19 Conclusions

In 2005, our work on the research intent essentially followed the accepted plan. The slight adjustments we had to make partly reflected new trends in information and communication technologies. However, we also had to react to changes in the budget structure resulting from a decision of the Ministry of Education, Youth, and Sports.

The year 2005 marked important changes in the research of computer networks, distributed systems and their applications. The GN2 project made significant progress in implementing the pan-European high-speed network for research and education (GÉANT2), which, apart from the traditional IP services, will offer its users dedicated 1 Gbps and 10 Gbps digital circuits (lambda services) across the entire continent. The GN2 project also coordinates research activities aimed at developing new services on GÉANT2 and support tools required for creating a transparent virtual environment for cooperation among research teams and individuals within Europe, and, last but not least, at building a testing infrastructure in several countries that will be used for testing new network technologies. Preparations for the successor project of GN2 have also started. Each NREN involved in GN2 is responsible for technical and organisational conditions ensuring the availability of GÉANT2 to their customers and users.

CESNET is one of the project members with major contributions in all activities. Our efforts focused on making lambda services accessible to academic institutions in the Czech Republic and on building the CzechLight testing infrastructure. We also significantly participated in the SEEFIRE EU project and the planned Porta Optica Study project. Both projects support implementation of optical networks in the countries of South-Eastern and Eastern Europe. The deployment of affordable technologies for lighting dark fibre proves to be one of the key factors that may help to overcome the "digital divide" between countries and alleviate the related economic and cultural handicaps.

A very important development happened in the United States in 2005: The National Science Foundation launched the GENI (Global Environment for Networking Investigations) initiative that aims at establishing a global network research environment. The NSF has thus possibly returned to its original significant role in supporting the development of the Internet. This global environment should facilitate research and deployment of new network architectures and distributed systems. In contrast to Europe, it is foreseen that commercial companies will be significantly involved in research projects (not only through selling equipment and services). This tradition actually dates back to the Abilene and NLR (National Lambda Rail) networks. It is worth noting that these two networks are likely to merge as their user bases mostly overlap. The CEF-based NLR network provides 10 Gbps lambdas that are core building blocks for Abilene. Therefore, in the future we can expect two complementary trends in research and educational networks:

Production networks for research and education, primarily oriented on providing services to researchers from various application fields, and differing from their commercial counterparts mainly by the cooperative development and implementation of new services required for research, and also by state-of-the-art parameters that are needed for certain highly sophisticated research applications - all that on the global, continental, national, and local scales.
Environment for investigating possible future trends in data communications that should catalyse progress in research, innovation and economic growth. Such an environment will also consist of global, continental, national and local experimental networks.

The latter trend is mainly represented by the GLIF initiative and experimental infrastructures at all levels. A significant contribution to this research environment should be the GENI initiative. It is thus very important that CESNET already has experience with establishing and using GLIF and various test beds.

Important components of the planned research environment are the new optical and wireless technologies. Realistic predictions expect a rapid growth in the use of PC workstations and servers equipped with 1 Gbps and 10 Gbps interfaces in 2006 and 2007. This, in combination with relatively inexpensive optical transfer systems, will have impact on the structure of networks, in particular in their peripheral parts. However, the term "peripheral parts" is rather misleading: these parts may be quite large thanks to the distance that modern optical transceivers are able to span (over 100 km). These expected future developments also seem to favour the CEF approach to optical networking. Indeed, such networks are no longer limited to research and education, as we have recently seen CEF deployments in health care, military, municipal administration, industrial, and commercial sectors.

The creation of a sophisticated environment for investigating networks and their elements will also change the attitudes towards commercial and industrial organisations. Such an environment will allow for testing products and services in nearly realistic conditions, regardless of their character and/or origin. This is quite different from the situation in the past where the cooperation between the research and commercial worlds was limited to the standard supplier-consumer relations. CESNET is already developing and testing several prototype devices and intends to hand them over to a commercial subject by means of a license or spin-off. While this process is far from smooth, we believe it is important to work patiently on overcoming the obstacles and thus establishing a much closer cooperation with commercial partners, to the benefit of the economy and society.

