1. Project Summary (Contract Preparation Form A2) *

2. Project Objectives *

3. Participant list *

4. Contribution to programme/key action objectives *

5. Innovation *

6. Community added value and contribution to EC policies *

7. Contribution to Community social objectives *

8. Economic development and scientific and technological prospects *

9. Workplan *

Appendix A - Consortium description *

Appendix B - Contract Preparation forms *

Project Acronym 2

ANDROID

Proposal No 3

IST-1999-10299

A2.	Project Summary 20

Objectives (maximum 1000 characters)

The principal objective of this project is to prove the feasibility of providing a managed, scalable, programmable network infrastructure capable of supporting the future needs of the Information Society. This will be based on active networking. To fully realise the benefits of an active network infrastructure, a flexible management system is required, combined with a common information model for resources - both processing and network. The management of active networks, the extent to which management information must be shared, and the communication mechanisms required by the management system are currently open research issues. The detailed scientific objectives of the project are to develop innovative solutions to the open research problems identified above, and test the solutions through a combination of implementation experiments and modelling.

Description of the work (maximum 2000 characters)

This project will use an existing, state of the art, active network platform prototype. This is component based, and consists of network machines such as routers and associated processing platforms. It supports the dynamic addition of software components such as programmes, protocols, and policies. The project will develop innovative extensions to improve the manageability and scalability of the platform, including extensible multicast protocols, dynamic resource discovery to assist with mobility, information sharing protocols providing reliable efficient delivery, protocols providing security, policy exchange protocols and message and policy handling at active nodes. It is expected that many of the above will form the basis of new standard proposals. No existing active network implementation has more than a rudimentary management system. The project will design, build and test a suite of policy driven management components to overcome this deficiency. It will consider system and management architecture, information models and component models, management tools and mechanisms, security components, billing tools and mechanisms and middleware support for management. The project will be primarily concerned with resource management and efficient information transfer. It is expected that the most significant innovations will be the management architecture and the proposed policy sets. The policies are expected to influence providers and management fora such as the TMF. There should also be significant innovations in the areas of policy syntax and policy interactions, both of which are currently unresolved research issues. An integration and testing phase of the project will show for the first time that it is possible to build an advanced active infrastructure, integrated with an active management system, that is effective on a wide-area network. This will significantly influence early deployment of public service networks based on the technology.

Milestones and expected results (maximum 500 characters)

Identification of key mechanisms and management functions required. Novel architecture for active networks and their management. Proof of benefits of extensions to existing basic active network prototypes. Proof of feasibility of policy-based approach to active network management. Proof of benefits of modular multicast stack Predictions of system performance, novel modelling tools. Proof that project results work in an integrated system.

Internet technology will form the basis of an Information Society in the early years of the next century. Multiservice networks will be built on IP packet switched technology and traffic will be predominantly data. High capacity communications links are being deployed at an increasing rate worldwide and will form the basis of the future network. The widespread availability of these high bandwidth links will stimulate new ways of using communications, but transmission capacity is just one aspect of the infrastructure required to support fully the expectations and aspirations of all users.

There is strong pressure from users to introduce new services rapidly and to enable dynamic customisation such as the use of personal profiles for mobile users. There will also be a need to support the definition of services by users including information filtering, transcoding and the use of security proxies. Flexible and scalable management solutions must be developed to support the dynamic user services and the flexible infrastructure required to support them.

The infrastructure will need to provide:

• a programmable processing platform - to run user defined services and customisation components
• a flexible management system capable of adapting to rapidly evolving circumstances
• quality of service - to provide predictable performance for business applications
• security - to ensure protection of business transactions
• resilience - to allow users to depend on it
• scalability - to cope with the expected growth in demand as a result of the wide range of new services which will be available
• mobility support - to allow universal access
• directories and information retrieval tools - to locate resources effectively

The infrastructure will be highly heterogeneous because it must be able to accommodate new technologies incrementally as they become available while still supporting existing components. The infrastructure will be provided globally through cooperation between a large number of competing enterprises. These factors mean that interoperability between components of the infrastructure is also essential.

The principal objective of this project is to prove the feasibility of providing a managed, scalable, programmable network infrastructure capable of supporting the future needs of the Information Society outlined above.

Active networks is a recent, and highly novel proposal that aims to offer network users greater flexibility. A recent EURESCOM strategic study concluded that active networking is the only route to adding integrated mobility, security, quality of service and management services to existing networks [Ref 1]. Active networks is also an ideal vehicle for integrating caches and resource discovery with the network. This project will therefore use active networking to provide most of the features, many of which have not previously been realised in an active network implementation.

To fully realise the benefits of an active network infrastructure, a flexible management system is required, combined with a common information model for resources - both processing and network. The management of active networks, the extent to which management information must be shared, and the communication mechanisms required by the management system are currently open research issues.

The detailed scientific objectives of the project are to develop innovative solutions to the open research problems identified above, and test the solutions through a combination of implementation experiments and modelling. Specifically we plan to develop and test a novel management architecture together with active network versions of the key Information Society infrastructure features (such as resource discovery) that have not previously been validated elsewhere. Other aspects will be considered only to the extent required to prove the novel developments. Modelling will concentrate on dynamic interactions between components, with the aims of minimising integration efforts and maximising performance.

An early milestone is to identify in detail the key mechanisms and management functions that must be provided to enable the communications infrastructure to have the characteristics identified above, and recommend a suitable management and communication architecture.

Key developments will address the efficient sharing of information between active nodes. In particular we aim to define, and competitively evaluate the performance of, mechanisms that allow managed, secure cooperation in resource allocation and execution of user defined programs together with dynamic resource discovery across the infrastructure. Results will be delivered at the end of year 2.

The project will construct a simple active networking infrastructure as an overlay on the existing public Internet. This will enable integrated evaluation of the developments. In addition the project will modify some existing multimedia conferencing tools so that they can provide effective test scenarios to exercise the active infrastructure and its management. The objective is to show that all the project developments can work together at the end of the project

Where possible, work will be based on proposals of the IETF, but it is expected that the project will be a source of new proposals. The primary exploitation objectives are to contribute to standards where appropriate (e.g. in the IETF and TMF), and to build industry competence to a level where European products based on the results of the project can be developed and used in a deployed infrastructure. We expect the final project report to provide some indications of exploitation status.

The project also intends to use existing work wherever possible in order to focus effort on the areas where innovation is genuinely required. With this in mind we have established strong links with the active networks research community, and with some other fifth framework proposals (FAIN, FANA) which are complementary. We also view the information ecology projects in FET.P.2 as a useful source of detailed requirements. A final objective of the project is to continue this process by identifying work elsewhere in the fifth framework that can accelerate progress as soon as possible. The project deliverables will enumerate the work in other projects that was considered during deliverable preparation.

The ANDROID project clearly addresses two of the priorities identified for the 1999 IST Workprogramme:

Specifically the project is well aligned to Key Action IV - Essential Technologies and Infrastructures.

The main contribution is to:

This project will also contribute to:

This project will use an existing, state of the art, active network platform prototype. The platform is component based, and consists of network machines such as routers and associated processing platforms which execute user defined programmes or services in a protected virtual machine. The platform supports the dynamic addition of software components such as programmes, protocols, and policies. Network machines are normally modified by administrators, who can enable new protocols and network layer services, by simply updating the control policies at the node. When the policies are interpreted by the node, any new programmes required are retrieved from a trusted source and added to the nodes code base automatically. The associated processing platforms are frequently modified directly by users, who supply personalised policies, and/or programmes of their own. The policies can request an existing programme to be used in a new way or request a new programme to be added from a cache or from a server of their choice. The project will develop some innovative extensions intended to improve the manageability and scalability of the platform, including:

• extensible multicast protocols
• dynamic resource discovery to assist with mobility
• information sharing protocols providing reliable efficient delivery
• protocols providing security
• policy exchange protocols
• message handling at active nodes
• policy handling at active nodes

It is expected that many of the above will form the basis of new standard proposals.

An active platform clearly requires an extremely flexible management system, but no existing active network implementation has more than a rudimentary management system. The project intends to design, build and test a suite of policy driven management components to overcome this weakness. Development of novel management solutions will require the project to consider:

• system and management architecture

• information models and component models
• management tools and mechanisms
• security components, billing tools and mechanisms
• middleware support for management, including messaging, security policy synchronisation and information distribution

However, it is expected that innovation will be primarily concerned with resource management and efficient information transfer, as service management will be left to the users. The project will focus on policy based management solutions and it is expected that the most significant innovations will be the management architecture, and the proposed policy sets. The policies are expected to influence providers and management fora such as the TMF. There should also be significant innovations in the areas of policy syntax and policy interactions, both of which are currently unresolved research issues.

The project intends to minimise integration effort, and maximise system performance through dynamic modelling of component interactions. It is expected that this will require the development of new modelling tools which could be exploited for other purposes. It may be necessary to develop new modelling techniques, but we consider it more likely that we will introduce complex system modelling techniques to communications that were originally developed for other application areas.

