Slashdot Mirror

← Back to Stories (view on slashdot.org)

Samsung Demos PCIe NVMe SSD At 5.6 GB Per Second, 1 Million IOPS (hothardware.com)

Posted by samzenpus on from the greased-lightning dept.

MojoKid writes: Samsung decided to show off their latest SSD wares at Dell World 2015 with two storage products that are sure to impress data center folks. Up and running on display, Samsung showcased their PM1725 drive, which is a half-height, half-length (HHHL) NVMe SSD that will be one of the fastest on the market when it ships later this year. It sports transfer speeds of 5500MB/sec for sequential reads and 1800MB/s for writes. Samsung had the drive running in a server with Iometer fired up and pushing in excess of 5.6GB/sec. The PM1725 also is rated for random reads up to 1,000,000 IOPS and random writes of 120,000 IOPS. The top of the line 6.4TB SSD is rated to handle 32TB of writes per day with a 5-year warranty.

88 comments

  1. Pricing? by smallfries · · Score: 2

    How many gigadollars?

    --
    Slashdot: where don knuth is an idiot because he cant grasp the awesome power of php
    1. Re:Pricing? by Anonymous Coward · · Score: 0

      Less than 1, for sure.

    2. Re:Pricing? by sexconker · · Score: 1

      Your name better be "Dell" or "Apple" or "HP" if you want to buy one of these things.
      Samsung has been parading around their OEM-only SSDs for about 3 years, and rarely are consumers able to get their hands on them.

  2. inb4 by Anonymous Coward · · Score: 0

    Eagerly anticipating a poster saying it costs too much money and that hard drives are still better bang for your buck.

    1. Re:inb4 by jon3k · · Score: 1

      I can hardly wait for the "SSDs are unreliable!" guys to show up. They always give me a giggle.

  3. Re:Impressive by botfap · · Score: 0

    Clearly you thought it was Apple when infact its Samsung who not only design develop and make the best consumer SSD drives outside of Intel, they did the same for hard disks before that.

  4. Rating vs. Warranty by Anonymous Coward · · Score: 3, Interesting

    Samsung PRO line offers 5 years or total bytes written, whichever comes first, as a part of their warranty package:
    http://www.samsung.com/global/...

    While this drive "is rated to handle 32TB of writes, every day for five years without failure" - I want to see a warranty to go with that. That's ~58400TB total, about 200 times higher than their best warranty offers right now at 300TBW.

    1. Re:Rating vs. Warranty by swb · · Score: 5, Interesting

      Given the endurance test people have peformed against "consumer" SSDs, it sure seems like the expected endurance exceeds the warrant by a lot.

      This guy:

      http://blog.innovaengineering.... ...has 7 PB written to an 850 Pro and it's still going (last blog update was more than a month ago).

      I'd be awful curious to see what the actual durability of an 850 Pro would be in a real production SAN. My suspicion is that the better-than-rated endurance coupled with the low replacement cost might make it worthwhile when you consider the staggering performance you would get.

      There might even be gimmicks you could apply on a per-disk basis to improve durability, such as underprovisioning each drive by 25% so that you could wear level across more capacity.

    2. Re:Rating vs. Warranty by TheRaven64 · · Score: 3, Informative

      You don't need an explicit warranty anywhere with sensible consumer protection laws. The Sale of Goods Act in the UK (and equivalents in most EU countries) allow you to return the goods for a full refund if they do not meet the promises made at time of sale. I had a battery fail in an Apple laptop after four and a half years, but within the number of charge cycles that their ads claimed. They replaced it (couriered out a replacement that arrived at 9am the day after I called them at 3pm - better service than I've ever had from them for anything under warranty) as soon as I mentioned the Sale of Goods Act.

      --
      I am TheRaven on Soylent News
    3. Re:Rating vs. Warranty by CastrTroy · · Score: 2

      The trick is that it's a 6.4TB SSD. So that's really only equivent to writing to each sector 9125 times. Seems reasonable to expect that each sector could handle 10000 writes easily. You won't see that kind of load advertised on a 256 GB drive, as it would require the disk to endure over 220,000 writes per sector.

      --

      Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
    4. Re:Rating vs. Warranty by JBMcB · · Score: 1

      The Sale of Goods Act in the UK (and equivalents in most EU countries) allow you to return the goods for a full refund if they do not meet the promises made at time of sale.

      So what's the difference between a 1 year replacement warranty and a promise that a product will last for a year under a Sale of Goods Act?

