Drobo

Thunderbolt on Drobo

Will it happen?

http://newsroom.intel.com/community/intel_newsroom/blog/2011/02/24/thunderbolt-technology-the-fastest-data-connection-to-your-pc-just-arrived

why would you want or need it?

thunderbolt is 10 times faster than drobopro…

You usually offer very sage advice, but that comment smacks of the infamous ‘no one will ever need more than 640k’ statement! :smiley:

With the speed at which Data Robotics adopts new features it will be a while.

See the 3TB drive thread(s).

Also the implementation of iSCSI is half-baked. Only working in 32-Bit mode on OS X, 18 months after the release of OS X 10.6.

The Drobo’s would require a major overhaul to even take advantage of the speed offered by Thunderbolt.

An SSD Drobo with Thunderbolt could be blazing fast, but imagine the cost.

Damn the cost! Bring me a DroboElite with 1TB SLC SSDs! :smiley:

I took docchris’s comment to mean that it would be completely wasted if incorporated to Drobo, due to the s-l-o-w-n-e-s-s of the Drobo.

Since the fastest SATA-III drive has max 6Gbps thruput, so SATA-III is too slow for Thunderbolt and SSD is currently the only option for the future Drobo Business models equipped with the Thunderbolt but will that drive up the price of the B1200t to over US$10K w/ 12TB or less and that’s just a DAS w/o SAN capability. I rather stick to 10GbE.

Yeah, why would anyone want a simple, universal, future-proof IO interface that could scale with future disk performance gains?

Me, I’d make do with maxing out on disc speed for now, and look for faster disks in future!

Yeah, don’t forget “amazing”, “revolutionary”, “magical”, “re-invented”, “redefined” :stuck_out_tongue:

Let the Apple-effect ensue…

you forgot expensive…

Right :slight_smile: But why would I wait until the current interface I’m using is not sufficient for some future speeds of my future storage and until the new technology gets cheaper, while I can future-proof myself TODAY. You know - unplugging one cable and plugging in another is such a pain… :wink: The only reason I could think of for going for Thunderbolt now is that the $1800+ laptop has paltry TWO USB ports. But I wouldn’t buy $1800+ laptop with two USB ports in the first place…

http://www.drobospace.com/forums/showthread.php?tid=2254
:smiley:

You guys realize Thunderbolt is an Intel product, right? Apple is just the 1st adopter.

One of the amazing things being overlooked is the Thunderbolt port on the new MBP supports damn near ANY standard with a adapter. Currently, there are HDMI, USB3, Firewire, eSATA and gigabit ethernet adapters available. Not to mention it can be used with any DisplayPort device with no adapter.

While the current ‘limit’ on Thunderbolt is 10Gbps (each way) it’s real strength is you could put 4 ports on a laptop or desktop and have whatever combo of ports you wanted for your setup.

doesnt affect the fact that its going to be 1)relatively expensive and 2) complete overkills compared to the speed of any drobo

I agree with mgriffin… there are too many different interfaces and never enough ports on laptops, much less the port you need. Like everything else it will get cheaper over time and should become ubiquitous. That MBP would have been $1800 with or without that new contraption :slight_smile:

http://www.theregister.co.uk/2011/03/03/drobo_thunderbolt/

I concur.

A lot of posts seem to indicate that this is a wasted effort because Drobo is too slow. I would rather have Drobo offer a device that can take advantage of the interface speeds and then sit back and wait for the drive companies to keep up. You can get SSD drives with SATA interfaces, not cost effective yet, but it’s coming.

just to get back to basics here for a second… :slight_smile:
what actually causes the drobo to have slow limit on its throughput?

asuming its “not” the speed of the drives inside it… then…

  1. is it the onboard cpu which is calculating parity etc?
  2. is it the amount of onboard ram/cache

if 1), what is the current cpu specs, and why not simply increase it?
if 2) what is the current value, and why not simply increase it?

eg, there are special gfx cards which have tons of cpu power and ram… (i know they cost an arm and a leg depending on what brand (nvidia cough) you go for… but surely there must be a basic dual axis chart or graph like moores law that shows the current cpu power, and increases of it… vs cost and actual speed increase? :slight_smile:

(maybe it was a no-brainer, and someone at dri made one and found that to double the speed you quadruple the cost or something… but either way even it is does or doesnt exist i sure would like to see such a graph from DRI, or someone in the know :slight_smile:

From what I understand, it’s both #1 and #2, and the reason is the same as why routers usually don’t have “overkill” RAM and CPU - manufacturing cost.

I miss the days when things were “overbuilt” - nowadays you’re lucky if they’re barely built well enough to work.

I have hope that the more expensive buisness-class Drobos are built more along these lines though (to compete with Synology and others), but only time will tell whether the pricing and features are competitive enough to give them traction.

Don’t think you’re quite giving full credit there.
The information seems to be that Apple brought the idea over to Intel in 2007 and worked on creating it with them.

Thunderbolt is the productized copper implementation, with the optical version still to come.

A few links to be read, but engadget’s is representative:
http://www.engadget.com/2009/09/26/exclusive-apple-dictated-light-peak-creation-to-intel-could-be/