The integration planned for the final year of the project will show for the first time that it is possible to build an advanced active infrastructure, integrated with an active management system, that is effective on a wide-area network. This will significantly influence early deployment of public service networks based on the technology by improving credibility and understanding, outside the active network research community. Whilst the integration in itself may not be seen as innovative it is an essential part of ensuring that earlier innovations have maximum impact on society.

We also expect the project to enable significant innovation in the provision of network services. Active networks is intended to give everyone the freedom to develop their own services and our experience is that developers in more application oriented projects benefit considerably from improved understanding of what active networks can offer them. We plan to publicise the possibilities through presentation of the project's aims and results in accessible journals and at European conferences with a strong attendance from application developers.

In addition, the project intends to publicise its results vigorously through publication at international research meetings. In this way the results can have the maximum possible impact on the research community, and the researchers can have the earliest possible feedback on those ideas that are destined to be less useful.

The ability to introduce innovative information services rapidly and at low cost is essential for the unrestrained development of a user friendly Information Society. The Internet is increasingly becoming a ubiquitous resource providing basic network connectivity across Europe and beyond. Unfortunately, the need for backward-compatibility and the lengthy standardisation process inhibit the introduction of new technologies, such as multicast and quality of service controls, at the network level. Active networks address this problem by making the network itself programmable. This allows individual users to configure their own use of the network and define their own services.

This project will investigate the feasibility of deploying a managed active network infrastructure across Europe. Such an infrastructure will present citizens with the ability to access customised multimedia services, regardless of their location and exploiting the full capabilities of their access devices.

A common European approach is required to ensure that maximum benefit is derived from these advances throughout the EU.

The project brings together partners with considerable international influence in telecommunications and information technology, each contributing recognised expertise in complementary areas. This will rapidly develop a significant capability in an emerging strategic technology area and will help to close the gap between Europe and the USA and Japan in understanding how to make effective use of the full potential of the Internet.

Project partners are drawn from a range of European countries, ensuring that national considerations will not unduly influence the scope of the work. The market strength of the industrial partners in both public networks and information technology allied to the influence of the consortium in standards bodies such as the IETF will maximise the chances of a common European approach being adopted to the benefit of all.

The programmable network infrastructure that this project will investigate will have significant impact on the way in which public networks are used to deliver future information services. The infrastructure itself and the services hosted on it represent a convergence of telecommunications and information technology. The supporting networks will be operated by a number of carriers using a heterogeneous mix of technologies. Results of this project will have clear influence on EU policy development on convergence, telecommunication regulation and component interoperability.

The project contains large industrial, SME and academic partners. This distribution accords with EU policy encouraging SME's. Collaboration within a developing European research community in active networks will stimulate the mobility and training of researchers.

The aim of the project is to stimulate the availability of programmable internetworking infrastructure. This will enable the deployment of a wide range of high performance multimedia applications, universally accessible regardless of location. Providers of these applications will include SMEs who will be able to leverage the programmability of the network to develop applications in a cost-effective way for niche and specialised markets.

Accessible, high quality services tailored to the needs of each user will improve quality of life in a number of ways:

Teleworking will be made more attractive, saving time and reducing unnecessary travel with its associated pollution, traffic congestion and use of natural resources.

Delivery of education to all, at all times of life will be achievable at low cost and suited to the circumstances of the student.

Ease of customisation will stimulate the availability of applications tailored to reflect linguistic and cultural diversity.

Support for a wide range of heterogeneous end systems will maximise the usability of services so that access to the Information Society is also open to people with disabilities.

The global economy is becoming increasingly dynamic. Successful organisations will need to respond rapidly to changing circumstances and new opportunities. Business processes and relationships with other organisations will need to be flexible. It is just this flexibility that an active network infrastructure can provide. Its availability to European enterprises will help them to remain competitive in the global market.

A further implication of an open programmable network infrastructure is that third party application developers can have access to the resources and customer base of public network operators when developing new services. Developing these services will provide significant opportunities for employment in companies of all sizes.

The availability of applications tailored to individual needs, rather than to the "standard user" will improve the employment prospects of people currently disadvantaged by disability.

The principal objective of this project is to prove the feasibility of providing a managed, scalable, programmable network infrastructure capable of supporting the future needs of the Information Society. This has major potential impact on a number of business types including network operators, vendors of IP networking equipment and providers of management tools and other software solutions. This project involves industrial participants from each of these key areas.

The subject area covered by this project is new and ideas are still developing. Significant advances are coming from the academic community, which is well represented in the consortium. The industrial participants in this project will be exposed to these new ideas first-hand and will be able to influence developments to meet their own needs.

The key feature of the infrastructure proposed here is the ability to support the deployment and execution of network programmes as and where required by network operators and users. This overcomes the problem of adding new features and technologies to the installed network base that is inhibiting the full development of the market in communication and information services.

The dissemination strategies adopted in this project reflect the need to influence the emerging market for flexible networks and services at a number of levels:

Members of the project consortium have existing links within the active networking research community and will continue to play an active role in advancing the state of the art. Papers will be contributed to research journals and there will be regular participation in research conferences, for example the Intelligence in Services and Networks conference.

The ANDROID project will participate in a cluster of active networking projects in Action Line IV.

In addition, appropriate results will be disseminated via CLIMATE (a project cluster on intelligent mobile agents supported by the ACTS programme) and via AgentLink (an ESPRIT funded Network of Excellence for agent based computing).

It is particularly important that impact is maximised among application developers to prepare the way for the emergence of a market in third-party service components. Articles in accessible journals and presentation at European conferences with strong attendance from application developers will be priority routes for highlighting the potential benefits of a managed, programmable network infrastructure.

A number of technical white papers and guideline documents on specific issues will be produced and made widely available throughout the industry. A public web site for the project will be set up and maintained to allow free and timely access to important project results.

Open standards will be required if the full benefits of the project are to be exploited in deployed networks. It is therefore important to build and maintain strong links with various standards organisations. Members of the project consortium are very active in standards fora and will be able to use their influence to ensure that future developments in these fora are consistent with the vision of the project.

In addition, the project will establish and develop mechanisms to promote and raise awareness of its results to organisations such as EURESCOM, TMF, IETF and OMG.

The major companies of the consortium have an active policy in patent application, IPR protection and licensing.

As a major international provider of networks and communications services, BT anticipates benefits from the deployment of a managed, programmable network infrastructure such as addressed in this project:

• increased network traffic as the network becomes more versatile and attractive to users
• reduced operating costs from automated configuration of management systems
• reduced central management overhead using policy-based management
• simplified service management by delegation to the user
• additional flexibility to allow rapid service introduction
• development of service components by third parties

This project provides the opportunity to work with major suppliers at an early stage in the development of a new approach to networks and service provision that will have significant market impact if widely deployed.

Compaq is a supplier of world leading management tools for communications networks. The infrastructure proposed here is an example of the convergence of wireline, wireless and IP networks which will have significant impact on the development of future management solutions. This project will bring together experts in each of these areas to develop the new skills that will be required to meet the needs of customers in the future.

Compaq hosts many conferences and fora. It is envisaged that they will provide a channel for the dissemination and exploitation of results for the consortium as a whole and for individual customers and partners of Compaq. In addition various established web sites and user fora will be used for publication and dissemination.

The IP Edge Device concept aims to satisfy the emerging needs of customers while offering efficient mechanisms to a provider for managing IP traffic on its backbone infrastructure, whatever the underlying technology might be.

The IP Edge Device is a Customer Premises Equipment which, at the end of the ANDROID project will integrate advanced traffic regulation capabilities, multicast services capabilities, security mechanisms and high level management interfaces that allow a provider to programme IP services easily and quickly to satisfy the users' needs.

The market for this type of equipment is proportional to the Internet or Intranet access points in the enterprises, that is to say greater than the number of sites connected to the Internet. In the world, this market is estimated to more than 2 millions in year 2000 and more than 3 millions in 2001. Thomson-CSF Detexis will develop during the ANDROID project a marketing and commercial strategy to address a significant part of this market. The targeted customers will be network operators and those users requiring managed access to advanced features of IP networks.

Secunet provides vendors of information systems and their customers with solutions to enhance security for various kinds of applications. In the ANDROID project the teams will be involved with the appropriate IETF working groups and will influence the standardisation process. This will promote the Secunet strategy by enabling the company to provide attractive security solutions for the next generation Internet. Secunet will invest in this technological approach because we believe that the Active Network approach reflects a broad range of requirements that are missing in current applications. Secunet is going to set up partnership agreements with major network providers and IT enterprises to implement fundamental components for a security infrastructure required for the information age. Through the ANDROID project it will be possible for Secunet to implement security enhancements for active networks. The participation in the ANDROID project will force the development of enhanced security solutions for inter-domain authentication in all fields of telecommunications business. The results of the ANDROID project will complement our activities and support the company in providing common and standardised security solutions for our customers.

MediaSec is one of the leading suppliers of watermarking technologies. Content security is expected to be of great significance across a number of industrial sectors. MediaSec will use the experience gained in ANDROID to improve its technologies and the competitiveness of European watermarking systems.

