Tuesday, October 09, 2007

APAC07 - APAC Conference and Exhibition

Advanced Computing, Grid Applications and and eResearch
8-12 October 2007 - Rendezvous Observation City Hotel, Perth, Western Australia

Opening Ceremony

John O'Callaghan, Chair talked about the origins of APAC and introduced the addresses. John Zillman, APAC board, welcomed everyone to Perth, bantered about the weather and spoke of the past, present and future of Apac. Francis Logan discussed the future of supercomputing. Finally Mal Bryce talked about the history of Internet. After proclaiming that he was the "oldest rat in the barn", reminisced about past technologies that he personally lived through. Mal describes iVec and Apac as some sophisticated computer concepts, big pipes, big grunt, big warehouses and a great deal of people development.

The first keynote was Thom Dunning, who also presented on the previous day at the Student Forum. Similar topics (in truncated form) were covered. The stress was on future strategic directions of super computing/data intensive computer on the petascale. Examples used were mainly Blue Waters and synoptic telescope.

The second keynote was Mike Netzband. He spoke about high performance computing at Chevron, the challenges of HPC, innovations and frontier projects. From a HPC point of view, Chevron most costly processes are seismic imaging and processing and reservoir modelling. Processing and storage started with IBM mainframes and punch cards in 1982, to super computers and now Linux desktops and other mixed environments with multiple tiers of vendors for equipment with dual, quad core clustering. Challenges that never seem to go away are expandability and improving the system is a constant process. Innovations need to be adaptive to constant change, effective with clear purpose. Current targets include collaboration with third party partnerships and more investigation in grid computer and alternative high performance processing.


The most obvious and distinctive aspect of this conference was the large number of industry sponsors. While I knew there would be some key industry representatives present, I was not prepared for the large number of displays, tables and marketing stands at the venue. Lunches was sponsored by Intel, IBM had a strong presence, SGI were giving away USBs, caps as well as ice cream while ISA seemed to be everywhere. While the first three I had heard about (big names), the third I wasn't so sure. These were the gold sponsors and after lunch, they was a whole session dedicated to presentations from the Gold sponsors. Of course there were other sponsors, the most memorable for me were Cray and Sun Microsystems. After some discussions with other attendees, it was suggested that APAC conference attendees come from the companies that account for majority of the big buyers of large processing power. So it makes sense that many of the suppliers of large processing power would want to be here promoting their hardware. It is mostly hardware on display/advertised. We decided the organisers have done very well. Marketers get what they what (advertising to their target demographic) and conference attendees get what they want (sharing of work/research, more information product availability that may help their work/research and free ice cream, golf tees and usb keys).

Other than the marketing from sponsors, the conference ran as per norm: parallel sessions, pre and post conference workshops, presentations of work, promotion of collaboration... It was very interesting to see what other people were doing in industry with the techniques developed in research. Even before the conference started, there were iVec and AARC open houses where we were shown demonstrations of some existing projects and hpc applications. It's a very good conference for large scale applications, processing of large amounts of data and distributed management of information and resources in industry.

No comments: