UTF-8 http://feraios.blogspot.com/feeds/posts/default

Monday, September 17, 2012


Position paper on the Core Platform needed to support the Future
Internet Projects

1 Motivation The European aspiration of creating the most competitive and dynamic knowledge economy and  rapidly raising standards of living must be based on the Future Internet being integrated in the  business processes of all industrial sectors with a special focus on helping new and innovative  SMEs and start-ups to succeed by providing a more flexible and less costly IT infrastructure and  SW. This will enable these sectors to be more efficient in terms of using resources, reductions of  emissions, improving their business processes and helping people.
To achieve this, there is a need to confront, understand and support the expectations and
requirements of the different sectors in order to deliver a common, standardized, efficient Internet.
This challenge requires bringing together the competence of the application sectors (Usage Areas) and the ICT sector to work out common solutions from an end-to-end perspective. This entails a  multidisciplinary and integrated approach, where massively distributed services and applications  are run over large scale and secure internet platforms. This is the only adequate way to deal  effectively with the increasing complexity of intertwined application and service demands and still  ensure compatibility and interoperability of the solutions.
The EFII has analysed a number of representative “Usage Areas” and determined sets of
applications that may share common domain expertise, technological and services requirements so  that they can, in turn, share framework, technologies, generic common enablers and architectures  in the provision of working systems.
This concept of maximum commonality across application sectors is crucial to facilitate the creation  of a critical mass for advanced services and the creation of European-scale markets for smart  infrastructures, with integrated advanced communications functionalities.
Hence, The EFFI is proposing to develop and deploy a platform that instantiates a unified and consolidated open architectural approach that globally enables the creation, deployment and  execution of applications by using hardware, software, network enabling capabilities etc. This is  what we refer to as the Future Internet Core Platform.
2 Overall Approach and Rationale
The Future Internet must be based on shared frameworks, technologies, common enablers and  architectural principles. These will be defined through the Future Internet Core Platform (FI-CP)  concept. This is built on the need to evolve from a set of communication-centric, content-centric,service-centric, resource-centric isolated infrastructures to a polymorphic infrastructure, where the  boundaries between systems are more flexible and allow blending, and where the emphasis is on  the integration, interrelationships and interworking of their elements through new service-based  interfaces
Future Internet enabled Applications will typically be built on top of Future Internet Core Platform  Instances (FI-CPI). A set of Generic Enablers is provided with the aim of building such instances in  an easy and affordable way, and with the intent to support sharing of the same architectural  principles across all instances, enabling interoperability between applications running on top of  different instances.
Generic Enablers, selected for a given platform instance, are configured and “glued” together to  provide an appropriate set of functionality. Future Internet Application Providers can use standard  interfaces provided by such a Platform to deploy and monitor the type of applications for which the  platform was designed.
The benefits of such an approach would be to:
• Analyse and classify the needs from Usage Areas and identify commonalities
• Define open interfaces between enablers and assembled functions
• Develop or integrate functionalities which meet the needs of Usage Areas according to a
sound methodology
Such a Future Internet Core Platform would not be built from scratch but would be assembled on
an iterative basis to meet design specifications coming from Usage Areas. This assembly will also
make use of existing results coming from various Framework Programmes, National Initiatives or
companies´ internal developments.
3 The Core Platform Architecture
3.1 Basic Architectural Principles
The Future Internet Core Platform provides the means to support many different applications and  business processes, which may have different requirements on the Core Platform capabilities.
Therefore, different instances of the Core Platform can be established comprising different Generic  Enablers (GEs). However all instances should adhere to a common architecture and some basic  architectural principles.
Each Future Internet Core Platform Instance will be created based on the combination of a number  of the Generic Enablers defined above, targeted to cover the needs of Future Internet Applications.
These GEs should be designed to support interoperability of applications as well as portability of   a given application across alternative implementations of the supporting FI-CPI and allow the  evolution and adaptation of the each FI-CP and FI-CPIs to the changing needs of the Usage Area  and the new advancements in the state of the art.
Portability and Interoperability  Enablers available for the creation of FI-CPIs will provide open interfaces that applications will  employ in order to benefit from the Core Platform services and functionality. They may also require  that applications provide some data in compliance with some metadata definitions. Therefore,
• A common communication model will be defined based on the requirements of the usage
areas (e. g. Quality of Experience, mobility and security issues) and operational constraints
following the next generation network architecture (data control, management plane).
• A unified set of interface specifications – based on the communication model and where
needed and appropriate- should be defined that will be adopted by all GEs and FI-CPIs.
• Interfaces must be invoked in a standard manner to ensure a sound model and a coherent  usage of enablers.
• Open standard metadata specifications have to be defined for all those pieces where
standardization of metadata will be required to support portability and interoperability (e.g.,service specifications to be used during marketplace registration)
In order to meet its objectives, the FI-CP will have to address this issue in a pragmatic way, selecting and promoting those solutions that best solve the requirements coming from the usage  areas and minimizing fragmentation in terms of the GEs implementation and integration. Service  orientation, Web oriented architectures, REST vs. WS-*, commonly accepted ontologies and XML  standards, new solutions proposed by many parties like USDL will all be analysed and adopted  where appropriate.
Evolution and adaptability
A FI-CPI can be considered as a composition of enablers into a platform that supports a particular  family of smart FI applications and services. This composition evolves over time, i.e. enablers can  be added, removed, reconfigured, and managed in a continuous platform life-cycle. From an  engineering perspective the focus must therefore be put on supporting the controlled evolution and  change of FI-CPIs over time.
4 What will the Core Platform Offer – Initional Set of GEs To tackle the provision of services at a global scale, FI-CPIs might support the following principles at global level:
• Accessibility: Every entity in the Future Internet must be made accessible anywhere-anytime.
This is related to ubiquitous connectivity.
• Identification and Naming: Every entity in the FI must be uniquely identifiable.
• Run-time explorability: Every entity in the FI must be self-describing. This is essential to allow  the exploration of entities at run-time – not only at design time.
• Data: All data exchanged between applications must be self-describing.
• Trustworthiness customization: every entity must be able to adjust or to expose adjustment  rules regarding security, privacy, integrity, compliancy and certification aspects based on the  interactions in which it operate at run-time in a global level
• Non-Functional Aspects: Elementary non-functional aspects such as quality, governance,
accountability, resilience, availability, and integrity must be properly supported in various
degrees by entities in the FI globally.
The FI Generic Enablers Development activity will focus on how the above core platform objectives  can be implemented my means of a core set of Generic Enablers. The architectures, interfaces and  information flows will also be defined to allow these GEs to be developed, integrated and managed  to create FI-CPIs. The intention is to standardise the interfaces to encourage global take up of the  results of this initiative.
The identification of the highest priority Generic Enablers in accordance with the overall goals and  objectives of building the Core Platform will be based on which enablers can support the most  generic sector requirements. Obviously the enablers that are critical to all sectors will be highest  priority.
The iterative exercise to identify, define and characterize the generic enablers is already taking  place. The emphasis is to selecting a realistic set of Enablers and to find pragmatic ways to  implement them in the given time frame. Use of available results from previous research work is  part of the strategy.
The following sections provide more detailed information about the overall vision and motivation  behind some of the Generic Enablers identified so far.
4.1 End-user access, adaptation and composition of Resources This GE encompasses specific enablers responsible for empowering users in the usage and
adaptation of the different available resources. By resources, we mean any end user valuable  software or content, that is data sources (RSS, OpenData, web contents…), Services (WS*, REST,IMS, Telco Services, … ) or Applications (Web Clients, Gadgets,… )
In other words, the ultimate goal of this GE is to offer the necessary technologies to allow endusers  to access and adapt every resource to any delivery context, situation and process, taking into  account a wide range of scenarios covering any daily-life common tasks and the fusions of roles  users take at any moment and any place (work,entertainment, family, home etc.)  End-users of any application must be able to act as prosumers, creating new content and  applications and sharing their knowledge in well controlled crowd-sourcing environments.