This can be illustrated by our own development of optical amplifiers CLA (CzechLight Amplifier) designed for the CEF networks. In 2005, we successfully tested first prototypes on the 1 Gbps line between Prague and Hradec Králové and on the 10 Gbps line between Prague and Brno.

Apart from our original development, we also significantly enhanced the CESNET2 production network. In 2005, we integrated the core line between Prague and Brno into a DWDM ring Prague-Brno-Olomouc-Hradec Králové-Prague. This DWDM infrastructure represents a major step forward in extending the network capabilities and making the advanced services of the GÉANT2 network accessible to our users. In the future, we plan to extend this network further, including connections to foreign networks.

Global trends also support our vision of applying programmable hardware for networking applications. A number of commercial products based on Field-Programmable Gate Arrays (FPGA) that appeared recently on the market prove that FPGAs currently offer a better price/performance ratio compared to both ASICs and network processors. We intend to capitalise on our strong position in this field by further improving the family of COMBO cards. Following the recommendations of the final review of the SCAMPI project, we are also working towards making products based on our prototypes.

Along with the network infrastructure, the national research and education networks nowadays allocate considerable capacities to specific network services and applications that are generally referred to as middleware. Beside building a communication environment for cooperation, these services are steadily gaining in importance, in particular for the purposes of service integration.

A major issue in this area is to ensure a secure access to network resources for mobile users. In October 2005, CESNET Certification Authority became one of the founding members of an international federation aiming at improving the integration and compatibility of authentication systems and thus making it easier for users to access Grid resources in particular. To this end, CESNET successfully implemented the eduroam roaming system.

Network security is another global issue that requires coordination on the international level. Therefore, apart from developing technical tools for detecting attacks, we are involved in the international task force TF-CSIRT (coordinated by TERENA), and during 2005 we started to build, in cooperation with the CESNET member institutions, a formalised hierarchy of the security teams. An important part of this activity is the development of methodologies for effective communication within individual security teams and among the teams.

Great attention is also paid to network monitoring and optimising the available bandwidth and other transmission parameters. We developed tools for efficient processing of traffic data and extended the monitoring infrastructure. Our association participates in international bodies dealing with network performance issues, such as PERT.

In terms of applications, the most intensive users of high-speed networks proved to be mainly applications transferring large data volumes in real time: distributed computer and storage systems (Grids), high-quality video conferences and remote data-intensive experiments. It is positive that CESNET and its member institutions are involved in all three areas mentioned above.

Thanks to the MetaCentre activity, we play a major role in the European EGEE project that builds a large Grid oriented on demanding scientific computations. In the upcoming successor project EGEE II, the role of CESNET should be even more important. Our working Grid platform allows us to get involved in experiments with top-quality video conferencing. In 2005, we had a great success with our demonstrations at the iGrid 2005 and SuperComputing 2005 conferences. These accomplishments naturally lead to an increased interest in our videoconferencing activities on the part of our international partners.

We are also involved in several distributed scientific experiments, mainly in the area of particle physics. While Czech physicists do not have their own equipment for large-scale and expensive experiments, they are able to participate in experiments performed on foreign equipment and use own computing facilities for processing the measured data. We feel that our network could be used more intensively in this field. Therefore, we tried to attract new users by organising a seminar titled High-speed Networks for Science and Research, and we are preparing a similar seminar for mid 2006.

One of the strategic goals of CESNET is to accelerate the penetration of information and communication technologies in medicine. As an acknowledgement of the role CESNET plays in this area, Milan Šárek, CESNET staff member who coordinates these activities, was invited to join the Czech expert panel on eHealth. This initiative of the World Health Organisation (WHO) investigates the options and methods for utilising information and communication technologies in health care.

contents