GridCoord d 1

Annex K - Grids in Industry

Industrial involvement in Grid research is difficult to estimate. GridCoord therefore organized a workshop on industrial Grid usage. The workshop was held at HLRS Stuttgart/Germany on March 7-8, 2005 (Task 4.1.1). More than 40 participants attended the workshop and participated in the discussions. The workshop was organized in presentation sessions plus two extensive panel discussions. The findings were substantial to get a better understanding of industrial Grid research and approaches in developing business plans and integrating Grid into the everyday business and operation. The workshop also was contributing to our understanding of industrial integration with public research in different member states.

When aiming to get a better understanding of the level of funding for Grid activities in industry we are confronted with the following challenges:

• 
  The term Grid is well known and well established in the scientific community. It has, however, not yet fully become a term generally used in industry.
  
• 
  We need to distinguish between types of industry with respect to Grid activities. There are at least the following:
  

• 
  Industries that supply Grid software - like Sun or IBM.
  
• 
  Industries that supply Grid based services - like SAP or science+computing.
  
• 
  Industries that make use of Grid services – like DaimlerChrysler or British Aerospace.
  

• 
  Industry has become decentralized in recent years establishing so called cost centres which are independent of each other and handle their own budgets.
  
• 
  Industry is very sensitive about confidentiality.
  

Assessment

Funding for Grid activities in industry is difficult to assess. The consortium was not only confronted with the usual problems of confidentiality but industry typically is also reluctant to give numbers on money spent for certain technologies that are considered to be key to their business.

The funding for Grid activities in industries cannot even be assessed within any reasonable margin. The key problem is that IT developments in industry are usually evolutionary processes while the Grid is a revolutionary concept. Industry tends to integrate components of the Grid that are considered to be mature or to pick up concepts of the Grid that are considered to be well established. On the other hand the creation of so called cost centres in large industries makes it more difficult to assess the amount of money dedicated to a given specific IT development. It would therefore take a separate very intensive study to get information that is detailed enough to make any estimates. Such studies are conducted by companies like International Data Corporation (IDC) but are beyond the scope of any funded project activity.

Usage of and demand for Grids

However, there are some indications as of the demand for and usage of Grid computing in the market that can be described here and reflect the work of the project consortium with industry in the last 12 months.

Companies like IBM and SUN have set up separate activities to provide their customers with Grid solutions. The fact that the Grid has risen to this level of visibility inside such large companies indicates that there is enough demand from customers and hence enough willingness to spend money on the issue. Furthermore there are commercial activities not called Grid that by their nature indicate that in the scientific world they would be considered to be Grid activities. Some examples are:

• 
  “On Demand Computing” by IBM
  
• 
  “Adaptive Enterprise” by Hewlett Packard
  
• 
  “Dynamic Data Centre” by Siemens-Fujitsu
  

Large companies like DaimlerChrysler, Porsche, Bosch, and others have their own internal programs of IT consolidation. Many of the activities set in the frame of such a consolidation have to be considered Grid activities. The most notable example for automotive industries is the setting up of a virtual market place for the supply industries. This concept has all the features that come with a Grid although it was never called a Grid project.

Similar findings are valid for the aerospace industries like EADS and British Aerospace. These large industries are distributed by nature and hence require a certain level of distributed collaboration between resources and staff. Many of the issues considered to be important for the Grid (data handling, security, dynamic virtual organizations) were long tackled inside these companies.

Perspectives

The basic problems in industry are similar to those in research. The background of Grid research in high performance computing has for some time hampered the pickup of the concept in industry. Gradually, however, the potential of integrated solutions that go beyond the computing aspect have gained ground. The perspective of industrial research in Grid has to be considered rather positive. The following findings could be extracted from the workshop held in Stuttgart (Task 4.1.1):

• 
  End user industry has a variety of problems that required Grid solutions.
  
• 
  Software industry has started development in this field and has reached a level of maturity that will result in further expansion of the field both in terms of activity and in terms of revenue.
  
• 
  The weak collaboration of public research and industrial development was long hampered by the notion that Grid is something of purely scientific value. This misconception has overcome and a closer collaboration of public and private research can be expected in the future.
  

Summary

It has turned out to be extremely difficult to estimate the monetary funding for Grid in European industries. The key problem is the fact that Grid is mainly a scientific concept not yet deeply rooted in industrial IT language. On the other hand it is clear that industry has faced and is still facing problems similar to those tackled by scientists based on the Grid concept. It is reasonable to state that most innovative activities in IT in industry are to be considered Grid activities and hence the mastering of such Grid technologies is vital to the future development of European industry.

1 It must be pointed out that some changes have recently been made to the structure of the funding bodies. Prior to 1999, research funds where basically delivered by the ministries of Education and of Industry. At that point, the Ministry of Science of Technology was created, which took over most tasks related to research and its funding (among other things) previously in the hands of the other two ministries. After the elections of March 14, 2004 (and the change of government), there has been a switch back to the previous situation. The ministry of Science and Technology disappeared, but some of the internal structure and organizations were inherited (in a split form) by the Education and Industry ministries. However, part of the internal layout was also recovered from the previous two-ministries state. Despite its impact on the development of research itself, this can account for the lack (or inaccuracy) of information publicly available at some of their web pages.

2 The FECYT is an “independent, non-profit entity, with the aim of serving the Spanish science-technology-business system in an on-going, flexible way. It also helps to identify opportunities and needs, while proposing methods of action for scientific research and technological innovation.”

3 Jesús Marco started the IRISGrid initiative, previously mentioned.

4 Juan José Moreno, together with Jose María Carazo (from the National Biotechnology Center, CNB) devised the action line e-Science in the latest NRP.

5 The names we will cite here are the current ones. They are, of course, subject to changes.