59.

Tanenbaum, A. S. (2003). Computer networks, 4-th edition.

Xylomenos, G., Ververidis, C., Siris, V., Fotiou, N.,

Tsilopoulos, C., Vasilakos, X., Katsaros, K., and Poly-

zos, G. (2014). A survey of information-centric net-

working research. Comm. Surveys Tutorials, IEEE,

16(2):1024–1049.

ESaaSA2015-WorkshoponEmergingSoftwareasaServiceandAnalytics