      --
      My Other Computer Is A Data General Nova III.
    5. Re:Rating vs. Warranty by TheRaven64 · · Score: 1

      The main difference is that the warranty is usually offered by the manufacturer, whereas the Sale of Goods Act governs interaction between people who buy and sell things. It's often easier to get a refund / repair directly from the manufacturer, but this requires a manufacturer warranty. If you don't have one, then you can return it (under the SoGA) to the retailer, they can return it (also under the SoGA) to the wholesaler, who has to return it to the manufacturer. This can take a lot longer. A lot of store-provided extended warranties are a complete scam, because they will charge you for things that the law requires them to provide anyway.

      --
      I am TheRaven on Soylent News
    6. Re:Rating vs. Warranty by Anonymous Coward · · Score: 0

      So what's the difference between a 1 year replacement warranty and a promise that a product will last for a year under a Sale of Goods Act?

      Firstly, the Sale of Good act has been replaced by the Consumer Rights Act 2015.
      Secondly a promise that a product will last for a year "under the sale of goods act" is meaningless. The whole point is that the Sale of Goods act effectively requires that products last for a reasonable time *regardless of any indication to the contrary even if 'agreed' by the consumer*. The consumer cannot give/sign/whatever this right away, and the supplier cannot limit their rights to a year or any other specific term. What constitutes a reasonable time depends on the nature, cost and usage of the product, but a hard drive in normal consumer use would probably be expected to last considerably longer. After six months there is an onus on the consumer to show that the item was faulty when supplied - but this could be as simple as: Provide some evidence that a typical consumer hard drive lasts four years whereas your drive failed at 18months, so therefore it must have a defect; providing evidence of other such drives (same batch, same model, whatever) failing prematurely would also be helpful in showing a manufacturing defect. You don't have to identify what the defect is, just the visible symptoms ('drive does not power up or whatever).
      In practice if most suppliers will replaced/repair as appropriate within a reasonable time when the relevant Act is mentioned.

    7. Re:Rating vs. Warranty by JBMcB · · Score: 1

      The main difference is that the warranty is usually offered by the manufacturer, whereas the Sale of Goods Act governs interaction between people who buy and sell things.

      I understand legally what the difference is, I don't understand how it's different for the consumer.

      --
      My Other Computer Is A Data General Nova III.
    8. Re:Rating vs. Warranty by Billly+Gates · · Score: 0

      Speak for yourself.

      Last weekend I just replaced my 2nd sansdisk ssd after constant disk corruption issues. Only 1\2 a petabyte was written. I replaced lots of components trying to find the culprit as the other drive had the same problem.

      Another slashdotter a few months ago mentioned his team downgrades all his industrial equipment with mechanical disks as the ssds always fail out in the environment. Sure benchmarks show how great and reliable. Real use dictates otherwise. I only buy Samsung pros now for reliability but I give credit where it's due.

      Ask hairyfeet or anyone working in an it shop? They do fail randomly was and for some still is a problem. So take a risk if you need the speed

      --
      http://saveie6.com/
    9. Re:Rating vs. Warranty by Archangel+Michael · · Score: 1

      Size matters. Having plenty of empty space is a huge advantage for SSD drives.

      --
      Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
    10. Re:Rating vs. Warranty by Anonymous Coward · · Score: 0

      TechReport did something similar and finished killing all their SSDs. Granted, their article included only a 840 (Pro and "non Pro"), but it is still significantly more than what they were rated for.

      I am glad more people are conducting these experiments regularly to make sure that the manufacturers do not relax their QC.

    11. Re:Rating vs. Warranty by jon3k · · Score: 1

      Most likely the PCIe SSD is using SLC NAND. The Samsung Pro line uses some form of MLC (eMLC, HET, whatever).

    12. Re:Rating vs. Warranty by jon3k · · Score: 1

      If it's measured in drive writes per day, why would the size matter? My guess is that the PCIe SSDs are using SLC NAND which has a much higher write endurance.

    13. Re:Rating vs. Warranty by BronsCon · · Score: 1

      It's different in that what the warranty oh so graciously offers to cover is often vastly different from what the product is claimed to be capable of in advertising. My understanding is that SoGA entitles you to a full refund if the item does not live up to its advertising and not just the portion the manufacturer states in the warranty. Additionally, as explained by the poster before me, the ability to take the product back to the store rather than having to deal with the manufacturer, regardless of the store's return policy.

