The very first sentence of this article mistakes Terabytes and Petabytes. I used to dismiss the entire article as poor quality on seeing a mistake like this. But these days it also feels like an indicator the article was written by a human and might actually have something interesting to say.
Sadly not in this case though - the Kioxia drives are interesting, but the fact that Dell has put some in a box is much less so.
There's been a lot of talk about orbital DCs lately, but with these levels of density, orbital CDNs might be a more obvious usecase. It would be interesting to see if something like Starlink can use something like this to cache media content and reduce their overall data moving through the constellation. It could even be worth it to have some satellites in higher orbits (even GEO if the ground hw can reach it) dedicated to streaming media content. You can tolerate higher RTT for content that doesn't need to be real time.
What exactly is the selling point for an orbital data centre?
Satellites for connectivity to ground-based data centres seems to make sense, but that seems like it's relatively light hardware compared to a rack full of compute and storage.
If we're looking for apocalypse-scale redundancy, we already have all sorts of ground-based infrastructure and multi-continental designs on earth already.
I wonder if it's some weird play in the seasteading style-- if they put the data in space, it's outside the reach of any government except ones willing to jam the airwaves or fire missiles capable of reaching orbit.
At current enterprise NVMe prices, the drives alone for this must easily push past the $500k to $1M mark. It's fascinating to see this level of density, but it’s strictly going to be hyperscaler or high-end defense/research budget territory for a long time.
This is one of the case limited by PCIe speed, sharing it with SSD so Network could only do 5x400Gbps Network. This is on PCIe 5.0, luckily we have 7.0 spec ready and 8.0 is even in 0.5 draft status.
If we could somehow increase the density further by 5x, we would be able to store 1EB in a single rack.
The most interesting part to me is the last sentence.
>Scality tells us it’s working on supporting a future nearline-class SSD from Samsung, viewed as an HDD killer, with similar or even larger capacity and a roadmap out to a 1 PB drive.
Finally a HDD killer. May be in another 5 - 10 years time. The day of everyone having an SSD NAS / AI Cloud at home will come.
10PB is probably the amount of data a medium sized country can collect about its all citizens (basic details, work history, all taxes, all financial records, all medical records, all police records, all biometric records and more) for their lifetime.
I think development like this might get many public sector focused firms sweating.
All the increases in density are impressive, but they come with the downside of repairability and recycling difficulties. I hope we can still repair this when parts of it break or at least recycle it properly. No matter how high tech it is, eventually this will break.
Look, out of all the BS that's come from the current AI CYOA (bubble, revolution, scam, whatever floats your particular boat), the one thing I'm optimistic for is the absolute glut of memory that'll be available once this cycle wraps up. I'm salivating at the thoughts of finally producing memory in such quantities that we can retire HDDs for bulk storage in all but the edgiest of edge cases; of more solid state tech seeping into consumer and business devices, enabling us to compute locally again instead of leaning on cloud providers for extra storage. Hell, I'd just be happy to see 4TB TLC or QLC SSDs coming in under $100 with endurance ratings that aren't garbage.
C'mon, 60TB SSD NAS for my media collection, or 20TB external SSD for backups. Get that density up and those costs down, already.
24 comments
[ 2.4 ms ] story [ 45.9 ms ] threadI feel like we’re in that season.
Sadly not in this case though - the Kioxia drives are interesting, but the fact that Dell has put some in a box is much less so.
Satellites for connectivity to ground-based data centres seems to make sense, but that seems like it's relatively light hardware compared to a rack full of compute and storage.
If we're looking for apocalypse-scale redundancy, we already have all sorts of ground-based infrastructure and multi-continental designs on earth already.
I wonder if it's some weird play in the seasteading style-- if they put the data in space, it's outside the reach of any government except ones willing to jam the airwaves or fire missiles capable of reaching orbit.
Satan’s NAS!
If we could somehow increase the density further by 5x, we would be able to store 1EB in a single rack.
The most interesting part to me is the last sentence.
>Scality tells us it’s working on supporting a future nearline-class SSD from Samsung, viewed as an HDD killer, with similar or even larger capacity and a roadmap out to a 1 PB drive.
Finally a HDD killer. May be in another 5 - 10 years time. The day of everyone having an SSD NAS / AI Cloud at home will come.
I think development like this might get many public sector focused firms sweating.
C'mon, 60TB SSD NAS for my media collection, or 20TB external SSD for backups. Get that density up and those costs down, already.