BAE explores new Grid horizons in Vogon projectsAmsterdam 20 May 2003 Computer aided engineering is one of the first industries to take-up Grid computing, is it not? Well in fact it is not. Collaborative engineering is a way of live for large engineering companies like BAE that is involved in many aerospace development projects like Airbus and the JSF. So much of the collaborative environment and software already predates the Grid. So it is more that engineers were doing Grid computing before the word came en vogue. Of course, as Alan Gould from the Advanced Technology Centre of BAE Systems explained last week at the HP Grid computing conference in Amsterdam, BAE is exploring new Grid in a number of VOGON (Virtual Organisation using Grid-enabled, Objective-led Networking) projects. The trial to use untapped PC-resources for large jobs did not work: the BAE applications are not suited for that. Of interest is the global resource model he developed. Showing what has been done to date - a little -, what will be done in five years - a little more -, and what must be done to make virtual organisations work - much more.
BAE works in a lot of international projects with hundreds of partners. Examples are the Joint Strike Fighter (JSF) and several Airbus A380 projects. In some projects partners are working together that are in other developments part of competing projects. BAE itself has some 80 offices in a dozen countries. For some project export restrictions apply, so even within BAE not all information can flow freely.
Hence the ultimate goal of the BAE Grid research programme: create a Virtual Organisation using Grid-enabled, Objective-led Networking. The networking that BAE, according to Alan, is concerned with, is not primarily file exchange, but rather direct access to computers, software, data and other resources including people, in order to perform a specific collaborative task.
This nature of the sharing is a little different from that in e-Science. In e-Engineering, the sharing is highly controlled, with resource providers and consumers defining clearly and carefully just what is shared, who is allowed to share, and the conditions under which sharing occurs. BAE hopes that ultimately, Grid technology will allow them to make a quantum leap in efficiency, effectiveness and agility.
To get experience, BAE set up an internal Grid, called - yes, you guessed right - BAEGrid. This gives them the opportunity to support internal R&D initiatives and be able to access the UK e-Science and EU IST programmes. BAE also partners with and tries to influence the vendor community.
In 2002, BAE did a foundation study trying to make IT work harder by federating local systems. The idea was to use grid software to hunt down and use idle computing time. This looked like a good idea at the time. BAE owns around 63,900 desktop machines, said Alan. At an average of 200 Mflop/s and 256 Mbyte this is a distributed resource of 12.7 Tflop/s with 16.3 Tbytes of memory. Acquisition cost for a facility of this size would be GBP 150 million. So every percent of better utilisation would represent GPB 1.5 million.
Well, it does not turn out to be that way. The applications are not ready for such an environment, neither are the processes in BAE. BAE is working in a number of Grid research projects with partners, mainly in the UK. One of the short term issues that Alan finds relevant is the lack of Grid-enabled collaborators. This is much like the first phone syndrome.
Semantics is key for automating data transfer and business processes. What is an airplane to one person is an aircraft to someone else. What is an inch to one individual is a centimeter to the other. You better do the concersions automatically and you better do them right. Business processes must also be Grid-enabled before Virtual Organisations can function.
Alan summarised this in a global research programme:
Most of which is "unexplored territory". And even in five years time, Alan expects not more that half of the necessary reasearch will be finished. But he is not discouraged by this and considers it a very good time to get into exploring the Grid. There are short term benefits that can be delivered. There are grants available and at this stage of development, there are opportunities to monitor and influence the emerging Grid future.
Ad Emmen [News on Advanced IT] [Calendar] [Analysis] [IT in Medicine] |
||||||||||||
 |
||||||||||||