      --
      APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
    14. Re:Rating vs. Warranty by thegarbz · · Score: 1

      Another slashdotter a few months ago mentioned his team downgrades all his industrial equipment with mechanical disks as the ssds always fail out in the environment.

      Ahh yes. I'm sure he completed a thorough RCFA on the issue and didn't just say "these drives are shit let's throw these others in there".

      The thing about random failures is that they are not related to end of life and SSDs are no different from conventional HDDs in that regard. You ask anyone about any failure and he'll have a story.

    15. Re:Rating vs. Warranty by CastrTroy · · Score: 1

      It's measured in bytes written per day. Let's assume that the disk was really small, like 1 GB. If you wanted to write 32 TB to that disk every day, you'd have to write to the same sector 32000 times every day. If the Size were 1 TB, you could spread the writes out more over individual sectors, and you would only have to write to the same spot 32 times a day. When you have a 6.4 TB disk, you can write 32 TB a day, and only need to write to the same sector 5 times a day. If you had a 32 TB drive, you could write 32 TB per day and only write to each sector once. Since each sector has a limited number of writes, spreading the writes out over a larger number of sectors will increase the life of the drive.

      --

      Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
    16. Re:Rating vs. Warranty by tlhIngan · · Score: 1

      You don't need an explicit warranty anywhere with sensible consumer protection laws. The Sale of Goods Act in the UK (and equivalents in most EU countries) allow you to return the goods for a full refund if they do not meet the promises made at time of sale. I had a battery fail in an Apple laptop after four and a half years, but within the number of charge cycles that their ads claimed. They replaced it (couriered out a replacement that arrived at 9am the day after I called them at 3pm - better service than I've ever had from them for anything under warranty) as soon as I mentioned the Sale of Goods Act.

      So basically, everywhere with such an act (Europe, Australia) is effectively buying an extended warranty by default.

      Instead of say, in North America, where you get what the manufacturer gives, then Best Buy asks if you want to buy an extended warranty for another 15-20%. So instead, goods cost 15-20% more because the extended warranty price is built into the cost of the good.

      And you guys still complain you're overpaying for stuff and being scammed. No, you're not - you're paying for taxes built into the price, you're paying for extended warranties, and probably a few other things that people in North America have a choice or option to not buy with the item.

    17. Re: Rating vs. Warranty by Anonymous Coward · · Score: 0

      Extended warranties are just profit centers for retail. They pocket 50%, give the employee who sold it a couple cents, then subcontract out the service/repair.

      Your position has nothing to do with "market forces." It's about how (large) companies should not be required to have truthful advertisements.

  5. Can the SSD stand the heat of Data Center? by Anonymous Coward · · Score: 0

    Servers inside data centers can turn very hot, very very hot, and SSDs tend to fail much faster in heated environment

    No matter how fast it is rated for, a dead SSD is just a very expensive brick

    1. Re:Can the SSD stand the heat of Data Center? by Anonymous Coward · · Score: 0

      The majority of consumer SSDs are rated to run at 0-70C.

      Granted heat is an issue, but will a company willing to buy multiple 1k-2k$ SSDs be skimping out on the high end cooling?

    2. Re:Can the SSD stand the heat of Data Center? by dreamchaser · · Score: 1

      The majority of consumer SSDs are rated to run at 0-70C.

      Granted heat is an issue, but will a company willing to buy multiple 1k-2k$ SSDs be skimping out on the high end cooling?

      Not to mention skimping out on things like RAID, High Availability configurations, etc.

    3. Re:Can the SSD stand the heat of Data Center? by Anonymous Coward · · Score: 0

      I agree with the first part of your argument, ( Servers are in data centers.) The part about servers running hot sounds suspicious, and the part about SSD's failing in a "heated environment" is unadulterated bullshit. Careful, someone might find out you have no idea what you are talking about. Maybe you should STFU.

    4. Re:Can the SSD stand the heat of Data Center? by Anonymous Coward · · Score: 1

      Actually SSDs work better at high temperatures (they are easier to write). They only really dislike being stored at higher temperatures than they were written at.

    5. Re:Can the SSD stand the heat of Data Center? by Archangel+Michael · · Score: 1

      Cooling is not a problem for real datacenters. Real data centers have dedicated cooling systems, piping cold air into the racks directly.

      --
      Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
    6. Re:Can the SSD stand the heat of Data Center? by ShanghaiBill · · Score: 1

      Cooling is not a problem for real datacenters.