The main objective of NTUA is to use research results to advance and proliferate scientific knowledge. Exploitation of research achievements is carried out through education and consultancy. The results of the ANDROID project will play an important role in both maintaining the expertise of NTUA researchers and advancing the knowledge of students in the crucial technologies of Active Networks. The achievements and results from this project will be disseminated in top archival journals and international conferences at once fulfilling both the objectives of the European research and NTUA's exploitation strategy by enhancing its knowledge base and prestige.

As a university, UCL has no direct plans to exploit the results of ANDROID commercially. It is, however, active both in multimedia conferencing and in the development of high performance communications infrastructure. It expects to profit directly from the results of the project in the way it organises its multimedia conferencing activities in the future, and the way it assists in the development of UCL, British and European research networks. UCL has a tradition of using the results of one project in subsequent (or even parallel) projects. Thus the applications work from earlier MECCANO, ICECAR, DARPA and other projects are being used as a basis for many of the UCL contributions to GIANA. We expect to use the results in the same way, and thus feed our results indirectly into the commercial activities of other parties. Finally, we work closely with many companies, and receive substantial equipment donations from them. We expect to exploit the results of the project directly by making us a more attractive partner with such partners.

The project workplan is structured around nine workpackages, leadership of each being determined by the principal interests and expertise of the project partners. This clear technical leadership of each workpackage will make the project management efficient.

There will be close interaction between related workpackages throughout the early phases of the project and the iterative development of prototypes so that innovation in one workpackage can be exploited and validated in others. The final phase of the project consists of a single technical workpackage in which all partners participate to integrate earlier developments.

Coordination between workpackages will be handled in WP1: Project Management in which all partners participate.

Each workpackage has participation of at least two partners. This will ensure that the project gains full benefit from its collaborative nature.

It will be an integral part of each workpackage to address exploitation and dissemination of results. Project-wide coordination of dissemination, liaison with other projects in complementary areas and with appropriate standards bodies are of particular importance. These activities will be managed in a separate workpackage (WP9) and will ensure that the results of the project are visible throughout the industry. This will maximise the impact of the project.

The project management workpackage will include overall technical coordination of activities across the project. All partners will participate in this coordination activity. The aim is to minimise direct dependencies to reduce risk, but to maximise interactions between workpackages. Specific interactions between workpackages are the responsibility of the individual workpackage leaders. This workpackage will have a coordinating role and will organise meetings to resolve any differences between workpackages.

A number of innovative technical developments is required across the workplan. This means that there is a level of risk associated with each workpackage. This workpackage will manage that risk and plan for any necessary changes in the technical direction of the workplan in collaboration with the individual workpackage leaders.

Continuous assessment of the progress made by the Project is required to allow necessary changes to the technical direction to be identified in a timely fashion. This workpackage will be responsible for on-going evaluation and reporting to project management.

The principal objective of the Project is to prove the feasibility of providing a managed, scalable, programmable network infrastructure. This will be done through experiments which integrate output from different workpackages. Each of the main technical workpackages (WP3, WP4, WP5, WP6) will carry out testing of the components they develop before making them available more widely in the Project. This workpackage will perform limited integration tests, assessing components from more than one workpackage together so that possible integration issues can be identified early. Integration of the management system with other components in controlled network conditions, prior to full wide-area integration is of particular importance.

Feedback from evaluations will be made available to the appropriate workpackage leaders and to the project management.

In addition to the internal evaluation activities, periodic evaluation by external expert reviewers will be arranged by this workpackage. This workpackage will have the responsibility for producing regular reports on the progress of the project, as required by the Commission.

Active nodes support dynamic addition of executable code to customise the network behaviour to the requirements of a particular service instance. Realisation of an open, programmable network requires that users are empowered to define the behaviour of their own services. This will only be acceptable to network providers if the integrity, performance and security of the network infrastructure are not compromised. Active networking at the transport layer - injection of code into router kernels [Refs 2-16] - is attractive in the long term but faces major problems in these areas. In addition, integration with existing networks would require changes to software in existing routers.

A solution carrying lower risks, while still delivering the required flexibility is application layer active networking, ALAN [Refs 17-22]. ALAN redirects active packets to processing platforms that are situated at the network nodes, but are logically end systems. This avoids the need to dynamically alter the forwarding rules in the network and minimises the performance impact on non-active packets. Integration with existing network routers is straightforward as the active nodes can be added as an overlay without modification of the existing router software. This project will therefore use ALAN.

This workpackage will use an existing, state of the art active network platform prototype. It will enhance and extend it to include management and communications capabilities. Expected extensions include dynamic resource discovery [Ref. 23], security measures, component relationships, and the effective use of existing approaches such as CARP [Ref. 24], ICP [Ref. 25] and DEN [Ref. 26] to share state between nodes. Each node will run an operating system, which is responsible for allocation and scheduling of access to node resources such as computing power, storage and network connectivity. The node will also provide common facilities for essential features such as multicast and management. These will be addressed in other workpackages (WP4, WP5).

Two types of node are envisaged, those associated with conventional routers and those hosted on IP edge devices bridging from client networks to public networks. An IP edge device is a generic IP building block that can offer enhanced IP services to its clients while making optimum use of the public backbone infrastructure. It is capable of offering advanced traffic regulation, security mechanisms and high-level management interfaces, allowing IP services to be quickly and easily configured according to policies defined by a local user or remotely. An IP edge device can provide a flexible environment supporting all IP-level functions including diffserv and IPSec as well as providing the basis for message handling and policy handling at an active node. Ten IP edge devices will be available for use within the project.

The kernel of an active node is the dynamic proxy server (DPS). This is a software component that can dynamically load other components (proxylets). These can then be executed to meet the processing requirements of active packets. Currently the DPS obtains proxylets from a cache hierarchy using HTTP on the basis of policies. The policies can be passed by client or server using HTTP POST, passed by the client using HTTP GET with an embedded cookie, and passed by the server using HTTP GET_RESPONSE metadata. Policies can also be obtained from caches by the DPS. The policies can apply to any flow (identified by a port number) passing through the server. Packets in the flow can signal changes by setting a flag pointing to a new policy held by the DPS, or sending a URL for a new policy. The policies are expressed in XML in accordance with a defined schema accessible to the server. The proxylets are implemented as java beans. The existing DPS will be extended to offer mechanisms for dynamic discovery of proxy servers and proxylets to support mobility and changes in demand. This will probably be achieved using the DEN proposals for directory servers, in conjunction with a hierarchical search mechanism enabling searches to begin at the local node and broaden as required.

A service component model that defines the structure of a component that can run in the execution environment will be developed. A component will be self-describing according to an information model that will also be specified here. This will include information such as a component's security policies, resource requirements, authentication information and available payment mechanisms in addition to its functional characteristics and its facilities for communicating with other components. The service component model essentially expresses both which components are capable of working together and which components are allowed to work together. The goal is support secure, predictable composition of services from components developed in an open market. Techniques for guaranteeing code safety and integrity will be incorporated. The execution environment will require mechanisms to interpret and enforce management policies, including the resolution of conflicting policies, as the execution environment is a network resource shared by many users. It is anticipated that the component model will make extensive use of policies expressed in component metadata using XML.

An active node is a network resource shared by a large number of users and it is essential that processes initiated by one user do not affect the performance perceived by any other user. Exhaustive testing of all proxylets and their potential interactions is probably not feasible in a dynamic environment. Adequate security is achieved by confining components to a constrained execution environment. This isolates proxylets from each other with the active node controlling access to the shared node resources. However, the DPS must still ensure the integrity of the active node and hence the network. This requires extensive security measures which will be addressed in this workpackage. Proxylets need to be authenticated on arrival at an active node to grant access to resources and to ensure proper billing. The integrity of proxylets must be monitored to prevent execution of invalid or expired code. This will require a proxylet certification mechanism, probably based on IKE and JDK1.2 security provisions. Resource negotiation with each proxylet is policy driven. Strict isolation between proxylets will seriously affect flexibility. Where services are being dynamically composed in unpredictable ways, it is important to be able to define relationships between them and support cooperative working. This will be done using policies associated with proxylets and specified in the proxylet metadata using XM. This package is responsible for the schema. Policies, their management, and interactions with the certification authority will be specified in WP4.

The active networking infrastructure will be realised as an overlay of active nodes on the existing passive Internet. The architecture of the active nodes themselves will offer scalable communications throughput.

An important feature of the infrastructure developed here is that it must consider interactions between a set of active nodes that will need to communicate to support cooperation in the allocation, caching and execution of service components.

A simple object communication model will be required to support the interactions between active nodes. This will have security features to prevent fraud and to allow policy driven resource negotiation. The protocols used must support efficient, reliable information sharing and must be appropriate for wide-area, scalable deployment in heterogeneous networks. The requirement is for a simple, ubiquitous protocol that does not impose a rigid set of assumptions about its usage. Existing middleware solutions such as CORBA and DCOM do not meet these requirements [Ref. 27]. This workpackage will extend the use of XML and http, which have already been used in the existing DPS, together with state of the art techniques used for state sharing in Internet caches [Refs. 24,25,28]. In addition the active nodes will be integrated with a scalable, store and forward messaging system for reliable delivery of policies and events. The messaging system will be based on specialisation of existing prototypes and on emerging IETF protocols for efficient messaging.

