As the Internet continues to shape developments in IT and as PCs become more powerful and cheaper, the WMO GTS is being presented with alternative, more cost-effective means of meteorological data handling.

The WMO GDPS has been under "attack" by forces of technological change for quite a while. The results of which has been the emergence of automated GTS and computerised data/products handling systems. Before the advent of Internet the WMO data communication system was based on the "store and forward" procedures. Under these GTS procedures it was not possible to "fetch " data/products from a source. Internet has changed all that. Using Internet technology it is now possible to configure systems which facilitate two-way data communication. This document provides details of some cost-effective implementations of such systems that integrate automated message switching and telecommunication functionality with computerised meteorological applications. Systems that have clear migration paths in response to GDPS requirement increases in scope and sophistication. The modular approach to the design of the systems playing a major role in the definition of such paths.

The most important aspect of the above system concept is migration which simply put means the ability to upgrade the system without discarding what is already there. The ability to build on what already exists. When NMCs try to constitute a system ensemble some of the following factors are to be considered:

• A migration path which takes into account the hyper evolution of Information Technology (IT) and at the same time, retains GDPS’s fundamental philosophy namely, timely meteorological data and products reception, processing and dissemination.

• Migration, which should, as far as possible, includes system elements presently on hand. RMC/NMCs should not be expected to throw overboard all and sundry of the GTS infrastructure they have in favour of new technology. Centres in RA I have already invested a lot of their meagre resources to get to where they are. So migration need to be gradual and should also provide a reasonable overlap period.

• Where there is a possibility of choosing migration paths, the path with a lower learning curve should be preferred. In this connection, not only does training need to be emphasised, but also the notion of the "black box" technology must be discouraged. Centres receiving new technology must be provided with the needed expertise to handle it, if that technology is to make the expected impact. There are cases where foreign experts have designed and installed meteorological GTS related systems without the full participation of the local personnel. In some other cases the foreign system installer has left behind no adequate manuals. All this makes maintenance of the new system a nightmare.

• So much is presently happening in cyberspace to the extent that the direction of change is unpredictable. The Internet is a case in point. It may therefore not be prudent to completely tie the GDPS GTS to the Internet. While the GTS, like commercial closed systems (CompuServe, AOL ) should continue to collaborate with the Internet, but because of the Internet’s unpredictability, the GTS should remain an Intranet, at the same time, a pace manageable to the NMCs. Internet technology should be used to provide the services and to meet the increased needs for meteorological products.

• There are many advantages in the WMO GTS collaborating with the Internet. The advantages include:

• exchange of data and products as an adjunct of the GTS;
• exchange of non-time critical information such as climate data;
• dissemination of information to the full spectrum of users;
• collaboration between Members;
• downloading and exchange of software.

• Access to information provided by the WMO web site.

• The hyper growth of Internet in terms of users and material carried on it, does not seem to match development in telecommunication infrastructure, especially in RA I. As of now (year 2000), the Internet is not particularly suitable for real-time meteorological data/products exchange. However, for non-real time, and in some cases even for real time data/products exchange, the Internet may provide a cost effective mode of exchange.

• Imagine the main RTHs being disabled through the "denial of service " or "love bug" viruses. The resulting GDPS meltdown would unleash catastrophic global consequences. Serious security considerations must be made before NMCs decide to interconnect their GTS with the Internet because the strength of a chain is determined by its weakest link. WMO would have to ensure that all NMCs connected to the Internet are secure to guarantee the security of the GTS as a whole.