Well, some details were announced this week for September’s GridWorld conference, and it seems like there will be no let-up in the Grid community’s push to achieve commercial acceptance. After all, even though this year’s conference will combine GGF18 and GlobusWORLD, two events that could very easily make for a very academic conference, GridWorld 2006‘s theme will be “Achieving Agility, Efficiency and Innovation in the Enterprise.” It should be very interesting and, with the GGF/EGA merge expected to be complete by then, we should get our first glimpse of Grid’s new unified front.
Also in this issue, I got a chance to ask some questions of Wilson Rivera of the University of Puerto Rico, Mayaguez, about his upcoming presentation at Gelato ICE (April 23-26 in San Jose, Calif.) and about the Grid testbed his Parallel and Distributed Computing Laboratory has deployed. It’s an interesting Q&A, and a nice little glimpse into how Grid technologies are being used, and in this case researched, in Latin America.
On the distributed computing front, good samaritans around the world have a new project on which to donate their PCs’ spare cycles, as the Rothberg Institute for Childhood Diseases announced the release of the first avian influenza target to its Drug Design and Optimization Lab distributed computing project. This announcement provides a great background on the project, and I’ll be speaking with someone from the institute next week, so stay tuned for more details.
And, in the world of SOA, we saw an interesting announcement from HP this week. Specifically, we found out the company will be rolling out industry-specific SOA solutions for the financial services, manufacturing, public sector, and network and service provider communities. Considering how many companies have already deployed SOAs, it seems unlikely this venture won’t be successful, especially as newcomers try to improve upon the successes of the early adopters by deploying even more targeted solutions.
Finally, although I spend nearly every day of the week reading and speaking about Grid computing, I rarely have a use for it in my day-to-day life (believe it or not, running Word, Firefox and a publishing solution simutaneously does not warrant a dedicated cluster). That said, I spent some time last week experimenting with the Sun Grid, specifically on the Cepstral text-to-speech application currently being hosted on it, and I was impressed with the results: I was able to translate a 2,300-word article into an mp3 file in well under a minute. To be honest, I don’t know how long it would take to complete the same task without a grid, and I still need to learn SSML to add some humanity to the voice, but it was impressive nonetheless. And, here’s the real kicker, it used a total of .027 CPU-hours. While Sun isn’t gracious enough to charge me 3 cents for this particular job, a buck is still a pretty good deal. I’m inclined to believe your average novel could be translated for under $2 (something to think about in a society where podcasting is so popular). While I can’t comment on the price and time expenditures one might experience running a common HPC application, my personal experience was satisfying, albeit on a very small scale.