Requirements for managing interactions between active nodes and potential conflicts will be considered. Interactions between the active nodes and the existing passive network elements will also be investigated, with particular consideration given to services requiring end-to-end broadcast and multicast. In a network containing multiple active nodes, naming, addressing and routing must also be addressed. Active nodes must be identifiable within the infrastructure. Directories must also hold information about the network addresses associated with each active node, whether unicast or multicast. Routing requirements between active nodes will be addressed in this workpackage. Neighbour discovery using local broadcast could be used, combined with multihop policy-based approaches.

The emphasis on demonstrably scalable technology solutions within this project highlights the need for an extremely flexible, lightweight management system [Refs. 29-32]. This will be fully distributed and based on autonomous elements within the network making decisions based on policies defining their authorisations and obligations in the context of knowledge of their local environment (i.e. no global knowledge required). The management system must allow secure exchange of management information in a heterogeneous environment with multiple points of control. The management system will be policy-based [Refs. 33-36] and the definition of the required policy sets is an area in which significant innovation is anticipated. This workpackage will focus on defining management policies for a network containing active nodes, and enabling the actions requested in the policies.

The proposed ALAN approach itself simplifies the management task facing the provider of the active networking infrastructure. The infrastructure is generic and is unaware of service-specific issues. The management of specific services will be included in the service definition using service management proxylets, as required. The remaining issue is management of the active network infrastructure itself.

Management of processing power and its distribution will be considered. Management of network capacity is adequately dealt with elsewhere and will not be repeated here.

An essential part of the active node architecture is the secure execution environment. This is a constrained environment offering controlled access to the resources of the active node processing platform according to a set of policies. A variety of independently developed programmes must be able to run simultaneously in unpredictable, dynamic combinations in the execution environment. Interactions between components must be supported if required by a user, but there should be complete protection of other users from unwanted interference. This requires controlled allocation of all shared processing platform resources in which there should be support for differential quality of service.

Management features required to protect the infrastructure from abuse and to guarantee proper operation include:

• code authentication and certification
• resource negotiation
• secure management
• secure logging
• protection against denial of service
• billing and accounting
• proxylet mobility, including cross-certification

The management system will focus on resource management in the active network system and will use efficient information transfer mechanisms. The key non-functional requirements on the management system are that it should be distributed, reliable, scalable to global dimensions and that it should offer operational simplicity and flexibility. Support for the management of mobility, security and multicast will also be explicitly addressed. This workpackage will produce a management information model and a specification of requirements for a management system. These will be taken as the basis for design and implementation in WP3.

In a system based on active nodes, any active node can, in principle, be used to host service components. There is therefore scope for managing which nodes are involved in running a set of proxylets to deliver a particular service. The basis for selecting active nodes may be either static (e.g. location, physical network characteristics) or dynamic (e.g. processor loading) considerations. This workpackage will study techniques for performance management across a group of active nodes. Mechanisms for messaging and distribution of management information between active nodes will be developed in WP3.

Security management of the infrastructure is a key issue. Authentication and authorisation of both users and code, negotiation of resource usage, auditing against policies and enforcement will be necessary. The area of integrity, including component validation, certification and conformance will be important. This will include both digital sealing and signing, and the capabilities and obligations of a component. The approach adopted here will be to draw on work elsewhere rather than to attempt full coverage in this project. These management services must be extremely stable and robust since information about changes in policies and cancelled certificates must be reliably available and distributed in a timely manner to all relevant nodes. Synchronisation of policies across distributed network resources will be addressed.

Management of interactions between the application layer and the network layer will be addressed. It is possible that user services operating within allowed constraints at the application level could result in excessive use of network resources (by inappropriate use of broadcast techniques, for example). This could then result in degradation of service for other users. There is also a strong requirement to manage policy interactions to provide coherent rules in different network elements to support end-to-end services requiring support for security, mobility, quality of service and multicast. This may be further complicated where the same devices are shared by multiple organisations. Edge devices will require the capability to interoperate with the management system in the interpretation and implementation of policies.

A design specification for the management system will be produced which will be a public deliverable from the project (D5). This will be an important activity in disseminating project results and raising awareness of the issues addressed. Activities in this workpackage will have significant impact on network providers and in external management fora such as the TMF.

The design specification will provide a basis for selecting the components necessary to enable the actions required by the management system. Existing components will be identified where available. There will be a need to develop new components in this workpackage and also to exploit prototypes developed in other workpackages.

The central benefit to users of the infrastructure proposed here is the ability to dynamically define and compose communication services. This workpackage will develop a set of modular transport protocol components that will be integrated with the active networking platform so that they are available for use in new user services. The components will be based on existing and emerging IETF protocols

The services to be supported will be multimedia, distributed applications each requiring different end-to-end communications characteristics. The set of protocol components will include:

• support for quality of service using a differentiated services approach [Refs 37-40].
• extensible multicast protocols allowing multicast filtering, error detection and correction and heterogeneous end-systems [Refs 41-46]
• mechanisms for secure interactions based on IPSec [Refs 47,48]
• support for binding of mobile nodes
• support for service mobility

Protocol components will be required in the management of the infrastructure itself. Specifically there is a requirement for information sharing protocols that can provide reliable efficient delivery for group management, including policy exchange. These are addressed in WP2

The key innovation in this work package is likely to be the modular multicast stack, which is intended to be dynamically configurable in response to policies supplied by the administrator. Stack components will not be proxylets as it not intended they be supplied by users, they will conform to a service provider interface provided by the socket manager. Flexible multicast is a key requirement, as it will be used intensively by both users and managers. Multicast is also resource hungry and exposes a feature to proxylets which must be managed very carefully to avoid excessive resource uptake by individual users. We envisage temporarily removing multicast from those places; it is not needed to simplify this management.

Service mobility will be addressed in this workpackage by an investigation of the use of active smart cards with Java capabilities. A user could hold his personal policies (security authorisations and specific service requirements) on such a card in a portable format. Accessing the network via an end system equipped with a card reader, the user could trigger automatic configuration of local network resources.

All other components developed in this workpackage will have the flexibility and resilience to be used in various combinations in trial applications but will not be dynamic.

This workpackage will produce a public deliverable (D7), which is a report on IP layer transport services for use in an active networking environment. This will include a description of a set of prototype modular components to be delivered to WP8 for integration. The modular components developed in this workpackage will be made available to several other workpackages during the lifetime of this workpackage in accordance with the project's iterative approach to development.

The principal focus of this project is not on the development of end user applications. However, the main advantage of the approach we are proposing is the flexibility it offers in the deployment of a wide range of new service types. This must be clearly shown and so a small, carefully selected set of test applications will be identified and deployed. The applications will support a variety of usage scenarios and be sufficiently demanding to stress the infrastructure effectively. Test applications will therefore be selected in close conjunction with other aspects of the project. In particular, it will be important to include support for mobility, heterogeneous multicast to a variety of terminal equipment types, security, quality of service and transcoding at the boundary between networks using different transmission media.

It is apparent that secure multimedia conferencing is an application that can test all these aspects in various scenarios. Components to support this will therefore form a particular focus of this workpackage. A set of media tools including support for audio, video, shared workspaces and media processing will be adapted to work within the ANDROID infrastructure and initial trials will be performed [Refs. 49-55]. A component allowing dynamic digital watermarking will also be adapted to the platform [Refs. 56-58]. This will use most of the security primitives and basic service components provided by the infrastructure. It will expose facilities for management control and remote monitoring of its operation. The components developed here will be used further in WP8 as part of the overall system integration and testing.

Dedicated test applications designed to validate the functionality of the IP edge devices developed in WP3 and the transport protocol components from WP5 will also be produced in this workpackage. These applications will probably not be used further.

The suite of test applications will be delivered to WP8 for integration and testing. A report describing the applications will be released as a public deliverable of the project (D6).

This workpackage will develop and apply a principled approach to the modelling of active networks. The contribution of this workpackage to the overall project will be to apply modelling techniques so that the overall system integration effort is reduced and the system performance is maximised. New modelling tools will be developed to elucidate both the static and dynamic characteristics of the active networking infrastructure. Novel modelling techniques may be required but more likely, existing techniques will be applied to the specific problems of interest here.

Areas to be addressed will include:

• static and dynamic aspects of the active network. This will include the topology of the active node network and the protocols for inter-node communication. Candidate cooperation protocols between active nodes will be modelled and refined
• static and dynamic aspects of an active node. A functional component model of an active node will be developed allowing predictions of performance and scalability
• security components
• dynamic modelling of service composition and component interactions
• policy handling
• local to global policy transforms

Modelling will be used to validate and refine the architectural choices and design decisions made in other workpackages and also to design overall system test scenarios for WP8 so that scalability to very large systems can be deduced from measurements on the prototype infrastructure.

There will be one formal deliverable, D7, from this workpackage. This is a report describing the design of the active networking models and the results of applying the models to components and deployment scenarios.

