SOHand

SOHand: an ontology-based platform for building services to exploit contextual handovers information

Project leader: Edson dos Santos Moreira
Intermídia Lab – ICMC - USP
Home Page: http://www.sohand.icmc.usp.br
Duration: 2007-2010

Introduction: the user is the centre

In future communication scenarios, mobile devices with multiple wireless interfaces will be able to seamlessly roam around, hoping from a network to another (and using different technologies) without losing their IP connection, as illustrated in Figure 1 below[1]. A combination of technologies which includes Local Area Networks (WLANs), Wireless Personal Area Networks (PANs), Cellular Networks (GSM, GPRS, UTMS) and Community Area Networks (WiMax, 802.11) will provide the infrastructure needed for this environment. One challenge faced for the designers of this scenario is to minimize the perception of the change in quality when a handover (the process of changing the point of network attachment) occurs. Handovers can happen between 2 domains using the same technology (horizontal handover) or from different technologies (vertical handover). Moreover, the vertical handover is further split in downward (when the handover allows for a decrease of bandwidth) and upward (when the new bandwidth is larger). Such a rich environment could bring up a new world of business possibilities and the proposal described here tries to create the proper conditions to exploit those innovations. The Service Oriented Handover – SOHand platform adds to the handover process the awareness of the ecosystem in which the event is embedded [14 ].

Figure 1 - A heterogeneous mobile overlay network[1]

The vision behind this proposal is that in the future mobile environments, such as the one above:

Project description

The figure 2 shows the overall architecture of the system, an abstract platform to support advanced network services. The basis of the system is a good, versatile, information structure, an ontology, which could be shared by the providers and users, a kind of Independent Register Profile – IRP. The ontology [15] will be the unifying technology, providing a common understanding of the terms and relationships which could be jointly maintained by the providers. The IRP would be an ontology server, populated with the user, application and environment contexts. This ontology could be beneficial for the deployment of innovative business models and, if linked to knowledge management engines, would provide important information for the corporate decision making processes from the providers.

The user side of the applications would contain agents able to access the environment sensors, the provider’s information and the IRP so the user could benefit from the combination and correlation of the information to enrich its experience. Multi-homed devices equipped with several interfaces can use the ontologies to assist the roaming, being able to dynamically influence the handover process (e.g. network selection). Reduced network management complexity both in terms of support and delivery would then be achieved.

The providers could use the information on the IRP to dynamically improve both the online experience of the users and the long term relationship with them. An empowering role is given to the users on their capacity of controlling the interaction with the net and having more choices, both in terms of connectivity and the services. In the handover events, the session entity will take care of the change of the policies, making the proper mappings and adjustments.

Uses of the Platform

1. From the network management perspective (the access provider):
• It may be interested on grabbing usual information which would be interesting to improve long term relationship with the user;
• It may collect information about the handovers of a user and relate them with positioning information;
• It may correlate information about routes and timing of accesses.
2. From the content provider perspective:
• It may adapt the delivery of media to specifics of devices, location, timing, type of user, etc
• It can understand the criteria by which the user chooses the access providers;
• It can explore the contextual information to add value to the content (advertisement, linking to 3rd party products, etc);
• As the user has the control over the mobility aspects, the access provider can focus on providing better and more varied services[4];
• It can provide brokerage services based on the common information available.
3. From the user perspective:
• It can choose the provider based upon several criteria:
• Best price
• Better response
• Best matching to his/her requirements
• Economy (use a home user-owned WLAN, during a traffic jam)
• It can use contextual information to adapt the user device profile of usage:
• Power management (CPU, memory, display, network interface)
• Streaming control (proximity to known blind spots such as tunnels)
• Optimization of content delivery to a new profile, on vertical handovers with a lower bandwidth provider.

Figure 2 - The archtecture and relationship between the entities

The Relationship with PROTON

A Mobile IPv6 environment, connected to a Vodafone’s GPRS network, has been set up at the Computer Laboratory of the University of Cambridge in an effort to demonstrate a 4G mobile scenario[1]. The main aims for the testbed were to assess the performance issues involved in the wireless overlays in a heterogeneous environment and to improve the seamlessly capabilities of the handover process. Figure 2 shows the system implemented at the Computer Lab.

