Posts by ManOnTheHill 11 publicly visible posts • joined 27 Feb 2013
Yeah, it'll probably cost that much, but this isn't aimed at individual users - at least not until flash prices drop pretty dramatically. But the corporate/enterprise world will eat these up, even at current prices; there is a hunger for flash in that environment that is truly astonishing.
It used to be true that a laptop drive couldn't take that kind of treatment, but most of them today will withstand 100G while operating. But your point that the SSD is more reliable in a more dynamic environment is a good one.
Problem with SSDs taking over the storage world is that there isn't enough FAB capacity to replace all the rotating rust being built out there, at least not for the next 5-10 years (given current growth projections for storage and flash fabs coming on line). Look at the details of the numbers on rotating drives sold; total hard drive sales were around 500 Exabytes in 2015 (http://www.anandtech.com/show/10098/market-views-2015-hard-drive-shipments), with the average disk size increasing from 1TB in 2014 to 1.4 TB in 2015 - with the per-unit cost staying approximately the same. Probably not a huge monetary growth industry, but the growth in storage demand will keep HDDs in the market for quite a while.
Total shipped SSD capacity in 2015 was somewhere around 40 Exabytes (hard to find trustable numbers on this), with the SSD penetration being primarily desktop/laptop (for all the reasons you suggest).
SSDs are getting much greater adoption in enterprise (server) environments these days (density is one of the big reasons; there's talk about 30TB 2.5 inch SSDs before the end of 2017) but there still just isn't the fab capacity to replace rotating drives for all storage yet. Seagate is behind in this particular race to flash, and may be ageing out of the race as a consequence; WDC made a smart move in picking up SanDisk, now to see if they can leverage that to stay ahead of the game. (BTW, Samsung has something over 50% of the SSD market, and virtually none of the rotating disk market; arguably they made the best move to flash of all the HDD manufacturers).
>> They could also stop that "calculating" bollocks on file transfers and simply display a throughput graph.
Have you looked at Windows lately? Not only do you get a throughput graph for a file copy, you have the ability to pause the transfer - which is very useful if you're doing multiple transfers at the same time to a slow destination (or from a slow source).
I find the current file copy progress indicators on Windows to be far more useful than the stupid analog progress bar Linux provides...
Actually, there has been a TCP stack in Windows NT since 3.51 (maybe even 3.5); it was an OEM-ed stack back then, but it was there. I don't recall whether or not you had to pay extra for it, but...
(That stack, and all of the other OEM TCP stacks on Windows (NT and otherwise), ultimately lead to the development of the Windows Sockets interface spec, which we should all be happy about - it may not be quite the same as *nix sockets, but it unified protocol access on Windows, which was a Very Good Thing)
*(and I couldn't agree more about NetBIOS and NetBeui (NetBIOS is the interface, NetBeui is the protocol used in early Windows NT versions). NetBeui is a ghastly protocol (holes in the state machine!) from a bygone era that we're well shed of, and NetBIOS being retired is a good thing as well.)
I agree that the time for the rugged individualism of IT workers (an oxymoron if ever there was one) is probably at an end. The laws have definitely moved in the direction of favoring profit for the company over worker rights and freedom, and need to be changed. Organizing IT/Tech workers and using the organization to enact local-worker-friendly legislation would help move hiring and retaining locals in the right direction, as long as it made offshoring less attractive than hiring locals.
That said, finding local workers (US in my case, but it's an issue in the EU in places as well - Ireland, for example) is difficult, because in many cases the workers-to-be have decided that tech (or STEM, if you wish) is not where they want their future to be and have hared off to MBA studies or such. Until there is a program in place to get locals trained up to the necessary levels, the friction between the offshore and local workers will remain; it's hard to say you've met your fiduciary responsibility to shareholders if locals cost you twice or three times what the offshored work does. Especially if you can hide the screwed up results of offshoring or hiring unqualified but cheap workers behind reorgs/rebuilds for years.
How about this? Keep the H1B/L1 programs in place, but for every worker hired that way the hiring company must fund a training program for locals (standards and length to be set by an independent organization) and graduate (not just start training) at least one worker. Details to be worked out, but the local should get a minimum wage while in the training and be given some preferred interview placement once graduated successfully. The companies don't necessarily have to replace the H1B head with the local, but they must at least interview the local for the jobs the H1B workers are being considered for. Tax breaks for all, of course, to make the programs at least marginally acceptable to corporations and politicos.
I don't like the idea of completely cutting the H1B programs, because I have worked with some very good, very smart people who got here that way, and I think it's very worthwhile to keep that door open. But I also think that offshoring should be discouraged as much as possible; companies say the local talent isn't there, let's grow the pool.
You need to look at the Kinetics documentation (it is available at the Seagate website). The Kinetics stuff actually implements an object store for the Ethernet attached drives, and will do HA and so forth on them. I don't know how easy it would be to make a Ceph cluster out of them (haven't spent much time looking at how Ceph uses drives).
I suspect that Ceph would be better served by the HGST version of this, called Open Ethernet Drive Architecture, which simply pushes a Linux distro down into the Ethernet-attached drive, where you can do whatever you want with it. So maybe each drive can become a member of the Ceph cluster...
Oh, and the drives that I've seen all have the standard SATA/SASA connector with Ethernet mapped to them, and implement the SAS-style multiple connections for the single connector.
My vote will be get rid of the lot of them. But it would be nice to see this lead to something like reform in the way the board works - maybe a non-executive employee member of the board? Or two?
Wait, that can't work, none of those HP employees are smart enough to understand how the board (and executives!) are protecting shareholder value by buying 11 billion dollar companies that are worth 3 billion.... Or was it reducing the value of an 11 billion dollar company to 3 billion through mismanagement in a year and a half? I can't figure out which one really happened!
Idiots. Not much the board did made sense while I was an HP employee and nothing they're doing now makes sense.
Let's not forget that Meg was part of the board that approved the Autonomy acquisition, so she's got blood on her hands too. Seems only right to me that all of the board and upper management that had any approval say in the acquisition of these companys should pay the price for their failures - after all, they were more than ready to revel in how wonderful they were if they had been successful (and take more than their share of the profits from that success).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