This workpackage will occupy the final third of the project. It will prove the feasibility of building an advanced active infrastructure, integrated with an active management system and effective on a wide-area network. This activity is essential to ensure maximum impact of the innovations anticipated from earlier workpackages.

The integration phase of this workpackage will involve a number of aspects. Transport protocol components will be integrated with application layer components to define a set of proxylets that will be available for dynamic composition of new services. These will then be deployed across the active nodes. The IP edge devices will be fully integrated with the dynamic proxylet servers, providing mechanisms for efficient proxylet discovery, transmission and execution. A number of active nodes on various sites across Europe will be connected via the Internet as an active networking overlay in a live wide-area network. This will identify issues associated with deployment of the programmable infrastructure in real public networks. The most significant integration exercise will be to then integrate the active networking infrastructure with the management system.

Following a successful integration phase, a number of experiments will be performed to test the performance of various aspects of the infrastructure. The management system will be characterised in the wide-area environment in the presence of increasing noise and delay. The performance in these sub-optimal network conditions will be compared with the results of tests performed in WP3 in a controlled network environment. Various scenarios associated with dynamic service composition and customisation will be addressed. These will include wide-area multimedia conferencing applications. The performance of various subsystems will be measured and the results compared with predictions from the modelling tools developed in WP7. The efficiency of various network configurations will be evaluated across a range of service types. . Conclusions about the performance and scalability of the infrastructure will be drawn from these experiments. If necessary in the light of the results further platform improvements may also be implemented

This workpackage will produce the major public project deliverable, the active network prototype, together with results of testing in a wide-area network environment (D9). This will significantly influence early deployment of public service networks based on active networking technology by improving understanding of their feasibility and potential benefits outside the active networks research community.

It is recognised by the project participants that there is a considerable amount of related research work underway in other projects and further that adoption of the technology requires a coordinated approach and open standards. This workpackage will establish strong links between this project and the research community. This will enhance the presence that several of the project partners already have by regular interactions at the project level. Relevant projects will be identified and formal mechanisms for information exchange put in place. Participation in a project cluster in IST Action Line IV will be coordinated in this workpackage with the objective of harmonising technical approaches, such as the adoption of common middleware protocols. Key results will be identified and discussed across project boundaries to ensure that maximum benefit is derived from the IST programme. This workpackage will coordinate contributions to cluster activities in influencing standards bodies such as EURESCOM, TMF and OMG including regular presentations of the project results. This workpackage will include regular participation in research seminars on active networking to present an overview of the work of the project and will contribute actively to the annual Intelligence in Services and Networks Conference.

This workpackage will produce three formal deliverables according to the guidelines of the Commission: D1, the Project Presentation; D2, the Dissemination and Use Plan; D3, the Technology Implementation Plan.

The timing of the project workpackages is shown in the following table:

There are nine workpackages as shown in the diagram. WP1 has a planning and coordinating role over all the other workpackages. WP9 facilitates information exchange for the whole project with other projects and activities. WP3, WP4, WP5 and WP6 are responsible for the main technical developments. Exchange of prototypes between workpackages is mediated via WP2. There are links between these workpackages which then contribute components to WP8 for integration and testing.

The principal aim of project management of the ANDROID project is to provide a lightweight but effective framework to support collaboration in realising the goals of the project. Efficient decision making processes and responsiveness to changing circumstances are required.

BT, the Project Coordinator, takes responsibility for overall project management. This includes interactions with the Commission on contract-related issues as well as chairing regular management and technical coordination meetings.

Each participant in the project will nominate an individual to be a member of the Project Management Committee (PMC). The PMC will meet at regular monthly intervals to report and discuss progress, generally using telephone conferencing. Additional PMC meetings will be called as required by the Project Manager.

The PMC has responsibility for monitoring the overall progress and direction of the project, the resources used and the costs incurred. Reasons for any deviations from the project plan will be identified and the necessary corrective actions will be agreed by the PMC.

Any differences between participants will be resolved by the PMC as they arise.

The PMC is responsible for reviewing all project deliverables prior to making them available outside the project consortium.

The technical workprogramme of the project is structured as a number of workpackages. Each workpackage has a clearly identified leader who has responsibility for setting detailed objectives and milestones for the workpackage.

Each workpackage has the involvement of at least two participants. This will encourage good teamworking and maximise the benefits of collaboration.

The general approach adopted in realising the technical objectives of the project is based on prototyping. A number of workpackages will be active in parallel and frequent, iterative exchange of information and components between them is anticipated. These interactions between workpackages will be managed by the workpackage leaders, facilitated by WP1 with preliminary integration testing carried out in WP2.

At the outset of the project, there is a clear idea of the best technical approach to achieve the objectives. However, as significant innovation is required in a number of key areas, there is an element of associated risk and some flexibility is required in delivering the overall objectives of the project.

The workpackage leaders, together with the Project Manager and other PMC members to ensure coordination between workpackages, will identify technical options for achieving the major objectives of each workpackage and will plan changes to the technical approach in the event of adverse findings.

The project will hold three workshops per year at various locations in Europe. These meetings will cover all the work of the project and include a review of progress in each workpackage and detailed planning of the future direction of the project. Each workshop will also include a PMC meeting.

External evaluators will be invited to join the PMC to form a Project Steering Board. This will meet at selected project workshops to make an assessment of the project's progress and provide support for project planning.

Interim technical meetings will be held as required within specific workpackages. These will be called and organised by the leaders of the workpackages involved.

In normal circumstances, full use will be made of technology to support cooperative working and minimise travelling time. This will include routine use of e-mail and conferencing tools.

An internal project web site will be set up and maintained by UCL. This will facilitate effective exchange of information between all project participants. A public web site will be set up to disseminate results widely outside the project.

This project will participate in an Active Networks cluster including project FAIN.

Effective dissemination and exploitation of the results of this project will require regular reporting at selected international conferences. It is likely that some of these conferences will be located outside EU Member States and Associated States.

A significant strength of the ANDROID project is the composition of the consortium. It relies on a close collaboration between eight highly skilled research centres and industrial partners with wide-ranging technical expertise and global status. It includes a telecom operator (BT), a provider of IT services and solutions (Compaq), a vendor of networking equipment (Thomson-CSF Detexis), two SME's with particular expertise in security solutions (Secunet and MediaSec), and three universities (University College London, Technical University of Madrid and National Technical University of Athens). Several of the partners have already worked together in the COIAS and HIPPARCH projects, or under existing bilateral agreements so we are confident that collaboration will be unusually close.

BT will lead WP1: Project Management, WP2: Evaluation and Assessment, WP3: Active Nodes and WP8: Integration and Test.

In addition, BT will make significant contributions to WP4: Management System

Compaq will lead WP4: Management System and will make significant contributions to WP3: Active Nodes, WP5: Transport Services and WP8: Integration and Test.

Thomson-CSF Detexis will lead WP5: Transport Services and make significant contributions to WP3: Active Nodes , WP4: Management System and WP8: Integration and Test.

NTUA will make contributions to each of the technical workpackages.

University College London will lead WP6: Test Applications and WP9: Dissemination and Implementation and will also contribute to WP3: Active Nodes, WP4: Management System, WP5: Transport Services and WP8: Integration and Test.

SecuNet will make a major contribution to WP4: Transport Services and also contribute significantly to WP3: Active Nodes and WP8: Integration and Test.

Technical University of Madrid will lead WP7: Modelling and also contribute to WP4: Management and WP8: Integration and Test.

MediaSec will contribute to WP6: Test Applications and WP8: Integration and Test, with an emphasis on watermarking-based content security components.

BT is an international telecommunications and Information Technology group with a turnover of about £18.2 billion in 1998/9. BT is one of the world’s leading suppliers of fixed and mobile communications services. In the UK, we support around 27 million customers’ lines and, through our 60 per cent stake in BT Cellnet, over three million mobile connections. Our main services are providing applications, services and networks on a local, national, and international basis for residential and business users in the areas of voice, data and multimedia, through fixed, mobile and Internet communications. For further details see http://www.bt.com

BT’s main research site occupies 100 acres at Martlesham Heath in Suffolk, some 70 miles north of London. Approximately 4,000 people are based there; including 700 people dedicated to researching future technologies, systems, networks and services.

BT spent £268 million on R&D in 1998/99 this includes £54 million invested in longer-term corporate research. Of this around 75% stays with our in-house researchers, the remainder being invested in pre-competitive, university and other collaborative research initiatives including European funded programmes. BT is an active participant in European collaborative R&D both through its shareholding in EURESCOM GmbH and through its participation in EU Framework Programmes. BT has participated in all the EU Framework Programmes, mainly in the telecommunications and IT Specific Programmes. In the 4^th Framework Programme, BT participated in a total of 23 projects in the ACTS, ESPRIT and Telecommunications Applications Programmes,

The research team involved with this project have been responsible for active network developments in conjunction with academic partners since 1996. Currently the work involves a team of approximately 20 graduate engineers, about half of whom work in BT. The remainder are academics from institutions in UK, Spain and Australia. This team were responsible for the original ALAN prototypes. Results have been published at WWW conference, WISP, ICW, IWAN, ECMAST, IS&N and HIPPARCH workshops, and in the journals Computer Networks and ISDN systems, and British Telecom Technology Journal. The work is highly regarded by the international research community, and members of the team are frequently invited speakers at international events. Some of the developments have been undertaken in the ESPRIT project HIPPARCH and the ACTS project COIAS.