Figure 3 - Structure of Wireless MIPv6 Testbed [5]

PROTON (Policy-based system for ROaming Transparently among Overlay Networks) has been devised as a system which would assist the handover process, easing the complexity of the management, resolving conflict issues on the decision process and improving the general performance of the systems which makes the handovers more seamless. The capabilities added by PROTON to the tesbbed made it to have most of the characteristics being advocated by autonomic computing [3]. It is based in a novel model built on Finite State Automatas using a metric called Tautness Function and bringing up a new type of automata called Finite State Transducer with Tautness Functions and Identities (TFFST). This model was used to resolve, with good run-time performance, potential conflict resolution in policy based systems[6]. To cope with the processing/storage demands, PROTON is split in two parts, network-side and host-side. The architecture of PROTON as implemented is shown in Figure 3. The context is set up by incoming data collected by the sentinels (in charge of dynamic data acquisition: positioning, speed of displacement, adverts of network availability made by the providers) and the retrievers (responsible by the stored information context). This information is combined with the ones sent by the providers (regarding the availability of access points) and the policies (the rules which govern the handovers) by the Conflict Resolution Module which will then provide the elements to be enforced by the executors of the handovers.

In this way, PROTON looks at low-level contexts to improve the performance of the handover process.

Figure 4 - PROTON's architecture [6]

With this new platform, SOHand, we propose that the handover be influenced (and influences, in turn) high-level contexts, more detailed and rich in semantics, which would widen the possibilities of building more versatile service business models in this environment. The scope of the sentinels would be enlarged to get more information about the whereabouts of the user device. The role of the retrievers will be expanded to have information about past experiences (good and bad ones) of the user navigation, for instance. Additionally we understand that a new entity, the Independent Register Profile (IPR), or the Ontology Server in figure 2, will be necessary so to store common information (mainly definitions) which could be shared by the providers (to map different parameters of policies) and users. The instantiation will be mostly made and stored by the providers in order to protect the specifics of their business models.

The structure of the Ontology

A service is a facility (a video streamer, a voice channel or a game application) which a content provider offers, during a session, to a user through an access provider. One entity can offer both access and content at the same time. In the course of enjoying a service, the user seamlessly roams through a net of access providers. Context and Handover information, gathered by positioning sensors and other sources from the user device or by any other related service, can be used both by the session to frame Security, Privacy, QoS and other policies. Positioning information can also be used for the definition route patterns used for one user, which could lead to better pricing strategies for the user. Some sort of SLA is signed between all the entities involved in order to offer to the user some parameters by which they can measure his/her Quality of Experience while using the system – Figure 5.

Figure 5 - Classes and relationship

This ontology will allow for:
• The creation of a common vocabulary of terms which would easy the design and reuse in new services in the communications industry, with faster deployment and exploration in the added value chain;
• Definition of complex relationships between the terms which would make possible to correlate the business processes, exploring new possibilities derived from the positioning and context awareness technologies as well as from security/privacy policies;
• A structured integration of the Access Networks, Subscriber Profiles, Applications and Data both by a provider, or by a group of providers;
• Other ontologies in the IT management domain can be imported, increasing the management boundaries.
Each provider will have its own policies about IT Management. Some of these policies can be shared between providers to deliver ubiquitous services, and others can be protected or hidden for business reasons.

The Classes

The following classes are thought to be needed to manage the seamless roaming of a user connected to a service (for instance a video server) and using several access providers on the way. The key issue here is the possibility that the information in the Ontology can be shared by several entities involved in providing the services, in a service-oriented way. Parts of the ontology will be distributed between the following Profile Registers:
• Access Provider
• Content Provider
• The User Device
• The Business Device

User. This class will contain the profile of the user (name, address, age, etc) , as well as the information needed for authentication

User Device. This class defines the way the user access the net. Capabilities parameters (screen, network adapter, bandwidth, sensors, etc) are important information for the proper running of the service.

Business Device. This class will be the provider’s counterpart of the User Device. It will be responsible to gather information from the different entities to form the knowledge necessary for the continuity and improvement of the business.

Access Provider. This class presents information about the provider which are given the access to the user. The information here must be rich enough to allow for the identification of good routes and to troubleshooting eventual access incidents. The user must have the information needed to access the perception of quality of individual providers.

Content Provider. This is the class for registering information about the provider who are offering the content.

Service. This class presents information about services which are in the portfolio of providers, with the characteristic (QoS, privacy, security) needed for its run. Data regarded to intellectual properties of the content are stored in here.

Session. This class stores information about the period of time a user is connected to a provider’s network for accessing a service. The policies, context information and the handover information will be used to adapt the provisioning of the access.