      Many data centers run "hot" at about 100F. Modern servers can easily handle the heat, and HDDs actually have lower failure rates at the higher end of their temperature range. Some data centers are cooled to lower temperature, but that is usually based on superstition rather than any real benefit.

      Real data centers have dedicated cooling systems, piping cold air into the racks directly.

      What about unreal data centers? Or data centers run by Scotsmen?

    7. Re:Can the SSD stand the heat of Data Center? by sexconker · · Score: 1

      Show me a bootable RAID controller that lets me plug in 2-4 of these high speed NVMe (PCIe) SSDs for RAID 1 or 10, because I am NOT about to run software RAID on my ESXi hosts (and yes, I use local storage - it's buttloads faster and cheaper).
      Bonus points if it has battery backup.

      I had to stick with 845 DC Pro when we last bought servers because there was no reliable hardware solution for RAIDing the new PCIe based SSDs.
      The few I found were expensive, from no-name brands, and only supported (in total) half of the speed of (each of) the SSDs I was looking at.

    8. Re:Can the SSD stand the heat of Data Center? by Anonymous Coward · · Score: 0

      No true Scotsman runs a data center!

    9. Re:Can the SSD stand the heat of Data Center? by Chirs · · Score: 1

      It seems to me that as long as you have protection against power outage, it should be possible to get equal reliability from software RAID. Fundamentally a hardware RAID card is just a processor with a NVRAM or battery-backed DRAM cache, and it's limited to a single PCIe bus connection.

  6. Re:Impressive by Anonymous Coward · · Score: 1

    After the performance degradation bugs I wouldn't call them the best anymore. Cheapest yes.

  7. Re:Impressive by Anonymous Coward · · Score: 0, Flamebait

    No I'm sure I mean Samsung.

    “They never met a patent they didn’t think they might like to use, no matter who it belongs to,” says Sam Baxter, a patent lawyer who once handled a case for Samsung. “I represented [the Swedish telecommunications company] Ericsson, and they couldn’t lie if their lives depended on it, and I represented Samsung and they couldn’t tell the truth if their lives depended on it.”

  8. Re:Impressive by Anonymous Coward · · Score: 0

    Samsung and HDD's? Are you trying to sell Snake Oil.
    The stack of 1TB 3.5 in disks now in our recycling skip begs to differ with you. Out of 40 purchased 31 went dead inside 24 months.
    As a result we don't buy anything from Samsung any more.
    Sure their bits are in other items but we deal with those companies who are several orders of magnitude easier to deal with than Samesung.

  9. Re:Makes sense by ledow · · Score: 2, Informative

    With storage systems moving from the local system into the cloud, you have inherent bottlenecks on the networking.

    This might be able to do stupendous speeds but no amount of disk speed will help you get your files up and down. You might be able to process them in the cloud but actually moving them between the cloud and not is still limited by your downstream.

    As such, it's actually LOCAL users who are going to benefit more from fast storage, not cloud providers. Hell, just data protection alone is in the spotlight at the moment and has the EU and US arguing and we're on the verge of every cloud company having to have at least a European-only data centre storage (so all the advantages of cloud being a world-wide solution are nullified because you can only hold EU data within the EU).

    And, to be honest, Cloud is really just the new name for "external hosted". It's nothing fancy.

  10. TempleOS certifed by Anonymous Coward · · Score: 0

    You must provide an IO port simple block device interface. The clusterfucked PCI interface must be abolished.

  11. Interface by Kardos · · Score: 1

    PCIe is great and all, but when are we going to get one of these that fits into a DIMM socket?

    1. Re:Interface by Anonymous Coward · · Score: 0

      Kinda, Sorta already done, though not the way you probably mean:
      http://www.violin-memory.com/architecture/vimm/
      This is probably what you're looking for:
      http://www.simmtester.com/page/news/showpubnews.asp?num=180

    2. Re:Interface by Kjella · · Score: 1

      Probably never, DIMMs have the controller in the CPU and flash drives have the controller on the drive. Today the controller and the NAND chips are tightly paired, to separate the two you'd have to define a standard, get Intel/AMD to implement it CPU side and put NAND chips on a DIMM. If you don't to that, PCIe is a better protocol for talking to a controller. There's plenty downsides to that solution, mostly that you'll be stuck with whatever your CPU supports.

      --
      Live today, because you never know what tomorrow brings
    3. Re:Interface by jon3k · · Score: 1

      Why would you want it in a DIMM socket? You can already get M.2 drives that slot directly into the board and operate over PCIe.