BT brings to this project the extensive experience of working with most of the project partners that the active networks team has acquired through its collaborative relationships. It also brings experience of designing and building early ALAN prototypes [Refs. 19,20] which will be essential to this project since ALAN has been identified as an appropriate starting point. BT intends to provide the project with early access to the results of its ongoing work in this area. BT is also providing its extensive experience of building and managing large-scale complex systems. This will be particularly useful in the management and integration tasks

Ian Marshall is a technical group leader at BT Labs. Currently he is leading a BT project on active networks. He is also responsible for a major University Research Initiative funded by BT and is a PMC member in the ESPRIT project HIPPARCH, the ACTS project COIAS, and the Eurescom project P844 (active networks). Since becoming a group leader in 1992 his research has concentrated on distributed systems and the Internet. Previously he worked in optical networks, broadband networks and network strategy. He is the author of over 80 published papers in these areas. He is a fellow of the Institute of Physics, a chartered engineer, and a member of IEEE and ACM. He serves on several institute committees, on EPSRC and European research panels, and on the programme committee for a range of international conferences.

Dr. Mike Fisher received a BA in Physics from Cambridge University in 1984 and a PhD from University of Surrey in 1988. He is currently a team leader in the Distributed Computing and Information Systems research group at BT Laboratories and manages a project including University collaborations. His research interests include middleware technologies for heterogeneous distributed systems capable of deployment on a global scale. He has participated in a number of EURESCOM projects in these areas.

Mike joined BT in 1988 and initially worked in semiconductor materials and device research. He is the author of over 60 published papers and is a member of the Institute of Physics.

Paul McKee is a team leader in the Distributed Computing and Information Systems research group at BT Labs. He currently manages projects including collaboration with a number of Universities. His research is focused on large-scale distributed systems, particularly policy-based management and high performance event-based architectures for capturing and processing management information.

Paul joined BT in 1989 and initially worked on high-resolution optical devices before moving to a distributed systems group where he worked on autonomous replication and low overhead consistency protocols. He has published over 35 papers and is a member of the IEEE Computer Society.

Luis Velasco graduated in Telecommunications Engineering from the Technical University of Madrid in 1998. He worked for 2 years at the University of Madrid in Artificial Intelligence on an EC funded project developing an expert system for medical application. He joined BT Labs in 1997 and is now involved in active networking research at BT Labs. He is working on alternate path routing using ALAN as part of the ACTS project, COIAS.

Compaq is the second largest computer company in the world and the largest global supplier of personal computers. Compaq develops and markets hardware, software, solutions, and services including industry-leading enterprise computer solutions, fault-tolerant business-critical solutions, networking and communication products, commercial desktop and portable products and consumer PCs.

Compaq Services Division has over 29,000 Service professional resources in 550 locations across 114 countries. Within Compaq Services Division, Network and Systems Integration Services (NSIS) provides one of the most comprehensive portfolios of IT services and solutions focused on today’s most pressing business challenges.

NSIS focuses on customers in the financial, communications, public sector, and manufacturing industries. NSIS has over 7,000 professional resources worldwide. These provide expertise and competence to plan, design and implement Industry Solutions in Finance, Manufacturing and Communications.

Particular technology focus areas are Architectures and Application Integration, Internet technologies and network infrastructure, Enterprise applications and infrastructures.

Projects have been implemented on a global scale. Over 20000 networks of various sizes and complexity have been implemented to date. NSIS is the number one ISP integrator worldwide.

NSIS is organized by practice. Within NSIS the Communications Industry Solutions (CIS) practice is the focus for the engineering and integration of solutions to the Communications Industry.

CIS has over 800 Communications Industry professionals worldwide. It is from CIS that the resources required to support the project will be obtained.

CIS has expertise in the engineering and implementation of:

• Industrial-proof network and service management solutions that implement industry standards and are scaleable to the size and complexity of any environment.

• Enhanced Network Services solutions including Intelligent Networks infrastructure that can be rapidly rolled-out complemented with partner applications for services like prepaid calling service, integrated voice services and Internet telephony.

• System and Network Planning and Consultancy Services covering all aspects of planning, acquiring, installing and setting up complex network environments with multivendor equipment.

CIS Telecommunications Network Management Solutions have been implemented in over 150 Telecoms Operators worldwide, including the top 32. Overall CIS has 50% market share of Telcommunications Network Management solutions.

Within Wireless Operators CIS solutions are managing over 25 Million subscribers.

CIS has Software Engineering Centres of Excellence in Reading England and Valbonne France. It is envisaged that these Centres of Excellence will provide project resources.

In addition in Sophia Antipolis France is the worldwide competence centre for communications, with other regional centres in Texas USA, Sydney Australia and Sao Paulo Brasil. There is also an IP Centre of Excellence in Nashua USA.

Chris Bond received a BA in Physics and Computer Science from the University of York in 1978. He has particular experience in carrier grade management of IP and Intelligent Networks. He has been involved in design and implementation of fault management systems for Data and Intelligent Networks, the design of a cross-technology fault manager and the design and implementation of resilient Management Networks to support fault management.

Key skills include Network Management systems for large networks using SNMP and OSI compliant management systems. He has expertise in LAN/WAN networks involving multiple protocols including TCP/IP, OSI, DECnet and SNA as well as in complex multivendor messaging systems including directories.

Graham Criddle received a BSc in Mathematics and Computer Science from Manchester University in 1982. He has 12 years' experience within Compaq and is currently lead architect for a range of projects within the CIS practice. He has skills across a broad range of technology areas. These include Intelligent Networks, particularly Voice over IP and Intelligent Call Routing; Network Design and Management; and Internet technologies, particularly in relation to ISPs.

Thomson-CSF (France) is Number 1 in Europe and Number 3 worldwide in professional electronics for related commercial and defence markets. Thomson-CSF Detexis will provide Internet stack protocol and Internet Edge Equipment. The main interest is the extension of the functionality of the Internet software of the company, production of enhanced commercial edge equipment (IP, security, QoS, access interface) on the communication market.

Thomson-CSF Detexis and its subsidiaries have approximately 7,200 employees, with over 75% engineers and technicians. Its turnover exceeds FRF 8 billions. Its R&D capacity in the field of high-tech electronic systems is unrivalled in Europe. The Company will now develop its international leadership with its European partners in the fields of professional electronics for related commercial and defence markets.

Information Technology and Systems, one of the Strategic Business Units of Thomson-CSF Detexis, inherited Dassault Electronique leadership in the field of real-time networks. Thanks to this experience, it has diversified its activity in the industrial field with its expertise and involvement in the provision of Information Systems for various kinds of applications. More and more, these systems are based upon standard communication protocols and especially on Internet stacks. This led Dassault Electronique and now Thomson-CSF Dextexis to build an important Internet activity.

The network software specialists of the Strategic Business Unit Information Technology and Systems are currently developing a new generation of software based on the last IETF specifications. They plan to operationally implement this in various kinds of equipment. In particular, they are developing IP Edge Devices allowing provision of VPN services with bandwidth reservation. For this matter, a partnership policy with research labs, network and computer manufacturers and engineering companies has been set up. The company participates in national and international R&D projects that intend to develop the new capabilities currently under specification within the IETF including Quality of Service and Mobile IP. The company has been involved in the IETF standardisation process for more than three years now and intends to increase its role in the deployment of the new IPv6 standards suite with the creation of a new international IPv6 Forum.

Drawing on its expertise in networking and Internet Protocols, and its previous experience in various EC projects (including COIAS and Hipparch), Thomson-CSF Detexis will be responsible for the development of IP Edge Devices. These will include active node capabilities and will implement proxylets allowing the provision of added value services at the network access points. Thomson-CSF Detexis intend to make 10 IP Edge Devices available for use in this project.

The company will work on the design of the Network Architecture, and is especially interested in finding solutions allowing operators or corporate network managers to set up multicast secured differentiated services. Thomson-CSF Detexis will be the leader of the design and development of the transport protocol components necessary to provide multicast data transfers, security, adapted Quality of Service and mobility. The developments will be based on the current IETF specifications. In particular, the transport protocols components will be defined taking into account the recent IP standards on IPSEC, DIFFSERV, MOBILE IP on IPv4 and IPv6. Thomson-CSF Detexis is already involved in the IETF standardisation process and will participate during the project to works of the relevant working groups. The Company will set up a test and evaluation platform on its site of Saint-Quentin-en-Yvelines, this platform will be interconnected through the Internet to the partners platforms. The company will also study carefully the management capacities to include in the active nodes, and in particular in the Edge Devices.

Patrick Cocquet is responsible for the network research and development activities within the Technical Directorate of the Information technology and Systems Business Unit of Thomson-CSF Detexis. He has been involved in several different national and international R&D projects on Internet technologies. He has been a member of a number of IETF working groups as well as Vice President of the IPv6 Forum.