QoE. Information about the Service Level Agreement (SLA) signed with the content and/or access provider. It can also have information about the user’s experience on using the system. It will be managed by the user and by the providers in a shared way.

Positioning. This class is aimed at providing information on the localization of the user, as well as for the history of his/her roaming. Eventually this kind of information can be shared with other positioning-based services (see the context class). The patterns of tracing of a service could give valuable provisioning information for capacity management.

Context. The positioning is not the only contextual information important for the user’s navigation. Environmental data (who is around, is the user indoors or outdoors or in a car? is it cold or warm, is it a public place?, which kind of device is used?) can be useful for definition of sensitive services, regarding privacy, for instance.

Handover. This class contains handover information about the session.

Security. This class contains information about the way the provider deals with security. It can contain the restrictions derived from the providers Security Policy and the terms agreed on the SLA. It should also present relevant information so that the security profile can be mapped when the user crosses borders of domains

Privacy. This class will contain personal information about the user (preferences, agenda, etc). It will be managed by the user and, eventually, adjust itself automatically.

Methodology and adherence to standards

We have particular preoccupation on following established or emerging standards in the implementation of this project. The framework is based on SOUPA (Standard Ontology for Ubiquitous and Pervasive Applications)[7]. This is an interesting model which divides the system in core ontologies and extension ontologies – Figure 5. SOUPA Core defines wide range terms and relationships that are of general use for different ubiquitous applications. SOUPA Extension defines vocabularies for specific types of applications. We understand that, for this platform, the core ontology will take care of the common vocabulary which will be used by all the entities belonging to the environment. Some of the extensions will be designed by the providers in order to create their own private vocabularies, most of them will not be available to the other (most of the times competing) entities. The overall methodology which will be used for ontology development is the Methontology[8].

Figure 6 - SOUPA Ontologies: Core and Extension [7]

The ontology will be described in OWL[9] (Web Ontology Language), which is the prime language created for the Semantic Web. OWL is incorporated on the Protégé tool[10], which is a good platform to design ontologies. The Jena API[11] is used by Protègè-OWL for various tasks during the development and prototyping of the ontology and its applications.

The development of the applications that use the ontology will be based on Service Components, a set of programs and data that executes functionalities which are relevant in the context of a business. In this model, the activities can be set as having technical meaning (the updating of a table in a database) or a business meaning (the update of a customer address). A model which is emerging as a good standard for developing applications is the Service Oriented Architecture (SOA)[12]. SOA-based applications make available interfaces for other applications via service components. Through the pipelining of multiple components via request/reply remote calls more complex composite applications, a logical module in a larger business model, can be constructed. Another model, the Event Driven Architecture (EDA)[13], defines a model of developing application components which exchange events to implement business functions. It differentiates itself from SOA as in EDA all the components keep working, processing messages, while awaiting for response of a previous calls, while SOA will block until the query is answered. There are long arguments about the preferences on using SOA or EDA model, or if any of them should provide the right paradigm, however, we intend to follow at least the general ideas from them in developing the prototyping applications to test SOHand.

The environment has been installed composed by APs from CISCO, 4 PDAs from HP (5550), 2 notebooks, 3 GPSs, 2 Switches 3COM and 4 Desktops. Linux have been used on the notebooks and windows on the PDAs.

Funding

We have been granted funding from FAPESP for to build the basic infra-structure for the project, namely Acess Points, Switches, PDAs (802.11, GSM & Bluetooth), Notebooks, Desktops, GPSs, etc. Further on we will apply for extra funding from other sources to acquire WiMax systems.