Technologies that empower end users to create and compose new resources by means of lowering  as much as possible the technological barriers will be developed and integrated in this GE. At this  moment Mash-up technologies applied to Data, Services and Applications are the best candidates  to fulfill this need.
Moreover autonomous helpers are desired, able to solve generic issues that could appear during  the creation or exploitation phase, such as context awareness, including but not limited to delivery  context (device, network etc.), user context (identity, preferences, profile role, and social  information), environment context (location, sensor information, “things” offering information or  services etc.) or situation (time, date, season, at home or at work, on vacation, on a business trip,mood etc.)
4.2 Service Handling and SOA Support This GE will encompass components in charge of providing all the basic services required for a  next generation SOA approach, where all computing and networking infrastructure resources, IoT  elements, devices, SW functionality etc. are encapsulated and handled and offered as services, both internally within the platform and externally to the Usage areas.
Representative functionalities offered by this GE are (among other possible candidates):
• Uniform naming resolution for any Entity in the platform (services, things, devices, nodes, resources, etc.) that supports both public and private spaces, together with the ability to  handle large number of entities (millions)
• An efficient, scalable and distributed Service Repository that will also support management  and federation of various domains (public and private repositories).
• A common Application Communication Infrastructure that supports multiple paradigms,
such as request/response, publish/subscribe, multicast, etc., as well as transactional
• Meta-data indexing and inference to enable Entity/Service Indexing, Searching and
• Modelling, Composition, orchestration and execution of business processes and telco
4.3 Event and Data Management and Processing
This category will encompass all those enablers whose functions are related to handling,
composing, transforming, classifying and, in general terms, managing all kind of data and events:
• Event / data aggregation, correlation and filtering -This enabler will allow the aggregation
and hierarchical handling of data and events (e.g. event filtering and aggregation at
different levels).
• Event Transformation – This enabler deals with the inference of new events and/or
semantics; application of syntactic pre-processing etc. Real-time and data mashups will
also be elements related to this enabler.
• Data classification - Functions such as processing data, clustering and usage of semantics
on classification are provided by this Generic Enabler.
4.4 Multimedia analysis and semantic support 4.4.1 Multimedia analysis
We live in a multimedia world, where we daily interact with both professionally and user-generated  content. In particular, the amount of user-generated multimedia content has increased  exponentially in recent years, creating new challenges in multimedia analysis and the pervasive  problem of digital media overload, where users have a very hard time finding desired content in  their ever increasing repositories of personal media. In addition, an increasing amount of  multimedia content is stored in the ‘cloud’, thanks to pervasive Internet broadband access and a  steady reduction of storage costs.
In this scenario, new multi-modal approaches to multimedia analysis are needed, including novel  methodologies to fuse multi-modal multimedia content with its associated contextual information  (e.g. location, metadata, tags, user feedback, etc…). Automatic tagging techniques and innovative  interfaces to encourage users to add context to their content will enrich the multimedia information  and improve the results of multimedia search engines while, at the same time, increase the  knowledge about the users and hence enable personalization and recommendations (discovery of  new and relevant multimedia content). Given that most of the multimedia content is (and will  be) user generated, there are additional challenges that need to be tackled when building  multimedia management systems, including: efficient near-duplicate image and video copy  detection; algorithms and user interfaces for multi-modal multimedia summarization; computational
models of the aesthetic value of multimedia content (increasingly important due to the large  variance in the quality of user-generated content); and novel ways to create rich multimodal user  experiences by experimenting with new paradigms of consuming multimedia.