Philippe Lenglet graduated from ENSAE, France in 1979. He is project manager in the Networks area. He is currently responsible for the HIPPARCH project on novel network architecture based on the ALF (Application Level Framing) concept. He is currently working on mechanisms for managing QoS policy in heterogeneous networks.

He is responsible for the DEDICACE VPN manager product developed by Thomson-CSF Detexis and has worked on a range of different network engineering projects.

Vladimir KSINANT graduated from ENSERB, France in 1989. He is leader of the development team working on new generation IP technology. The Thomson-CSF Detexis IP technology, called MUSICA, is today developed under Windows NT and Unix environments. He participates in IETF working group activities on IPv6, Diffserv and IPSEC.

He is currently the project manager of the ACTS COIAS (Convergence IP/ATM/Satellite) project.

The National Technical University of Athens (NTUA) is the oldest and most prestigious technical University of Greece. It was founded in 1837 and has since then been contributing to the progress of engineering science in Greece, through the education of young engineers and its multi-faceted research and development activities.

The University consists of seven departments, each one covering a different aspect of the engineering field from electrical engineering and computer science to mechanical engineering. Over 600 staff members are involved in the education of more than 9,000 students in NTUA undergraduate, postgraduate and continuing education programs.

During the last decade, NTUA has developed an extensive research and development activity in numerous fields, with all departments contributing to the evolution of new theories, technologies and solutions for problems more or less related to engineering science. The achievements realised so far highlight the contribution of the Division of Computer Science (DCS) within the Electrical and Computer Engineering Department of NTUA as the leader in this R&D.

DCS is the most active research department in NTUA and incorporates 23 laboratories and study groups, currently involved in more than 60 national and international scale projects. These deal with systems and services in a wide variety of areas, such as communications, multimedia, computers, networks, control, expert systems, components, biomedical engineering, education and management.

The major role granted to the DCS in the R&D field is largely due to the evolution of the Telecommunications Laboratory, which is now the biggest group in NTUA in terms of staff volume and research activities. The Telecommunications Laboratory employs 13 members with Ph.D. degrees in the areas of information and communication technologies and supervises the postgraduate research work of more than 80 highly qualified graduate engineers. The group is entirely devoted to research and its members have extensive theoretical and computational experience in the fields of communication networks design, OSI architecture, protocol verification and simulation, analysis and performance evaluation, software development, distributed database design and management, ISDN, IBCN, LANs, WANS, mobile and personal communications.

The Telecommunications Laboratory of NTUA has extensive expertise in the following research areas particularly as a result of its participation in numerous RACE, ESPRIT and ACTS projects:

• Network Management and Service Engineering
• Mobile and Personal Communications
• Broadband ATM Networks
• Telecommunications Software
• Distributed Platforms
• High Speed Local and Metropolitan Area Networks
• Management of broadband internetworks.
• Telecommunications hardware
• Performance Evaluation and Teletraffic Engineering

Prof. George I. Stassinopoulos graduated in Electrical Engineering from the Swiss Federal Institute of Technology (ETH Zurich) in 1974 and received a Ph.D from Imperial College, London in 1977. He is Professor and Head of the Division of Computer Science, NTUA. His current research interests are in the fields of communication networks, LANs, MANs, broadband networks and services and applications of HPC in networks and switching. He has participated in many national research programs dealing with communication networks as well as in the European Union RACE and ACTS programmes, dealing with broadband ISDN in the areas of ATM Broadband Networks (RDP 2023, R1014, R1022, R63/233, R63/41), Service Engineering and Network Management (R1044, R1024).

Dr. George T. Karetsos obtained the Diploma in Electrical and Computer Engineering in 1992 and the PhD in Telecommunications Systems in 1996 both from NTUA. He is currently a Senior Research Engineer in the Telecommunication Systems Laboratory of NTUA. Since 1991, he has participated in various European and Greek research projects on ATM networks. His research interests include broadband communication networks, high speed LANs and MANs, broadband interworking units, multimedia communication systems and protocols, broadband terminals, performance evaluation and traffic engineering, network resource management, data structures for knowledge based systems, distributed processing and management of distributed applications.

University College London is one of the premier universities in the United Kingdom, with a strong emphasis on research. Its income in 1998 was £185M with an expenditure of £179M. It has some 5000 staff, 9000 undergraduate and 3000 postgraduate students covering all areas including Science, Engineering and Medicine. A set of five hospitals is linked to UCL; its strong audio-video department is sited in one of the hospitals, and is linked particularly to medical teaching. It has recently set up a Language teaching centre, and a distant learning centre. UCL has been running a video teaching network between a number of sites for 13 years. It is connected to at 155Mbps to the London Metropolitan Area Network, with its onward links both to the British SuperJANET ATM network and TEN-155. The ANDROID activity will be carried out partially in the Department of Computer Science (UCL-CS) and partly in the Department of Electronic Engineering (UCL-EE).

UCL-CS has been in the forefront of computer network technology development and applications deployment for many years. It operated the first international link to ARPANET (and later the Internet) from 1973, providing a gateway between the US and British networks, for some 15 years. It still provides technical support to the computer centre now running the service. It has pioneered the usage of most network technologies - both LAN and wide-area. Currently it is connected directly to the Internet, has a private IP/ATM link into the US CAIRN and the BT LEARNET (from its own ATM switches). Through the same switches it is connected to the London MAN via the College Information Systems Division switches. It operates its own primary rate ISDN, and has a direct link via ATM to two DBS up-links. It has been involved with European RACE projects since 1985 (e.g. the NEMESYS, CAR, PREPARE, ICM and DRAGON projects), several ACTS projects (e.g. VITAL. IBCoBN, PROSPECT, MISA, REFORM and COIAS), and a number of US DARPA-sponsored projects onIPv6, multimedia and security. Over the last five years, it has been the co-ordinating partner of the PARADISE Directory pilot, the MICE, MERCI and now the MECCANO multimedia conferencing pilots. It has also either been the co-ordinating partner or a participant in a number of European Commission security pilots (PASSWORD, ICETEL and ICECAR). The multimedia applications, IPv6 experience and security will be fed into the ANDROID project.

UCL-EE is one of the three top rated Electrical Engineering departments in the UK. It is the oldest in England, dating back to 1885 and Professor Sir Ambrose Fleming. The department is strongly focused on research across a broad front: Electronic Materials and Devices; Opto-electronics and Optical Networks; Electronic Circuits and Systems; Microwaves, Radar and Optics; and Telecommunications. In the area of communications research the department is involved from the lowest layers of radio, electrical and optical coding and transmission through routing and resource control at the network layer up to network, service and business management of communications networks involving distributed software systems. The department has been and continues to be involved in research funded by a number of sources: direct industrial funding, UK Government and European research programmes. Examples of EE's involvement in EU RACE, ACTS and ESPRIT projects include: INFOSEC, MODAL, MUNDI, DRAGON, MISA, VERTICAL, FRANS, MORSE, TRUMPET, REFORM, TOPS, MIAMI and AMASE. The department is connected to the BT LEARNET ATM and IP network.

Although Electronic and Electrical Engineering and Computer Science are separate departments within different faculties of UCL there is a high degree of collaboration and co-operation between them for both teaching and research. The BT LEARNET project is currently undertaken jointly by both departments as have a number of ACTS projects: e.g. DRAGON, MISA and REFORM.

Peter Kirstein was, for 15 years, Head of Department of the Department of Computer Science, where he is currently Professor of Computer Communications, and Director of Research. Professor Kirstein has been both Technical and Administrative Director of many European and National projects. He was Director of the MICE and MERCI projects, and is now Director of the MECCANO multimedia project. He is the UCL team leader of the ICECAR security and COIAS IPv6 projects. Professor Kirstein has a BA from Cambridge U, MSc and PhD from Stanford U., and DSc from the University of London. He has been awarded the 1999 ACM SIGCOM award for his contributions to the international development of the Internet. He has written some 160 papers and one book.

Jon Crowcroft is Professor of Networked Systems in the Department of Computer Science. Here he is responsible for a number of European and US funded research projects in Multi-media Communications. He has been working in these areas for over 15 years. He graduated in Physics from Trinity College, Cambridge University in 1979, and gained his MSc in Computing in 1981, and PhD in 1993. He is a member of the ACM, the British Computer Society and the IEE and a senior member of the IEEE. He is a member of the IAB and general chair for the ACM SIGCOMM. He is also on the editorial team for the ACM/IEEE Transactions on Networks. He has written some 50 papers. He is co-author of WWW: Beneath the Surf (UCL Press) and author of Open Distributed Systems (UCL Press/Artech House).

Lionel Sacks is a Lecturer in Telecommunications in the department of Electronic and Electrical Engineering. He received both his BSc in Physics and Philosophy and his Doctorate in high-energy physics from the University of London. Before coming to UCL he was at Liverpool University and in the European Laboratory for Particle Physics (LEPP/CERN). He is a Member of the Institution of Electrical Engineers. His research interests are in distributed systems and system dynamics for large-scale service and management architectures. Current research projects are with such companies as BT, Nortel and One2One. Specific ACTS involvement has included the MISA, TRUMPET and AMASE. Prior to working in Telecommunications, he worked on the ENVISAT project for UCL and the European Space Agency (ESA).

