It will address the following questions:

• What is wrong with things as they are now?
• why do we need to change?
• why is it necessary to push for this?
• why can't we just leave it to the market to figure it all out?
• why can't we leave it to incremental change?

Some of the following issues should be included

Abstract

We are facing a crisis due to a shortage of skills and knowledge in the industry. If this is addressed there is a win win win situation. Students gain employability. Employers gain critical-skilled staffed. Educators and Trainers gain a market for their work.

Crisis in Distributed Systems

We are currently facing a crisis in distributed systems akin to the crisis faced by software in the 70s. At that time programs were becoming very large, but techniques and methodologies for writing maintainable, well-designed software were not well known. University Computer Science departments were reluctant to teach these Software Engineering subjects however, as they did not see them as true Computer Science. Instead their focus was on algorithms, programming languages, etc.

We now have a similar situation where the techniques to provision, manage and adminsiter large distributed systems are not well known or understood, but universities and professional bodies have little interest in teaching these essentially System Administration topics.

In this document we will attempt to provide evidence for why this attitude must change.

Skills shortages in Industry

Current Industry Needs

{ taken from the draft document on t-Infrastructures by R. Barbera, E. Giorgio, M. Fargetta & G. Sipos }

Grid technologies are developed from University and Research centre that are also the main users of Grid infrastructures. To move the Grid into the business world it is very important to prepare the employers with ad-hoc training courses fitting the requirements of the companies.

A deliverable of EGEE NA3 working group [11] presents the results of a survey investigating the interest of companies in Grid and in particular in Grid training, with special questions to evaluate the feasibility of commercial training courses. The examined companies are differentiated per business area (IT, software, legal, Power, Oil, Finance, education, etc.) and size (from few employed to more than five thousands), and they are really representative of the United Kingdom economy, the country where the survey was made.

The majority of examined companies already knew about the Grid but only 38% use it, even though 96% of respondents assert that their business could have benefit from the adoption of Grid technology. Among the issues preventing Grid adoption, the lack of skills has an important role for 24% of surveyed companies. In more detail, the companies ask for courses aiming at teaching how to use this new technology and how to create an own infrastructure (installation and administration). In these two contests a permanent infrastructure where new users may have the opportunity to perform tests and understand how it work is seen as very beneficial. Moreover, a second issue for the companies is the lack of interoperability, so it could be useful to show how different middleware can communicate/interoperate among each other and how it is possible to move an application from one middleware to another middleware with a little effort.

Finally, it is surprising to see that a lot of companies prefer commercial training, even though free training courses are provided, because they are available to pay to train their employees with very “ad hoc” courses.

Considering this request for commercial training there are new Grid companies providing training on application development, site installation, business analysis and whatever could be related to Grid. One of the most important is the polish Gridwisetech [12] that provides standard Grid courses and special courses where they start to develop a Grid application with employees of the custom in order to realise the knowledge transfer.

We are seeing a high-turnover among senior technical staff because companies are not satisfied with their preformance. The inevitable Failures in large distributed systems and inconsistent behaviour in distributed systems, etc. are contributing to this.