Yes, I should have clarified. On site storage for the average user is decreasing, there is not question about it. Off site is increasing greatly though. What I meant to say was overall, local and cloud, the average users data "collection" is growing quite a bit.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Faster 'NVM Express' SSD Interface Arrives on Retina MacBook and OS X 10.10.3
- Thread starter MacRumors
- Start date
- Sort by reaction score
Yes, I should have clarified. On site storage for the average user is decreasing, there is not question about it. Off site is increasing greatly though. What I meant to say was overall, local and cloud, the average users data "collection" is growing quite a bit.
Hey guys wasn't this thread about NVMe support in OS X? That's a really good thing for the future of SSDs and it's great that OS X now supports it. It wouldn't surprise me that the next round of Macbook Pro redesigns come w/ PCIe SSD with NVMe drivers. Just think of how much money a desktop storage solution of the highest caliber would cost and imagine the next Macbook having similar speeds, but on a 2.5lb laptop. That's pretty cool.
Spot on. For the general consumer these days, SSDs are perfect for a main system drive and a secondary drive (like games). Otherwise, HDDs are perfectly fine for simple storage needs like photos, documents, and older programs.
Glad to see NVMe advance forward. The controller as been around for about a year, but has seen very limited implementation by mobo manufacturers. It's about time we left AHCI behind.
Prices WILL drop, reliability WILL increase. Capacity WILL increase. Durability, speed/performance, power savings I see no reason other then price to not go full SSD.
Sure HDDs are cheaper BUT in my opinion/case I would rather have the benefits of SSDs over rotational media.
I don't understand why anyone would want to stay with rotational media.
Sure HDDs are cheaper BUT in my opinion/case I would rather have the benefits of SSDs over rotational media.
I don't understand why anyone would want to stay with rotational media.
That's sort of my point, although it's as ridiculous to assume that everyone has the same 20TB+ home storage needs that I do as it is to assume that nobody needs more than 200GB of storage on their computer. It's a spectrum, but there are probably pretty large bins into which people fall.
My little statistical sample of about 32 business users says that about 12% have >1TB, moderate-speed data storage needs.
But here's an article from Business Insider that gives some actual statistical figures, the only one I managed to find. They say that in 1986 the average user had 500MB of data, and in 2007, the average was about 44GB.
If you do the math on that, it works out to about a 24% increase annually. There's no guarantee it's a linear increase, and that doesn't take into account the rise of streaming and cloud storage--which I suspect have a significant impact on local storage needs for average consumers--but if you assume the same 24% annual increase we would be at an average of 256GB per user today, and will be at 750GB per user in 2020.
Of course, that's an average--some people have less, and some have much, much more. But unless the curve has accelerated drastically in the last few years--which doesn't seem likely to me--this backs the general observation that a substantial fraction of consumer (and business) users need 200GB or less of local storage.
What would be really interesting is to see the bins--percent of users who use 0-100GB of local storage, percent who use 100-200GB, percent who use 1-4TB, etc. I'm sure somebody has data on that, but I can't find it online.
This 2014 article does have a graph showing price/GB of bulk HDD vs. "mission critical" (server) HDD vs. server SSD vs. consumer SSD prices. GB/$ would be an easier graph to read than $/GB, because of the asymptote approaching zero, but it does illustrate the comparatively steeper decline of SSD prices, and indicates that server cost parity should theoretically come around 2017, while consumer bulk storage parity will take longer. It's not clear to me from that graph whether SSD has been increasing the number of GB per $ by at least 24% per year, but if so, then their value proposition should remain viable.
This little statistical analysis is actually more interesting to me. It was done in early 2013, and if you skip down to the conclusions, in a "continuing the linear decrease" scenario, extrapolating 2 years to Feb 2015 should have seen 3TB HDDs costing $15 and 25TB HDDs costing $125. This obviously didn't happen. The more reasonable curve fit shows 2015 3TB drives costing $50 and 8TB drives at $125.
In reality, the cheapest internal 3TB drive on NewEgg is $90, and 6TB is the largest general-purpose drive available, costing $250 for the cheapest WD Green. The only 8TB drives on the market right now are the Seagate Archive, with its unusual performance characteristics due to early SMR, at $300, and helium drives, which start at $600 for 8TB.
