The Future of Internet
State of Play
3 State of Play: Running FP6 Projects
Some of the challenges of the future Internet are already considered by a number of EUu-funded projects. In particular the following projects were presented, emphasizing some of their main challenges and their similarity to those of the future Internet: Ambient Networks, Euro-NGI and Onelab.
a) In Ambient Networks (ANs)
(www.ambient-networks.org)
the basic idea is to have services supported from the ambient control space and the ambient connectivity. Ambient control space is the functionality that is required to establish, maintain and release connections, previously popularly called “call control”. Connectivity is the functionality related to the transport of user data. Ambient control space and connectivity are separated and are manifested in different dimensions. The reason for the separation is to increase the flexibility of the overall network.
The high level requirements of ANs are similar to those of the future Internet and include: heterogeneous networks, mobility, composition, security and privacy, backward compatibility and migration, network robustness and fault tolerance, quality of service, multi-domain/multi SLA support, accountability, context communications, extensibility of the network services provided, application innovation and usability.
The AN project is establishing a handful of basic concepts on which to base the AN architecture, such as: more powerful control layer within the network, overlay network, no special end-nodes (i.e. terminals) but all nodes are network nodes, self configuration, self management, inherent & integrated security infrastructure, mobility – sessions & moving networks, scalability - low entry cost, network composition, intermittent connectivity.
The scope of the Internetworking architecture dealing with naming, connectivity & transport level mechanisms includes: the separation of ID and location; new internetworking layer running on different networking technologies (e.g IPvXx, MPLS, PDP CTXxT); seamless connectivity across different technologies with self boot-strapping and self-management of inter-working functions; name resolutions and translating mechanisms; transport bearer functions across heterogeneous technologies. A common network layer is at the core.
The ANs project has taken a long term approach to advance the Internet and embed business level interfaces into the architecture. The proposed additions to the Internet architecture are fundamental, yet evolutionary. The full potential of an internetworking architecture based on locator/identity split and strong security is still to be exploited.
b) Euro-NGI (www.eurongi.org)
is an NoE on Next Generation Internet and more generally on future communication and information networks, formed by 59 institutions (from the academia and industry) from 18 countries, with research interests in network architecture and network theory, modelling and quantitative methods. The research has been structured around several architectural domains (core; fixed access; mobile access; IP networking and beyond; services overlays) and research domains (socio-economic aspects of NGI; modelling, quantitative methods and measurements; experimentation and validation through platforms; optimisation of protected multi-layer NGI (topology, layout, flow and capacity design); traffic engineering; network architecture evolution, technology integration, control, managing diversity).
Since its inception and initial deployment, Internet has moved a long way towards multi-service IP networks. Before 1992 it was mainly a low-traffic, low-user-population research network, supporting telnet, email and file transfer. In the ‘90s it experienced a traffic and user explosion due to the web and its commercialisation with the emergence of ISPs; it was mainly a best effort network with simple engineering design; its main issues had to do with capacity and addressing. The third generation has already begun and it is characterized today by triple and quad play, internet access, telephony, IPTVv, mobile convergence, multimedia digital content and final user empowerment. It is moving fast to encompass triple, quad play and beyond and global convergence, facing an unprecedented service evolution. Issues that have emerged and need to be addressed include: network and services ubiquity, location and context awareness; global mobility, vertical handover, always best connected (any terminal is your terminal -bio identification-, multi terminal handover); community services, spontaneous and opportunistic networking; self configuration capabilities, cognitive networks and self management; interconnecting/merging the real/physical and digital worlds-WSN’s; machine to machine and ambient intelligence (beyond RFID’s, from today’s information web towards a web of objects); home networking (beyond “just” networking); infrastructure services (virtualization -beyond VvPN’s-, services overlays, etc) - support new business processes and new business models between operators, service providers and other players.
The future Internet services may be designed by the final users! It happened in the past (web, chat, peer-to-peer, etc), just because the Internet was available. Consequently, if new network solutions based on new technologies are made available, new services and usages will emerge. The future Internet should hence be designed to integrate the previously cited concepts and be flexible for future evolutions.
The present Internet was not designed to be an essential part of the behaviour of the society nor of the functioning of the economy. It is facing difficulties to integrate “enhancements” (mobile IP, IPv6, multicast, security). When planning the future networked society, discussion on policy and governance is also essential. The new requirements for beyond IP networks are numerous: cope with the explosion of the number and diversity of connected devices; its architecture should be able to cope natively with mobility and infrastructure-less systems (contribute to ABC solutions); cope with addressing and routing evolution (addressing capacity and structure, separation of identification and location, new routing paradigms); be flexible to deal with the increase of capacity of devices interfaces and access networks;
Euro-FGI’s FP7 related contributions include: the organization of a networking session at the IST 2006 to facilitate the identification and preparation of relevant proposals; the planned organization of a workshop devoted to facilitating the participation of SME’s in those projects (from theory towards industrial innovation); establishment of strong links with external initiatives, organizations and companies involved in Future Internet, e.g. UuS GENI, Japan (Core), China (CAS, BUuPT); the organization of workshops, conferences, a PhD program and a summer school on this topic (to ensure the availability in Europe of a critical mass of top-level researchers on Future Internet – also supports the concept of a School of Network Architecture); a think tank; effort to coordinate Euro-FGI and EIFFEL activities.
c) OneLab (www.one-lab.org)
is a European project aiming to leverage the competence of researchers and federate experimentation facilities. It is not a testbed.
From past collaboration efforts among a large number of research organizations (NoE’s Euro N/FGI and E_NEXxT) over the last 5 years, the lack of benchmarking to assess research results was clearly identified. PlanetLab was seen as the beginning of such an effort with several weaknesses, such as the lack of wireless facilitation. Among the things to pursue was to add diversity to the PlanetLab, as well as establish a European Administration and a Federation. OneLab is a path towards experimental facilities, planning to extend, deepen and federate PlanetLab.
PlanetLab today is wired, based on interconnected NRN’s and hides the real networks. Several extensions are needed and are being undertaken through WiMAXx, multihoming, etc, although PlanetLab has limited ability to see the underlying network. Establish monitoring components could help in this direction.
Federation is a concept of growing interest and it is viewed as a promising way towards a global shared resource. Requirements for federation are: a usage model, specifying “slices” for each architecture or service and user request and reporting procedures; virtualization, allowing for multiple architectures on a shared infrastructure and possibly requiring some very high capacity backbone (e.g. GEANT); governance aspects, referring to APIs, policies, brokerage (trading resources) and trust relationships; analysis and archiving, that require some monitoring facility and a (central) facility to archive data. Furthermore, to achieve the necessary scale, incentives and international cooperation is necessary, as well as mechanisms and business models to ensure long-term operation and sustainability.
The Internet of the Future will most probably not be related to the current internet but it will rely on the Internet and on other network technologies. It will provide opportunities for new businesses and new business models. It will require more international exposition, federation and cooperation with major stakeholders including Technology Platforms, GEANT and users. OneLab is a concrete path exploring the way towards experimental facilities.
In conclusion, the session highlighted that there is already an important FP6 activity, be it from the view point of industrial research, of academic research or of testbeds that may be leveraged when moving into FP7.