    4. Re:Interface by Anonymous Coward · · Score: 0

      A DIMM type setup is probably the next evolution. Look at Violin Memory's VIMMs. Dell, now with the EMC acquisition, along with Samsung as its primary flash provider and whoever provided flash to EMC are in position to set a new storage interface standard.

  12. With the move to PCIe, how does this scale? by swb · · Score: 2

    More or less, most storage systems are SAS based and achieve capacity scale and IOPS with many units on a SAS bus.

    What's the scaling concept behind this? I'm not aware of a (commonly available) storage expansion system based on PCIe connectivity unless you start getting into something like VSAN or the buzzwordy hyperconverged model where compute nodes create a distributed SAN. But this usually requires a lot of nodes.

    This kind of storage seems to aim for single server gross performance, which I guess might be aimed at local caching or for DBs running on a native installed OS in most conventional senses. But if you're in a virtualized environment, this seems to run against the grain somewhat -- DBs utilizing local storage and pinned to nodes with the internal storage or if you're using it as a local cache against a more conventional SAN environment, crippling performance when you move a VM until the new nodes local cache catches up.

    I guess I'm not seeing how this is better (other than some gross numbers) than more conventional SAS bus aggregation that achieves IOPS through aggregating individual drives. A dozen conventional 1 TB SSDs will provide similar IOPS, greater aggregate storage and redundancy and with SAS-3 backplane probably even greater throughput.

    Eduncate me, please.

    1. Re: With the move to PCIe, how does this scale? by Anonymous Coward · · Score: 1

      Today: lots of spindles with a few SAS/SATA SSDs as cache.

      Tomorrow: lots of SAS/SATA SSDs with a few PCIe NVMe drives as cache.

      Think ZFS and ZIL/L2ARC.

    2. Re:With the move to PCIe, how does this scale? by goarilla · · Score: 2

      What's the scaling concept behind this? I'm not aware of a (commonly available) storage expansion system based on PCIe connectivity

      Isn't that what these things are for eventually: http://www.avagotech.com/produ....

    3. Re:With the move to PCIe, how does this scale? by Anonymous Coward · · Score: 0

      These are for work stations where you do lots of video editing or CAD/CAM files that are huge. Otherwise, it's not worth it from a DB Cache setup (depends on a single PCIE slot. Most DB Caches and such use high end Raid Cards with SAS connections (Keep in mind that a 12G SAS connection is as good as a 24G SATA Connection since it's bidirectional unlike SATA that's unidirectional).

    4. Re: With the move to PCIe, how does this scale? by swb · · Score: 2

      Tomorrow: lots of SAS/SATA SSDs with a few PCIe NVMe drives as cache.

      Still not seeing the benefits versus complexity and overhead. In a 24 drive shelf you're looking at close to a million read IOPS and sequential reads into the GBytes/second range for sequential reads *just* from SSDs on a SAS backplane.

      Maybe there's some exotic, single-host database environment that would benefit from this, but a SSD-only solution would saturate 16GBFC with multipathing. At the point of combining NVMe and SSD, you're now spending more on exotic interconnect fabrics to get the data off the host than you are on the storage.

    5. Re:With the move to PCIe, how does this scale? by BronsCon · · Score: 1

      Here's a solution for virtualized environments (e.g. the cloud):

      Think LVM + software RAID + PCIe disk + SAS. When attaching an volume from a SAS array, first allocate a matching partition on the PCIe SSD, then build a RAID 1 with the SAS volume and the local partition. It's on you to figure out how to prioritize the local disk over the SAS for IO, so all reads come from the SSD and all writes hit the SSD first for consistency, but there you go. The caveat is that, once the SAS volume is used in the software RAID, it can only be used as a member of a RAID array going forward, so taking an instance offline and spinning it back up from the same volume will require rebuilding the RAID. There are issues as well with using two different "disks" for the RAID array, but those are mitigated by prioritizing the local disk; the SAS volume can be used during the rebuild, so the instance can still boot and be used (albeit with degraded performance) while the RAID rebuilds.

      This is just my 2-minute-engineering solution to the problem. I wouldn't be surprised if a better solution already exists for this; in fact I'm 99% certain one does. I don't live in a datacenter, though, so I'm not aware of one.

      --
      APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
    6. Re: With the move to PCIe, how does this scale? by BronsCon · · Score: 1

      The million IOPS isn't for getting data off the host, it's for data crunching done on the host; often times a host only sends out a fraction of what it reads from disk. In most cases, simple gigE or 10gigE will be sufficient to get data off of such a system, but more IOPS will mean the data gets processed faster and is ready to be retrieved sooner. For datasets under a couple hundred GB, there's still nothing that beats a RAMdisk, but for larger sets this is where it's at.

