As promised here is the video:
And here are the slides:
(Refresh the browser if these doesn’t load on first go, WordPress seems to sometimes fart at iframes.)
These are from my inaugral research presentaion done last Thursday. (17/03/2012) You can tell, as previously admitted that I hadn’t sorted a script out. I still don’t think this is a bad thing however I do say ‘urm’ loads, which is something I guess I am going to have to work on! Also I need a haircut, yes.
Enjoy and feel free to comment, flame or ask any questions!
Finally managed to find the time to watch this all the way through, and I think you did pretty well mate! After all, what you discussed here is essentially based on a literature review, and its always a bit nerve racking to present on the basis of that… mainly because you never know what type of questions your going to get. I’ve seen a few people get a rough ride from questioners who point at all the stuff they haven’t read! The question about data transfer at the end was interesting, mainly because I’m thinking about similar problems in my work. Like CERN the data from Jodrell (and other observatories) is dispatched via tapes and delivered directly to data centres. This is really not ideal, but its going to get so much worse. When built the SKA will gather ~430GB of data per second. Thats a lot of tapes per day!
Cheers Robl. I watched it the once with a grimace. I suppose you can be more critical on yourself than others unless you’re deluded by yourself. I am not so sure the recent awareness of BigData is actually fit for purpose. As opposed to pointing out that we’re generating a lot of data now and asking, “How FartCorp and BigGov handle it?” we should instead be saying. “I need to push a exabyte from 200Miles above the surface of the Earth to my home PC. That’s over air, land and probably water. What do I need to do this? How soon can this be in my home?” I guess I am saying whatever we are trying to cater for now is only going to be a short-term solution to an evolving and escalating problem. Maybe we are posing this scenario and these questions already but all I see is people messing around with the frequency of light in open spaces and that just has too many associated risks and limitations. During the literature review I wanted to get into some papers on SoSs in deep-space communications but there were very few and the best weren’t very forthcoming.
I agree mate, the BigData challenges don’t appear to be getting solved quickly enough for us to reap the benefits any time soon. I’m glad you’ve said this, I thought it was just me that had these concerns. I keep reading things that assume the problem will solve itself will the arrival of better and faster disk media, bandwidth and algorithms. But the exponential accumulation of information can’t really be addressed by a Moore’s law style rate of progression. While the likes of Google can probably solve many of these problems with financial muscle, those working in science for example don’t have that luxury. Can you imagine the cost of upgrading the network infrastructure of most western economies to say 100 mb or more? I’m sat here with 10MB and nearly ten years ago now I had 1MB – pretty paltry increase considering whats coming our way! Definitely only a short term view in the literature I’ve come across.
This is it, regardless what our favourite TV physicist says about Moore’s Law being a thing of the past in a few years it’s demise doesn’t account for the sheer uptake of what we’re are able or prepared to collect, analyse, index, query and fuse. I was going to mention this earlier but thought I was brain-farting-on.
Complete re-engineering is the only way forward. Not talking about just speeding things up. Interesting, there’s a current project to link Europe via the UK to the pacific with a big pipe across the northern shores of Russia. When complete it’ll be the first time we’ve had a route into Eastern-Asia without any hops through America’s One Wilshire building!