Scientists Built the 'Hubble Telescope For the Ocean' Using the Cloud
New submitter stacey7165 writes "VMware shared the application architecture story of how they worked with the Ocean Observatory Initiative to build a 'Hubble Telescope' of the ocean. It's comprised of a massive network of global, regional, and coastal sensors that send information to a common framework called the Common Operating Infrastructure (COI). The COI resides in a hybrid cloud powered by VMware and Amazon. To cope with a total of 49 classes and over 700 instruments deployed off of 6 coastlines, and a variety of consumption use cases, the Ocean Observatory Initiative built out the system using a variety of sub-systems loosely coupled through a messaging system powered by RabbitMQ called an 'Exchange.' Organized into a system where message clients pubsub from 'Exchange Points' and 'Exchange Spaces', the system is easier to maintain, extend, and scale. According to the OOI's documentation on release 1, the Integrated Observatory Network uses AMQP 0.9.1 and RabbitMQ-Server v. 2.3.1 on CentOS 5.5."
I can't even IMAGINE that kind of scale. It's almost humanly impossible to conceive of handling that much data input with a mere computational device. Why, it must take an army of bright young men just to replace vacuum tubes in such a beast each day! Truly a miracle of American technology, know-how, and gumption! Let us see the Bolshevik match THAT!
They made an ocean-scanning telescope using nothing but clouds?!??!?
HOLY SHIT THEY'RE AIRBENDERS!
Hopefully, as scientists, they know that their 'cloud' is only as resilient as the underlying hardware, regardless of number of OS instances running on it.
Cool specs, but the real entertainment is trying to figure out how its like the :
'Hubble Telescope'
Do they mean the equations to design it were wrong and not tested to save money?
Or they mean the transport machine that installs and services it has been decommissioned?
It makes pictures?
Its gyros burn out on a regular basis (disclaimer, I've not been tracking this for years)
Ah they must mean its just one machine a single observational instrument. Oh that contradicts the whole rest of the write up.
Um... its a 1980s hardware design?
Oh no wait I know, it costs a billion dollars each time it's serviced, whereas it would only cost a quarter that to build another and launch it on a non-reusable booster. Yeah that must be it.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
So how will they get past Microsoft's copyright on "Exchange"?
I think I have hit buzzword bingo! where's my prize ?
Because they can, the CEO's (Internets, serious business! lads) love it and it's the fashionable thing to do.
Not because the aging Dell in a forgotten corner of the office is perfectly capable of handling the input from 700 measily sensors. For only $xx.99/mo you can cloudify it, CEO's will be 500% more impressed than if you pointed to a Dell in the forgotten corner because the cloud is so abstract, so future, so new and some annoying sales grunt from a cloud provider with a nasally American accent already convinced all the CEO's their organisation is incapable of running a server thats any bit reliable.
I understand people with little computer knowledge using the term cloud, but really can we stop using that stupid marketing term someone got from a flow chart.
It uses the internet maybe, but nothing uses "the cloud" as "the cloud" is just an abstract idea for marketing purposes.
I don't know if you guys know this, but advanced civilizations are easily destroyed by small amounts of water. Pls see this link http://www.imdb.com/title/tt0286106/ Stay away from oceans and clouds, ... and M. Night Shama-lama-ding-dong movies.
There is a special place in hell for people who mod down comedy.
While integrated data analysis can be amazing to work with in many fields, and having a data store and processing that is well organized is miles ahead of working with a poorly designed data store ... this seems like a lot for the sake of marketing or just overkill. Having worked on various plasma physics experiments, they need to relay and display data from as many sensors on a shorter time scale, and it was simpler than this.
Dude...
They said Hubble, Cloud, application Architecture and COI in the same paragraph. Drop your pants and suck VMwares dick immediately. They have buzz-worded you into submission.
in Soviet Russia, ocean watch YOU!
Yes, people often confuse inches with centimeters...
Better luck next time.
8 terabytes per day, according to TFA. Add the ability to effectively query, slice, dice and present that much data on a long-term basis... yeah, that's a hell of a lot of vacuum tubes.
For comparison, the LHC does upwards of 27TB/day[1], Hubble 3-5[2].
[1] http://en.wikipedia.org/wiki/LHC_Computing_Grid
[2] http://hubblesite.org/the_telescope//hubble_essentials/
~700 instruments, running 24/7, generating an aggregate of 8 terabytes per day means each instrument is generating about 142KB PER SECOND!
"You're Doing It Wrong!"
Well, looking at the actual instruments, things like hydrophones, pressure sensors, broadband seismometers, and such, what instruments actual generate that bandwidth? I just don't see it. Seems to me that they're at least a couple of orders of magnitude off.
Seriously, its not that hard:
Parts comprise a whole.
The whole is composed of the parts.
VMware shared the application architecture story of how they worked with the Ocean Observatory Initiative to build a 'Hubble Telescope' of the ocean.
Apparently you can build telescopes out of software now. No need for expensive optics.
It's comprised of a massive network of global, regional, and coastal sensors that send information to a common framework called the Common Operating Infrastructure (COI).
No it's not. It might comprise that though, but it's still not a telescope.
Cant it detect a Russian submarine at US shores?
I know VMware is well known for the vacuous and unintelligible products of its marketing department but this one really has to take the cake. How does this software/hardware deployment even remotely resemble the Hubble space telescope? How can this "cotton candy" style marketing actually be effective? Who could possible be willing to waste their time consuming such a large puffball of words for a few grains of substance? And again, how the hell does this resemble the Hubble space telescope?
Two of my imaginary friends reproduced once