The VITAL Project is an ACTS AC003, and has had duration of January 1996 to October 1998. The current Co-ordinating partner is Alcatel-Bell. The only overlap in partners with COIAS is UCL. The main part of the TINA project is the verification of the TINA architecture in a Broadband ISDN environment. The main part of the project is the provision of a series of demonstrations that show how sessions can be set up, resources managed, subscriptions managed and services provided using the TINA concepts. A recent addition to this project is the investigation of how the VITAL architecture can be extended to Internet applications.
The Internet application being used as a Case Study is the multimedia conferencing one of MECCANO (See Below). There are two overlaps with the COIAS project. First the use of an application level relay, the second the use of a DBS satellite in the demonstration.
Most of the MECCANO conferencing uses IP multicast, with users joining the conference when they wish. To meet the VITAL/TINA aim of demonstrating how Service Providers may offer future services, we use the MECCANO UTG as a converter between multicast and Unicast. The control aspects of this application level relay has been developed further in VITAL; it will be used also, and developed yet further, in the mobile aspects of the COIAS project.
The closest affinity between VITAL and COIAS will come in the final demonstration phase of VITAL. Like many they ACTS projects, the final demonstration will be carried out at one site - in this case Alcatel-Bell in Antwerp, Belgium. For most of that VITAL demonstration, the components will be integrated with an ATM infrastructure in Antwerp. For the Internet demonstration, this is hardly feasible. Most of the physical servers and the multicast IP infrastructure exist only at UCL. For this reason we have decided to demonstrate the Internet/TINA phase by showing the media interactions and the different phases of control remotely at Antwerp though the demonstration is running at UCL in London. The remote presentation will use one of the DBS configurations being considered for the COIAS demonstration phase. In this the output of the different workstations and the session monitoring will be sent as Unicast streams over a DBS system run by a combination of UK Internet Service Provider and DBS operator. All the early tests of the feasibility of the system has been done under VITAL. It is feeding in to our evaluation of whether the use of that up-station is feasible in the context of the later COIAS demonstrations.
MECCANO is a Framework IV Telematics project, which runs almost parallel with COIAS to May 2000 from June 1998, to May 2000. The overlap of partners with COIAS is INRIA and UCL. The main purpose of the MECCANO project is to integrate different multimedia tools, and show how they can be deployed over various heterogeneous networks in real conferencing applications.
There are a number of related activities between the two projects. MECCANO is planning to use DBS satellites as part of its network infrastructure; here it will eventually use the exact system being developed for COIAS. The development costs are being shared across the two projects. Second it is developing further the transport level relay which may be used as one COIAS component. Third it is partly developing and deploying secure conferencing, using security at the application level. Although this is not the only way that security will be used in COIAS, it is an option being considered for some of the COIAS applications. Another development in MECCANO is the UCL transcoding gateway (UTG); this provides multicast/unicast conversion and some transcoding. Finally, QoS is particularly important when a heterogeneous set of IP networks are being used carry the multimedia. For this reason use of IPv6 facilities in the underlying network level, and their use in the provision of QoS, are vital parts of MECCANO. Here there is substantial overlap in the activities of the two projects. We expect them to use at least compatible tools, and some of the QoS components will be shared between the projects.
The Framework IV ICECAR Telematics project follows from the ICETEL project. The first ran from December 1995 to November 997; the second will run form October 1998 to July 2000. The aim of both projects is to develop a heterogeneous key distribution and certification authority (CA) infrastructure, various secured applications, and deploy them in different applications. The only overlap of participants between ICE-CAR and COIAS is UCL.
In ICECAR, the MECCANO secure conferencing system is being integrated with a European security infrastructure, and being applied in certain applications. UCL will operate some of the European CAs, and GMD some of the German ones. We will explore the use of various components from ICE-CAR in COIAS. These will include the use of secured X.509 directories - which have already been recognised as a COIAS requirement. It may also include other security components, and possibly will use the CA and certificate storage facilities being set up under ICE-CAR.
Learnet is a testbed project entirely under BT funding. Its overlap of participants with COIAS is BT and UCL. Its main aim is to study new concepts in network technology and control. One aspect of this is a five node ATM network, with nodes at both BTRL and UCL. As part of this project, a test cell to look at QoS mechanisms has been set up. It is planned to explore the use of this test cell to investigate some of the QoS implementation and algorithmic issues raised in COIAS. One use of LEARNET will be with the CAIRN routers (see Section 3.5).
CAIRN is a testbed and project funded entirely by the US Defence Advanced Research Projects Agency. The only overlap in participants with COIAS is UCL, which runs a CAIRN node, and has a direct ATM link into the network. CAIRN is exploring concepts in the Next Generation Internet, including conferencing, security, IPv6 and QoS. It has set up a substantial network of nodes - all running the CAIRN routers. These routers are designed to provide a flexible architecture to allow QoS algorithms to be incorporated to investigate alternate mechanisms for providing QoS. It is also piloting a number of security implementations including secure conferencing, IPSEC infrastructures, secure DNS and secured routers. There are a number of mobile systems attached, and the piloting of Mobile IP is also part of this work. CAIRN has a significant size of the network (some 30 nodes), uses ATM in some parts, is flexible and experimental. Since it is probably available to UCL and BT, we expect to experiment with a number of COIAS QoS concepts on this network after they have been tried out on LEARRNET on a smaller scale.
