1 pound = 2 kg?
"An SS8460 enclosure full of drives weighs 107.6 pounds (237.2kg)"
Ah, so you plan to crash a probe onto a planet ;) (though in the historical example it was Imperial und US units that were confused I believe...)
For maximum big data density, you basically need a rack full of disk drives and little else, and DataDirect has moved in that direction by stuffing up to 840 disks into a standard rack cabinet. The idea is that HPC and big data environments need as much capacity in a storage rack as they can get. DDN reckons it offers up to 40 …
They have updated the article to say 48kg.
However, given that a 4TB WD Red drive is 0.68kg:
84 of those is 57.12kg, so this suggests the case has a weight of around -9kg :-)
Seagate 6TB is heaver at 0.78kg per drive:
So that's 65.5kg of drives per tray.
Super micro has a 4U 90 drive rack with IVB Xeons, lots of ram and flexible io. With all of the software defined everything going on, it's trivial to hook these up to a cloud control system. Throw in some Diablo MCS in the DIMM slots for flash cache, IB/RoCE for high throughput low latency and .... Tell me what DDN has special that I can't assemble from COTS parts?
Meh. More uninspired crap from limited thinkers. Chalk another one up for "groups think" failure.
Oh, and you can fit more drives horizontal than vertical. Ejecting a heavy drawer of spinning drives to replace a top load is just a stupid design. No chance of jolting one of the other drives in the drawer when the drawer comes out or clicks back into place <sic>. Where the heck did the fundamentals of physics go? Out the window with all the other basic logic and reasoning?
The vertical deployment of the drives looks sensible from the purely spatial point of view as it means all the non-drive space (power, data, cables and support) is put into one plane which should optimise the use of space. Vertical stacking would remove the need for cables in the same way that commercial removable external HDD units work (if you're in the business of swapping out HDDs, these kind of exposed external HDD "caddies" are invaluable).
I can't see any details from the picture, but if I were designing this I would combine the cooling and support elements into one form, a thin metal (e.g. thermally conductive) caddy that ensures that the drive sits true on the connectors and doesn't topple or otherwise shear or twist the connectors. It would effectively make the caddy a part of a monster heat pipe.
They will be a bugger to deal with though, particularly when you need to swap a drive in the top unit at the top of a 42U rack. Servers are annoying enough, any although these probably don't have a lid case on top to content with, the drives would have to be carefully removed to not interfere with the operation of adjacent drives.
EDIT: Just googled the SS88460 user guide and aside from the unit looking different to the datasheet model and the image here on El Reg, it has slots for pairs of drives and enclosed caddies for each HDD.
There was once a problem when a large multiple fixed disk unit was moved into place. Once the trolley rolled off the central reinforced aisle the false floor started to buckle under the 1500kg weight. Eventually that was fixed and the water cooling pipe connected.
Everyone admired the technical marvel of this mammoth measuring about 3m x 3m x 1m - and its enormous total capacity of 600MB in 1970. Archiving the contents to tape used to take 8 hours.
Apart from giving field engineers a hernia, you have to wonder what this box is for. It can't hold much SSD because there aren't enough interfaces. This means it is a secondary bulk storage system, which begs the question of why SAS redundant interfaces were provided for the drives. SATA would have been the right choice. Front to back depth is going to cause problems, as is the floor load.
This looks a bit like a solution looking for a rapidly vanishing problem. The trend to compact storage boxes with fewer drives, such as FaceBook's 1U 12-drive module, makes much more sense, both use-wise and economically.
I recently visited a large DDN installation, with multiple racks of their SFA12k storage. EVERY rack was only 2/3rds full = literally only 7 chassis not 10, with a large empty gap at the top of each (DDN supplied) rack. When I asked why, it was two reasons – floor loading of course, and also this customer deemed it TOO DANGEROUS to even attempt a disk swap requiring a whole 100kg chassis to be slid right forward on two thin rails near the top of the rack. So much for having the densest hardware…
while I also question the utility of such a heavy enclosure, any moron knows you don't design for SATA but rather SAS. personally I would have put a fan bank in the middle as opposed to relying exclusively on the ones at the back but the ODMs who build this stuff (DDN is just an OEM) have these nifty things called temperature probes and IR guns.
Horizontal drive placement is HORRIBLE for space usage and cabling but more importantly air-flow.
Nobody is going to be replacing drives in these very often. If you're not using 3x replication or erasure coding you're a world class idiot.
Bit unfair to compare to that Supermicro server as its not their densest solution available as it does have a MB. As far as I can tell this is a JBOD. Supermicro do a 90 drive chassis in a 4U. This gives you 5.4PB in a single rack (10x 4U + 2x 1U with 12 drives each). 5.2PB without the two 1Us.
World's densest storage solution? I think that one goes to Supermicro, not DDN.
Biting the hand that feeds IT © 1998–2019