None of which is to say that future developments like HAMR, MAMR, and TDMR might not re-start a sharper curve for HDDs, but by just about any metric the last few years have seen a considerable flattening.
----------
That, I'd certainly believe, and agree with--you'd be hard pressed not to. Although even with cloud storage, the fact that much of the media used is now "generic"--that is, Apple only needs to maintain one copy of a movie to stream it to a thousand "owners" watching the purchased video on their AppleTVs, or one copy of a song to stream it to a thousand listeners--the actual bulk data storage needs of cloud services probably aren't increasing as quickly as the theoretical data "ownership" or streaming use of consumers.
Personal photos and videos backed up to the cloud, of course, as well as the proliferation of shared media via YouTube, Instagram, et al, are unique and will continue to grow in size and overall volume.
Last edited:
It depends on your definition of "want". I have about 30TB of storage at home, roughly 50% of it full of data. Only a small fraction of it will benefit noticeably from the speed and random access advantages of an SSD, and I have that on internal SSDs. The rest, even a single HDD serving data at 100MB/s give or take is sufficient to not "feel" any difference.
For this use case, sure, I'd want to have everything on silent, super-fast SSDs. But the cost would be prohibitive--what cost me maybe $1500 including RAID cases to put together today would optimistically cost well over $5000 to do with SSDs. I have $1500 to spend on storage; I don't have $5000. And that extra $3500+ would get me marginal noticeable speed increases and a slightly quieter workspace for data that's "bulk, archival storage" anyway--I would rarely even notice the advantages afforded by SSDs. I would much rather spend the $3500 on a new 5K iMac and MBP.
I'm not an average user, however. And I'd be equally crazy to not use any SSD storage for OS and regularly-accessed data just because it was more expensive per GB--the speed advantages are huge for that segment of my use case, and well worth the cost.
Based on current trends, SSDs may eventually hit the price point where they have an advantage even in cost--it's impossible to say for sure, but the price curves look good. Currently, though, there are some percentage of users--no idea what it is as a fraction--who "want" rotating storage simply because it gives them the bulk data storage they want or need more than speed at a lower cost. Some people seem to have a dogmatic attachment to it and a blindness to how things might change in the future, but today, for people with a limited budget and large data storage needs, rotating storage has an advantage.
There's also one other niche use case of extremely high write throughput server environments, although modern SSDs are getting to the point where they're reliable enough that the replacement cycle is shorter than the expected media lifespan.
Back on topic: I have been playing with my rMB and I don't think the drivers for NVM Express are ready for prime time. I suspect the hangs on start up and couple of times during heavy disk access have to do with probably a "new" driver. It seems to lock up the kernel and cause the spinning beach ball of death...
I was over at the Apple Store today and one of the salesmen there told me that the new 13" Macbook Pro is using the same new faster version of the SSD. I have heard from others though that this is just on the 15" Macbook Pro... Confused 
I was over at the Apple Store today and one of the salesmen there told me that the new 13" Macbook Pro is using the same new faster version of the SSD. I have heard from others though that this is just on the 15" Macbook Pro... Confused
Four channel PCIe and NVM are not the same thing. All 2015 MacBooks except the rMB and the cMBP are now using four Channel PCIe SSDs, supposedly including the 11" MBA.
Last edited:
I don't know whether they're the same part numbers or even the same specs. Both the 13" and 15" rMBPs use four-channel PCIe now. The 15" gets higher sequential read and write speeds than the 13". Both are double the speeds with the pre-2015 two-channel PCIe SSDs.
The main (only?) difference between the 2015 13" and 15" is the that the 15" supports pci3.0 where each lane offers an 8/Giga-Transfers-a-second (GT) connection instead of a PCIe 2.0 5/GT connection.
That's not because it's new. The Haswell (and ivy bridge) quad cores support PCIe 3.0 while the dual cores usually stick at 2.0, although I do not know why.
(see http://9to5mac.com/2013/11/04/lates...ing-ssd-performance-thanks-to-4-channel-pcie/ for more)