Another aspect that can mark a significant evolution is the analysis of Digital Footprint - large-scale  human behavioural data. With this challenge we refer to capture, storage, visualize and analyze  large-scale multi-modal data coming from different source: Internet of Things, sensors, social  networks, mobile users geo-location, etc. From these data and with new algorithms for inferring  and predict human behaviour, CP can offer valuable information for building smarter applications in  all the usage areas.
The Core Platform will need to provide algorithms for multimedia content analysis that would tackle  the previously stated challenges, enabling content filtering and aggregation at different levels,personalization and creating unique user experiences suited to the needs of its users.
4.4.2 Semantic Support
One of its most outstanding characteristics of FI applications will be their capability to understand  the meaning of the information exposed or interchanged. This capability, which will be provided by  the so called Semantic Technologies, will make it possible to build a more human Internet, diluting  the current communication barriers between software systems and their users, as well as more  intelligent applications we are not even able to imagine nowadays. In this sense, Semantic  Technologies will revolutionize Internet as we know it today.
To be aligned with the previous statement and predictions, the Platform will need to include these  semantic capabilities from the core so that any other additional feature to be included in the  Platform directly benefits from them. These semantic capabilities will include not only mechanisms  to generate the semantic information, including mechanisms to automatically extract meaning from  already existent content as well as manual mechanisms to involve the final users in the semantic  annotation of the information they manage, but also mechanisms to store and to exploit the  semantic information available.
4.5 Preferences, profiling and context
The society is evolving to a more and more such a heterogeneous ecosystem: in order to cover all  the needs and provide goods and services customer-tailored, the knowledge of users’ environment  and preferences is a key factor. In this way it aill be possible to satisfying demand‘s diversity and  creating new business opportunities.
This enabler collects big amount of data from users’ activity and after applying ETL process and  specific Analytic Models, it infers information about users like users’ age, mobility patters, tastes,social interaction, etc. This valuable information is an input to the Recommendation Engine  together with contents provided by external suppliers. The results of these recommendations are  provided to services and applications located in the Usage Area.
Profiling allows services and content to be tailored automatically to the end-user's preferences (e.g.content consumption in the past) as well as the user's context (location, current connectivity, mood,etc…). Personalization is not just limited to on-demand TV, video or music, but also for life experiences like social interaction, education, leisure, shopping, etc.
The Context Awareness is much more than location awareness alone, or merely the immediate  situation. The vision is to have a middleware infrastructure capable of collecting all user’s context  information from devices, sensors, database and from the environment in general, in order to  decide what is worth for the user depending on her/his preferences and instantaneous state.
Privacy policy will be implemented for retrieving and analyze users’ information.
Main objectives of this working work would be to define (among others):
• A platform independent & distributed approach encompassing all types of services.
• A support infrastructure that is aware of the user’s situation (location-based information,
mood, etc.).
• Analytic algorithms for modelling users’ behaviour based on SNA technologies, Neural
Networks, etc.
• A recommendation engine supported by a Personalisation Engine as common enablers for
services and applications deployed in the Usage Areas.
• Additional mechanisms for managing users’ privacy and controls the information they
desire to provide to the platform, preserving rest of data.
4.6 Identity, Privacy, Confidentiality
Creating trustworthy and secure applications that provide at the same time an appropriate end-user  experience will greatly depend on the FI-CP offering a shared framework that will span  applications, services and enablers, giving easy and uniform access to authorization, authentication and identification features.
Under this initial perspective, advanced functionalities must be provided to solve issues like account management, single sign on, federation, profile and personal context sharing in social  environments, establishment of privacy and access policies. A common Identity management  architecture will improve the integration of all the core platform enablers, helping in the "one  platform" user perception. A common identity management framework should also be responsible of handling privacy constrains and all kind of legal issues related to personal data protection.
Convergence of IT standards and technologies and in this area with exist Telco capabilities (SIM  Cards, trusted connections between devices and network, established trust relationships with end  users etc.) will also be a focus of the work, in order to achieve the best possible compromise  between security and user experience.



Post a Comment

<< Home