Planned Activities and Deliverables

Following is the updated list of activities and deliverables planned for the project:
1. specification and buying of the equipment and definition of the environment and software tools to be used (April to June, 2007);
2. installation of the access points, the gateway with the fiber core and configuration of the mobile devices with haggle (a PSN prototype in construction by an European team (July to August, 2007);
3. follow up with the installation of the 4G testbed at the Computer Laboratory, Cambridge University (April to August, 2007);
4. the installation of a video server which will be used as a Content Provider model (July 2007). The server has been mostly designed, built and functionally tested during 2005-2006;
5. the design of an ontology for context-aware management of handovers. During 2006 a good effort has been made to get the ontology’s specification done (April 2007 to February 2008);
6. preliminary experiments on handover performance (September 2007 to February 2008);
7. specification and acquisition of WiMax capabilities for the project (Jan to Aug 2008);
8. experiments on video traffic distribution measurements. A focus point in here is to study the fiber-radio relationship. One characteristic worth to mention is the fact that the TCP/IP protocols deal with packet retransmission as a sign of congestion (causing the applications to slow down) whereas in wireless networks the retransmission could just be caused by momentaneous package loss in the air, not requiring big concern from the applications. We plan to compare behaviors of flows of video streams in wired and wireless accesses (October 2007 to September 2008);
9. programming of the management system based on the ontology (January 2008 to December 2008);
10. design and development of a few prototyping applications due to show the usefulness of the perception-driven network management (July 2008 to July 2009);
11. performance analysis and system tuning (January 2009 to December 2009)
12. Experiments and elaboration of papers on traffic engineering, handover management, video distribution, security and privacy (2, during 2007, 6 during 2008, 8 during 2008 and following years);
13. MsC dissertation (1, by the end of 2008, 2, by the end of 2009, 2 by the end of 2010); PhD thesis (1, by the beginning of 2009; 2 by the end of 2010);
14. Special courses and Seminars will be given throughout the duration of the project.

References

[1] P. Vidales, "Seamless mobility in 4G systems," in Computer Science Cambridge: Cambrigde University, PhD thesis, 2005, p. 141.
[2] L. B. Patanapongpibul, G. Mapp, and A. Hopper, "An End-System Approach to Mobility Management for 4G Networks and its Application to Thin-Client Computing," in ACM SIGMOBILE Mobile Computing and Communications Review, 2006.
[3] L. Stojanovic, J. Schneider, A. Maedche, S. Libischer, R. Studer, T. Lumpp, A. Abecker, G. Breiter, and J. Dinger, "The role of ontologies in autonomic computing systems," IBM Systems, vol. 43, 2004.
[4] D. Chalmers, M. Chalmers, J. Crowcroft, M. Kwiatkowska, R. Milner, E. O'Neill, T. Rodden, V. Sassone, and M. Sloman, "Ubiquitous Computing: Experience, Design and Science," Grand Challenge web site, UK 2006.
[5] G. Mapp, D. N. Cottingham, F. Shaikh, P. Vidales, L. Patanapongpibul, J. Balioisian, and J. Crowcroft, "An Architectural Framework For Heterogeneous Networking," in International Conference on Wireless Information Networks and Systems, 2006.
[6] P. Vidales, R. Chakravorty, and C. Policroniades, "PROTON: A Policy-based Solution for Future 4G devices," in Fifth IEEE International Workshop on Policies for Distributed Systems and Networks, 2004, pp. 219-222.
[7] H. Chen, F. Perich, T. Finin, and A. Joshi, "SOUPA: Standard Ontology for Ubiquitous and Pervasive Applications," in First Annual International Conference on Mobile and Ubiquitous Systems: Networking and Services (MobiQuitous'04), 2004, pp. pp. 258-267.
[8] M. A. Fernández, A. Gómez-Pérez, and N. Juristo, "METHONTOLOGY: From Ontological Art towards Ontological Engineering," in AAAI Spring Symposium, Menlo Park, California, 1997, pp. 33-40.
[9] L. W. Lacy, OWL: Representing Information Using the Web Ontology Language: Trafford, 2005.
[10] Protègè, "http://protege.stanford.edu/," 2006.
[11] "HP Labs Jena 2 Toolkit, http://www.hpl.hp.com/semweb/index.html."
[12] S. Jones, "Toward an acceptable definition of service [service-oriented architecture]," Software, IEEE, vol. 22, pp. 87-93, May-Jun 2005 2005.
[13] B. M. Michelson, "Event-Driven Architecture Overview: Event-Driven SOA Is Just Part of the EDA Story," 2006.
[14] MOREIRA, E. S. ; D. N. Cottingham ; CROWCROFT, J. ; Hui, P. ; Mapp, G.; VANNI, Renata Maria Porto . Exploiting Contextual Handover Information for Versatile Services in NGN Environments. In: International Conference on Digital Information Management, 2007, Lyon. Proceedings of the International Conference on Digital Information Management, 2007.
[15] VANNI, Renata Maria Porto ; MOREIRA, E. S. ; GOULARTE, Rudinei . DOHand: An ontology to support building services to exploit handover information in mobile heterogeneous networks. In: I2TS - International Information and Telecommunication Technologies Symposium, 2006, Cuiaba. 5th International Information and Telecommunication Technologies Symposium, 2006. v. 1. p. 105-112.