      --
      APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
    7. Re:With the move to PCIe, how does this scale? by BronsCon · · Score: 1

      Keep in mind that a 12G SAS connection is as good as a 24G SATA Connection since it's bidirectional unlike SATA that's unidirectional

      You're assuming an equal flow of read/write traffic and control commands in both directions. If you've got a workload that involves a lot of writes followed by a lot of reads, rather than an interleave of reads and writes, SATA will win.

      I'm sure, for your common workload, you are correct, but please try to avoid such sweeping generalizations as they really make you look like you have no idea what you're talking about.

      --
      APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
    8. Re: With the move to PCIe, how does this scale? by swb · · Score: 1

      So where are you still seeing monolithic data processing hosts?

      As far as I can tell, everyone has moved on to virtualization for all the usual benefits associated with scale out, high availability, disaster recovery. The clients I've run across that need DB crunching IOPS have moved to tiered storage where their DBs live on SSD (like 98%) and a handful which have invested in per-host cache cards, but they're still virtualized and the cache cards were a side result of bad storage/fabric decisions they were stuck with.

      I'm sure there are places with the 8U database server because the transaction capacity so very high that no shared storage can accommodate it, but I just don't see it very often.

      NVMe only seems to fit outside of the monolithic host if you're drinking the hyperconverged kool-aid and are doing some kind of distributed SAN (like VMware vSAN), where each node supplies part of it. The downside to this, though, always seems to be that the node count ends up being pretty high to get any kind of redundancy for the storage and this ends up being quietly expensive when you start counting node licenses.

    9. Re: With the move to PCIe, how does this scale? by BronsCon · · Score: 1

      If you're counting node licenses you're doing it wrong. If you're at the point of needing performance beyond what FOSS can provide, you're at the point where hiring a team to build a solution tailored to your data will cost less in the long-term than buying a solution that was built for someone else's. You'll get better performance that way, as well.

      But if you'd rather pay perpetual licensing to third parties for the rest of your natural life (as implied at the end of your post), you're right, I guess having a large amount of local storage doesn't make sense.

      And yes, I do realize I didn't answer any of the questions implied by your comment. Just because you don't have a use case (and nether do I or I'd tell you what that use case is) does not mean one does not exist; clearly Samsung sees a use case for this or they'd not have spent the money to develop it.

      --
      APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
    10. Re: With the move to PCIe, how does this scale? by Anonymous Coward · · Score: 0

      Monolithic data processing systems are par for the course (still) when you leave the bubble of the database & internet world.

      When doing large-scale modeling & simulation (think finite-element sim, fluid dynamics, weather prediction/modeling, synthetic aperture radar processing, etc.), the datasets are well larger than what can fit into memory, don't necessarily scale well to highly clustered systems because the communication/synchronization overhead kills you, etc.

      We'll probably be ordering a bunch of these when they are available, if they work out, and using them as primary storage for data processing hosts that use 4-way GPUs for number crunching: the idea being that one of these drives can saturate the PCIe link to each of the GPUs on the input side without the latency of going off-host.

    11. Re: With the move to PCIe, how does this scale? by swb · · Score: 1

      But if you'd rather pay perpetual licensing to third parties for the rest of your natural life (as implied at the end of your post), you're right, I guess having a large amount of local storage doesn't make sense.

      It doesn't make a lot of sense, regardless of your FOSS trolling and fantasies. Power, hardware and physical space all cost real money no matter what name is on the login prompt. You live in a fantasy world if you think that managing petabyte scale storage across dozens of independent compute nodes makes any sense at all with FOSS tools off the shelf.

      Sure, your specialized team and it's tailored solution could make it work, but I'd love to see the spreadsheet that explains the cost savings of basically going into the SAN business and re-inventing the wheel when you can buy it off the shelf with solutions for problems somebody else has already solved. Unless your team's name is "Google", "Amazon" or maybe "Facebook", few would ever consider "Hey, why buy Compellent or EMC, when we could get a whole bunch of Linux boxes and build our own storage environment. That will surely save money and get us the storage we need faster."