SECUNET - Security Networks AG, located in Essen – was founded in 1996 as a spin-off from TÜViT GmbH. This was in direct response to the rapidly increasing demand for consulting on information security. Long years of experience and an in-depth knowledge of all aspects of information security systems make our international team outstanding in this field. SECUNET's highly qualified engineers, information technologists, physicists and mathematicians combine their expertise to ensure that all relevant aspects of security are taken into account. The integration of components to improve Information Security has become a key factor in production and service provision.

The focus of SECUNET is as an independent IT security consultant with expertise in the following areas:

• Security concepts for network management,
• Network analysis, firewall concepts and penetration,
• Integration of security products,
• Development and integration of security platforms and components,
• Smart Card solutions,
• Trust centre, key and certificate management, digital signatures,
• ITSEC / Common Criteria conformity,
• legal conformance of security specifications.

Our success has been proven by the confidence put in us by leading industrial and service companies world-wide.

SECUNET will contribute to all security relevant technical aspects of the ANDROID project. In the active network node architecture WP02 we will contribute to general security mechanisms, authentication, code security and integrity, and customer subscription. The model of dynamic proxy server requires security measures to protect the active network against misuse and fraud, as well as disruption through invalid code. Also the network must be kept in a stable state by ensuring synchronisation of the proxylet policies and certificates.

Because of the influence on the IETF standardisation process we will provide inputs according the public key infrastructure and policy management. Results of this work will directly influence the IETF working groups on security. To integrate these security features properly in the platform we will participate in the management mechanisms and architecture design of WP05. For the management system of WP03 we will develop software prototypes to meet the specific needs of the related policy and certificate management. We will also participate in the trials of the ANDROID project to get feedback concerning the software prototypes, the security infrastructure and the management system. Because of our great experience with security infrastructures we expect to play the role of a certification and policy authority within the ANDROID project.

Dr. Michael Gehrke received a Dipl-Inform. in computer science in 1989 and a PhD (Dr.-Ing.) in 1994, both from Technical University of Berlin. Since April 1997 he has been Head of Engineering at SECUNET. Prior to this he was project manager at DeTeBerkom GmbH from 1994-1997 and was Deutsche Telekom representative on the DAVIC Security Technical Committee. His areas of specialisation are distributed systems, security protocols, architectures and concepts, open systems, management of distributed systems (TMN, SNMP, OSI), Internet security (incl. WWW, firewalls), advanced networking concepts (TMN, TINA, UPT, CORBA) as well as project management and controlling.

Klaus Lenssen graduated in computer science and communication technology from the Technical University of Aachen (RWTH) in 1991. As a technical member of the University he was responsible for the EUROBRIDGE project (RACE-II) and was an executive member of the IEEE task force "Multimedia Computing" and chairman of the Panel-via-email. Between 1994 and 1996 he was project manager at DeTeBerkom GmbH, with active contributiuon to several EURESCOM projects on ATM for realtime multimedia systems. He was a member of the Internet Services Strategy Board of Deutsche Telekom AG.

Since 1997 he has been Senior Consultant with SECUNET, responsible for security in networking and communications.

The participation of UPM will be realised through the Intelligent Systems Group. This is an inter-departmental group composed of members of the Telematic Systems Engineering and Applied Mathematics in Information Technologies Departments of UPM. The experience of the members of the group is centred on multi-agent systems, data mining, natural language processing and intelligent networks and service management. It covers teaching and research activities as well as R+D Projects at National and European levels.

The Intelligent Systems Group is an active member of AgentLink, ESPRIT network of excellence for agent-based computing, where it participates in the Special Interest Groups on Methodologies and Software Engineering for Agent Systems and Intelligent Agents for Telecommunications Applications and Telematics.

The members of the Intelligent Systems Group have participated in the following Research Projects covering the main topics of their fields of interests:

• SCARAB: Smart Card and Agent enabled Reliable Access (ACTS AC30055)
• ADVOCATE: Advanced On-board Diagnosis and Control of Semi-Autononomous Mobile Systems (ESPRIT AC28584).
• MIX: Modular Integration of Connectionist and Symbolic Processing in Knowledge-Based Systems (ESPRIT 9119).
• CRATER: Corpus Resources and Terminology Extraction (MLAP 93-20).
• Multivariate Statistical Analysis and Bayesian Belief Networks
• for the Development of Intelligent Decision Support Systems with
• Applications in Medicine and Agriculture (INTAS 93-725)
• AUTOCODE: Intelligent System for Automatic Processing of Design Codes of Practice (ESPRIT 2626)

• CORAGE: Optimisation and Control with Approximate Reasoning and Genetic Algorithms.
• SEIC: Civic Information Service.
• SICHAN: Intelligent System for Automatic Check of Standards

• ATILA: Telematic Applications in Linguistic Engineering
• TETRA: Analysis, Design, Implementation and Management of a Tele-working Centre.
• PROTEGER: Multi-agent System for Managing Network and Services.
• MITA: Integration Methodology and Advanced Technologies for the Intelligent Control fo Continuous Processes.
• M2D2: Distributed Meta-learning and Data Mining.

Dr. Carlos A. Iglesias graduated in Telecommunication Engineering in 1993 and obtained a Ph.D. Degree in 1998, both from the Technical University of Madrid. He was teaching assistant from 1993 to 1998 at the University of Valladolid. Since 1998, he has been assistant professor at UPM. He teaches at undergraduate, postgraduate and doctorate levels in telematics, software engineering and intelligent systems.

He has been actively involved in research projects funded by National and European programmes, as well as by private companies. He has published several papers in technical magazines, book chapters and international conferences. The main fields of these publications are agent-oriented software engineering, agent architectures and symbolic-connectionist systems.

Dr. Mercedes Garijo graduated in Telecommunication Engineering in 1972 and obtained a Ph.D. in 1982, both from from the Technical University of Madrid. She is Associate Professor at the Department of Telematic Systems Engineering of the Technical University of Madrid (UPM). She teaches various levels in computer science and communication topics. Her research interests are in software engineering, object oriented methodologies and techniques applied to agents design, and applications of intelligent agents on internet/intranet and communications network management. She has participated or managed several research and development projects funded by National and European programmes and has published several papers in technical magazines, book chapters and international conferences.

Dr. Gregorio Fernández graduated in Telecommunication Engineering from the Technical University of Madrid (UPM) in 1969, in Automatic control Engineering from the Université Paul Sabatier (Toulouse) in 1971 and received a Ph.D. in 1975 from UPM. He has been a lecturing and assistant professor at the Telematic Systems Engineering Department of UPM since 1977 becoming full professor in 1983.

His research interests are in the fields of Intelligent Systems (in particular, in knowledge discovery in databases and agent logics) and computer architectures. He has led a large number of European and National research projects in these fields.

MediaSec Technologies was founded in 1996 as a spin-off from Fraunhofer Institute for Computer Graphics (Fraunhofer-IGD) in Darmstadt, Germany. Fraunhofer-IGD is an Institute of the Fraunhofer Gesellschaft, which has headquarters in Munich, Germany, and is the leading organisation for applied research in Europe. The core watermarking and labelling technology (SysCoP™) of MediaSec is a comprehensive data-hiding tool that secretly embeds various types of information into multimedia documents. The technology was developed in the period 1994–1997 at the Fraunhofer Institute. MediaSec Technologies is backed by the Fraunhofer-IGD and its branch Fraunhofer Center for Research in Computer Graphics, Inc. in Providence, USA.

MediaSec Technologies is providing products, consulting and services in the field of

content security and multimedia security utilizing digital watermarking and labelling technologies. The products and services are applied for the protection of intellectual property and related rights, enterprise content security, privacy, integrity, authentication, and confidentiality of multimedia documents. Currently the watermarking technology supports image, video and audio data in various formats. New methods for marking and labelling of other media and data types (XML, PDF, Java Code, Word, Excel) are under development.

MediaSec’s watermarking technology is one of the first and leading commercial watermarking products worldwide. The products include an easy-to-use API interface, stand-alone commands, batch scripts, a Java Interface, and an experimental World Wide Web service. Software can be easily integrated or plugged into other products such as Netscape, Photoshop, DVD devices and digital cameras.

MediaSec’s watermarking technology has already been successfully integrated, improved and used in several European projects (e.g. TALISMAN, OCTALIS, IDEALS) conducting various field trials.

Dr. Eckhard Koch was head of department at the Fraunhofer Institute for Computer Graphics from 1993-1996 and has been Director of the Essen branch of SECUNET AG since 1998. He is the co-founder and Managing Director of MediaSec, a leading products and solution provider in the field of content and multimedia security, based on digital watermarking technology. He has published several international papers, holds patents in the field of security, has participated in standardization such as DAVIC, and was co-organizer of several international conferences (e.g. IFIP).

During the last 6 years Dr. Koch has participated in several European Projects dealing with security and digital watermarking such as: ACCOPI (RACE), OKAPI (ACTS AC051), TALISMAN (ACTS AC019), OCTALIS (ACTS).