PIMMS is a British EPSRC project in mobile communications. It is providing a number of mobile LANs attached to the UCL-CS terrestrial LANs. The project is looking both at the way the UTG of Section 2.2 must be extended for mobile working, and the way both the UTG and the mobile stations should run secure conferencing.
HICID is a British BT-sponsored project to investigate how QoS can be provided for multimedia operations. It will be operating with multimedia services from other projects, including MECCANO, which are able to label packets
UCL expects to provide a number of facilities to the COIAS project as part of the testbed. These include the following:
Each is considered below.
INRIA will be providing, with the assistance of Eutelsat, an up-link to a DBS satellite. The facility will include a limited amount of transmission time on the satellite. The up-link involves a transmitting station on the roof of a building in INRIA at Sophia-Antipolis, attached directly to the LAN infrastructure in INRIA. UCL hopes to provide an alternate, but compatible, facility through an Internet.
A schematic of the expected UCL topology is shown in Fig. 1.
Figure 1, Experimental DBS Configuration at UCL-CS
UCL will access the up-station via a router attached to the Internet via the SuperJanet. It will also have a steerable satellite receiver. UCL expects to put up both a BSD Unix satellite router, with INRIA code, and a NT one with the more standard Convergence-1 code. We expect to experiment with both systems.
One reason for using both, is the limited number of boards which exist for the current board used by INRIA. It is obsolete, and the newer board supports only NT at present.
UCL expects to discuss with several potential satellite Service Providers; it is not yet clear to which it will be able to obtain access.
UCL is part of LEARNET. This is a five-node dark-fibre network currently connecting BT Laboratories, Cambridge U, Essex U, UCL-CS and UCL-EE. One part of this network is that it can be reconfigured for research purposes. An ATM/SDH/OC-3 portion will be brought up front-ended by CAIRN routers. The routers are described below.
Figure 2 Schematic of the LEARNET Configuration
In any case LEARNET includes three nodes at BTRL, UCL-CS and UCL-EE that could be used for COIAS purposes. In addition, there is a three node ATM network inside UCL-CS that can be configured for experimental purposes.
It is possible, but not yet certain, that the CAIRN network will be used in the COIAS project. A schematic of the CAIRN network late in 1997 is shown in Fig. 3.
Figure 3. Schematic of CAIRN configuration at the end of 1997
UCL will deploy two sets of radio LANs connected to routers inside UCL-CS and UCL-EE. The configuration used is shown below:
Fig. 4. Schematic of Radio LAN
Here attached to the UCL-CS LAN are a number of radio LAN stations (S); these currently run the WAVELAN software. Mobile WAVELAN PCs (M) can communicate with unicast IP with the stations. The UCL-CS LAN is operating in multicast IP mode for conferencing and other applications. For instance there is a media server on the LAN running multicast IP (see below). The communication with the radio stations is through the UCL Transcoding Gateway (UTG) (see Section 3.6).
The CAIRN routers are BSD boxes, with specially built experimental software. Although these routers are not production items, they are much more flexible than production routers like those from Cisco. There are two QoS experimental platforms: the ALTQ one from Sony, and the CFNC one from CMU. The router and CFNC already support IPv6; the ALTQ one does not yet. We will be providing various QoS algorithms with the experimental platforms, which include CBQ and WFQ.
Coming from the MECCANO project is the UCL transcoding gateway (UTG); this provides multicast/unicast conversion and some transcoding. While the main part of the COIAS environment will use multicast, most of the mobile nodes, and some of the satellite versions, support only unicast. For this reason alone, something like the UTG has to be used. In addition, the mobile and satellites networks may be more suited to a different coding scheme than the terrestrial; hence it is desirable to have a flexible transcoding platform available if necessary.
From the ICECAR project, we will be able to provide CAs and a PKI infrastructure. This should include secured depositories to hold public/private group key pairs. It is probable that the PKI developments in this area for COIAS will be very similar to those for ICECAR. We hope also to be able to bring over some of the smart-card support. There will not be much use of IPSEC in ICE-CAR.
From MECCANO project, we will be able to bring over the secured media tools - RAT, VIC and NTE. There will also be the secure
announcement/invitation SDR. This will be accompanied by a complete set of utilities including PGP or S-Mime, which will allow conferencing groups to be set up, and group keys to be stored. The transfer of this system to the IPv6 world will be done under COIAS
We can provide MMCR, a multi-media server, from MECCANO. It will probably only be operating under IPv4; the porting to IPv6 will have to be done under COIAS. The MMCR can store all the streams of a multicast conference, and play back any specific sets of streams on command.