      I don't disagree that enterprise storage is costly, which is why hyperconverged compute+storage is becoming a thing (VMware vSAN, MS Storage Spaces, etc), but the reality is that marrying performance and reliability in node-distributed storage is hard, especially if you need immediate coherency between nodes (and not Google-like "fuckit, we're ok with slightly stale pagerank data").

      And really, we're talking block storage here -- I don't really know of many IT products more basic and essentially uncustomized than this.

    12. Re: With the move to PCIe, how does this scale? by BronsCon · · Score: 1

      basically going into the SAN business and re-inventing the wheel

      Local caches are useless?

      --
      APK quotes people (including myself) without context and should not be trusted. Just thought you should know.
    13. Re: With the move to PCIe, how does this scale? by swb · · Score: 1

      They're less useful as clusters grow and workloads migrate across nodes, especially if you use automation like DRS to maintain a self-balancing cluster that levels workloads across clusters.

      I've seen them work OK in very small clusters with more or less static node membership, but they're pretty uncommon outside of specialized markets. The last time I've seen one deployed was someplace where they had a crummy, specialized database application that performed poorly and they were tacked on in desperation. The software was the problem and no hardware would fix it.

  13. Re:Makes sense by Anonymous Coward · · Score: 2, Informative

    As such, it's actually LOCAL users who are going to benefit more from fast storage, not cloud providers.

    Only if you completely ignore all processing of stored data within the cloud infrastructures.

    Uploading and download files to/from a cloud isn't really where a super-fast SSD will be used, primarily.

  14. 850 Pro Enterprise Use by Anonymous Coward · · Score: 2, Interesting

    I'm testing this theory right now. I'm using 850 Pros in RAID 10 arrays. We'll see how it goes. My math says I should be fine for 7 years, but the anxiety after only 3 months is palpable.

    The issue isn't that a $500 drive might fail. The issue is that if the drives start failing, I have to chuck and replace $23,000 worth of drives.

    1. Re:850 Pro Enterprise Use by Anonymous Coward · · Score: 2, Interesting

      i've been running on 840pros in a datacenter enviroment and i haven't had one fail on me yet in two years. recently we've changed to ordering 850pros for new servers and in the 3-4 months we've done that i haven't seen a dead one either.

    2. Re:850 Pro Enterprise Use by swb · · Score: 1

      I'd love to hear more about how you're using them.

      The performance is generally so good that I might be inclined to use a double parity redundancy scheme and hot spare auto rebuild as a hedge against failure. I would generally expect double parity and hot spare rebuild to be fast enough to protect against all but the most catastrophic failures, like all drives somehow failing faster than you can replace spares.

      Do tell how you're actually using them and what kind of write usage you se.

      I built a tiered storage spaces array and have been mildly alarmed at how great the write rate is on the SSDs I use for the SSD tier, although I think some of that is a function of a shit ton of write churn as I moved a bunch of data around temporarily on that storage space. I need to start keeping a spreadsheet to monitor the TBW over time to see what my estimated lifespan is relative to warranty. If these things are capable of even 1 PB of writes, they should outlast this system's useful lifespan.

    3. Re:850 Pro Enterprise Use by CastrTroy · · Score: 1

      I think this is one of the biggest advantages of using SSDs in the data center. With spinning platters, rebuilding the array after a drive dies can take a serious amount of disk resources. SSDs are so much faster that rebuilding a drive can be done in a fraction of the time, and not put so much strain on the system when the rebuild is being done.

      --

      Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
    4. Re:850 Pro Enterprise Use by Qzukk · · Score: 2

      but the anxiety after only 3 months is palpable.

      I assume they're under warranty for the next while.

      My fear would be that there is an intelesque hard limit on writes that bricks the drive, and ALL of your drives in a RAID array will hit that limit simultaneously.

      --
      If I have been able to see further than others, it is because I bought a pair of binoculars.
    5. Re:850 Pro Enterprise Use by sexconker · · Score: 1

      That's why we ponied up for the 845 DC Pro, just in case.

    6. Re:850 Pro Enterprise Use by swb · · Score: 1

      My fear would be that there is an intelesque hard limit on writes that bricks the drive, and ALL of your drives in a RAID array will hit that limit simultaneously.

      All of the drives in that test indicated potential failure via monitoring way before they actually failed, regardless of the actual event of their failure.

      My guess is that part of the trick to using an 850 Pro-type drive is close monitoring of drive error status and aggressive replacement of drives showing an indication of failure. It seems unlikely that simultaneous failure would happen across an entire shelf, so closely spaced, as to prevent rebuilds to available hot spares (and hot spare replacement).

