For the Grid to succeed financial and political issues have to be solved

Amsterdam 07 March 2001 It is not so much technological issues that have to be overcome to make the Grid a success, but financial and political issues. Hence, Bob Raiken from CISCO proposed to use Postel's two additional layers to the seven techies use to describe the network: Layer eight (financial) and layer nine (political). The problems with for instance accounting in cross-border testbeds are very complex. To get all the organisations involved to co-operate on a political level is even harder.

According to Aiker, Grids are the next evolutionary step in the Internet. He points out that in the network evolution, from the bottom up the layers have become standardised with new technology. First there was IP with the TCP service level on top of it. Then the NSFnet, SURFnet, DFN and other national research networks could florisch. After these were established, a hierarchy of networks was formed. This allowed for the World Wide Web to flourish. A good basis for middle-ware like RFC, IAB and Globus. Grids, the new peer-to-peer applications and e-presence are the next step.

Remember there is already a lot off middle-ware available: e-mail (SMTP), www (HTTP) for instance.

What will the Grid add: Directories and schemas for instance, with naming and addressing. New location schemes to find resources and much more. The challenge here is to keep interoperability between all developed components.

The real issues, however are more social and political:

  • Costing and accounting: what should be grid resource units?
  • Resource policy management: who has authority
  • Security
  • Standards: who chooses and who decides

According to Raiken the Grid community is still dominated by the supercomputer and distributed computing community. It should pull in more people from the peer-to-peer and interactive collaboration communities

The Grid is also still mainly an academic and research undertaking. To make the transistion to industry a wide range of concerns have to be taken into account, including API's core capabilities and middle ware. Science applications are important, but are they relevant to industry?

Raiken sees that in the future there will be multiple heterogeneous Grids, just like there are many internets and intranets today.

Ad Emmen

[News on Advanced IT][Calendar][Analysis][IT in Medicine]