      The trick is the "seems" part -- AFAIK nobody has publicly done this, created a full shelf array and seen what happens. It would be a great experiment, especially if a lot of data was collected about the workloads the array supported and per-member workloads of the individual disks.

      My guess is that replacement rates would be higher than rotational media for ordinary workloads, but less than expected. The question is whether the replacement rate would be offset by the performance benefit. I suspect it might be.

      If you could buy a 20 TB all-flash SAN for $50k versus an all-HDD SAN for the same money, would the performance boost be worth it if you had to spend $2000 per year in disk replacement? What's the annual replacement cost (and maybe risk) where it becomes not worth it?

  15. Finally! by grub · · Score: 0


    The top of the line 6.4TB SSD is rated to handle 32TB of writes per day with a 5-year warranty.

    Finally something that can handle my torrent load.

    --
    Trolling is a art,
    1. Re:Finally! by Anonymous Coward · · Score: 0

      Your mom always handles my torrential load like a pro.

    2. Re:Finally! by grub · · Score: 1


      Your mom always handles my torrential load like a pro.

      She died in 1982. If you're that desperate, I will lend you a shovel. Just let me take video.

      --
      Trolling is a art,
    3. Re:Finally! by Anonymous Coward · · Score: 0

      She died in 1982.

      Nah, she just moved after she finally realized her son will always be an asshole.

  16. Re:Impressive by JBMcB · · Score: 1

    Seagate used to sell the best spinning disk hard drives until their quality took a nose dive, then it was Western Digital. Samsung was OK, hardly the best.

    --
    My Other Computer Is A Data General Nova III.
  17. Re:Makes sense by JoeyRox · · Score: 1

    Think about the problem a little more deeply. Cloud servers have tends of thousands of concurrent users and hundreds of thousands of concurrent transactions. They definitely stand to benefit from faster storage (particularly IOPs).

  18. Re:Makes sense by Archangel+Michael · · Score: 1

    SSDs do especially well for Database applications, considering they tend to be IOP bound. Give me a Million IOPs on certain databases, and they will scream.

    --
    Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
  19. Re:Impressive by fnj · · Score: 2

    In the days of the HD204UI 2TB, I want to say maybe 10 years ago, Samsung owned the disk drive reliability world. While they were sold, I wouldn't get anything else. Seagate and WD were crap even then. All of my 20+ HD204UIs still work flawlessly after 5+ years of largely 24x7 operation.

  20. Re:Makes sense by zlives · · Score: 1

    pigeon delivery service shill?

  21. Re:All SSDs must be loaded with this haiku by sexconker · · Score: 2

    Haikus are for cows
    Moo. Moooooo cows! Moo your haikus.
    You haikuing cows.

  22. Re: Makes sense by Anonymous Coward · · Score: 0

    Troll or idiot? Your data is already all over the cloud.

  23. Re: All SSDs must be loaded with this haiku by Anonymous Coward · · Score: 0

    It's you,moo conker moo !

  24. Re: Makes sense by Anonymous Coward · · Score: 0

    I'm saturating 2* 10Xbe links from my data generators to the storage.

    Is it possible to get a 40gbe uplink to Amazon ? Any cloud service ?

  25. Re:Makes sense by Anonymous Coward · · Score: 0

    thanks Captain obvious.

  26. warranty vs advertising by Chirs · · Score: 1

    It's not like buying an extended warranty, it's holding manufacturers/retailers accountable for promises that they make.

  27. Re:Makes sense by im_thatoneguy · · Score: 1

    And, to be honest, Cloud is really just the new name for "external hosted". It's nothing fancy.

    Externally hosted generally used to be co-located. You would either lease rack space or you would sign a lease on equipment and rack space. I would say that "Cloud" means software-defined networking combined with as low as per-minute usage contracts. I don't remember any colo facilities offering me the option of configuring the network remotely. Nor was it easy and quick to just literally move a slider and say "MOAR POWER!" without a lot of hassle.

  28. Re: Makes sense by im_thatoneguy · · Score: 1

    Is it possible to get a 40gbe uplink to Amazon ? Any cloud service ?

    Azure has ExpressRoute https://azure.microsoft.com/en... which looks like it would cost $20,000/month for 40gb/s. They have several ways to connect your network to their internal network.

    Amazon similary has DirectConnect which lets you plug straight into a 10gbe port in select buildings.

    Looks like 40gb/s of connectivity would cost you around $6,500 a month.
    https://aws.amazon.com/directc...