In our series of Hard Disk Drive guides, here’s the latest update to our list of recommended HDDs. All numbers in the text are updated to reflect pricing at the time of writing.

Best Consumer Hard Drives: Holiday 2018

Data storage requirements have kept increasing over the last several years. While SSDs have taken over the role of the primary drive in most computing systems, hard drives continue to be the storage media of choice in areas dealing with large amount of relatively cold data. Hard drives are also suitable for workloads that are largely sequential and not performance sensitive. SSDs are yet to achieve the low $/GB metric that makes HDDs attractive in that market segment.

From a gaming perspective, install sizes of 100s of GBs are not uncommon for modern games. Despite the falling flash prices, high-capacity SSDs still tend to carry a price premium, making hard drives attractive in this market segment. This guide will help readers choose the appropriate hard drive based on their workload, while also keeping the price factor in mind.

November 2018 HDD Recommendations
Segment Recommendations
High-Capacity Desktop 12TB Seagate BarraCuda Pro $430
Mid-Capacity Desktop 6TB Seagate BarraCuda Pro $220
High-Capacity NAS 12TB Seagate Exos $375
Mid-Capacity NAS 6TB WD Red $190

There are three active vendors in the consumer hard drive space - Seagate, Toshiba, and Western Digital. While Seagate offers hard drives targeting desktop workloads at their leading capacity points, Western Digital and Toshiba reserve the leading edge for enterprise and datacenter drives. The previous version of our purchase guide from January 2018 only dealt with Seagate and WD drives. Since then, we have seen a number of announcements in the consumer space from both Seagate and Toshiba. These have given us a few additional HDD models to consider in our Holiday 2018 guide. The Seagate Barracuda Pro 14TB is the only newly introduced drive for which we have a hands-on review. Instead of recycling the graphs and benchmark data points from our reviews, we will be adopting a slightly different format starting with this version.

Consumers looking to purchase hard-drives need to have a rough idea of the use-cases they are going to subject the drives to. Based on that, a specific set of metrics needs to be considered. We first take a look at the different metrics that matter, and how various hard drives stack up against each other. Since many hard drive families from different vendors can satisfy the requirements, it may all come down to the pricing. We will present a pricing matrix for various hard drive families against the available capacities.

The following hard drive families will be considered in our guide today. Note that this list does not include surveillance-focused drives such as the WD Purple or Seagate SkyHawk. We will also consider some of the enterprise / datacenter SATA drives that are available for purchase from e-tailers.

  1. Seagate BarraCuda Pro
  2. Seagate IronWolf NAS
  3. Seagate IronWolf Pro NAS
  4. Seagate Exos Enterprise
  5. Toshiba N300
  6. Toshiba X300
  7. Western Digital Gold
  8. Western Digital Red
  9. Western Digital Red Pro

A few notes are in order - the WD Gold line of enterprise drives seems to be EOL-ed - and replaced by the HGST Ultrastar DC lineup that is not yet widely available in the North American retail market. Toshiba's MN07 and MG07 series are the 9-disk models featuring industry-leading 14TB PMR capacity models. However, they are again enterprise-focused, and the retail market has to make do with the N300 and X300 drives for NAS and desktop systems.

Metrics that Matter

One of the easiest ways to narrow down the search for a suitable hard drive is to look at the target market of each family. The table below lists the suggested target market for each hard drive family we are considering today.

Hard Drive Families - Target Markets
Drive Family Target Markets
Seagate BarraCuda Pro Desktops and All-in-Ones
Home Servers
Creative Professionals Workstations
Entry-Level Direct-Attached-Storage (DAS) Units
Seagate IronWolf NAS NAS Units up to 8 bays
(Home, SOHO, and Small Business)
Seagate IronWolf Pro NAS NAS Units up to 24 bays
(Creative Pros, SOHO, and Small to Medium Enterprises)
Seagate Exos Enterprise Datacenter and Bulk Cloud Storage
Toshiba N300 NAS Units up to 8 bays
Toshiba X300 Professional Desktops, Home Media or Gaming PCs
WD Gold Datacenter and Bulk Cloud Storage
WD Red NAS Units up to 8 bays
WD Red Pro NAS Units up to 24 bays

After filtering out models that don't apply to one's use-case (as an example, for usage in a 4-bay NAS enclosure, one could rule out the Tosiba X300 straight away), it is now time to look at some of the metrics that bring out a drive's characteristics.

Hard Drive Families - Metrics of Interest
Drive Family Rated Workload (TB/yr) Rated Load / Unload Cycles Unrecoverable Read Errors MTBF (Hours) Warranty (Years)
Seagate BarraCuda Pro 300 300K 1 in 10E15 1M 5
Seagate IronWolf NAS 180 600K 1 in 10E15 1M 3
Seagate IronWolf Pro NAS 300 600K 1 in 10E15 1.2M 5
Seagate Exos Enterprise 550 600K 1 in 10E15 2.5M 5
Toshiba N300 180 300K 1 in 10E14 1M 3
Toshiba X300 N/A (72?) 300K 1 in 10E14 0.6M 2
WD Gold 550 600K 1 in 10E15 2.5M 5
WD Red 180 600K 1 in 10E14 1M 3
WD Red Pro 300 600K 1 in 10E14 1M 5

Based on these metrics, it is clear that the enterprise drives (Seagate Exos Enterprise and WD Gold) are going to be more reliable in the long run over a big sample set. However, the 180 - 300 TB/yr workload rating is acceptable for most consumer workloads when the drives are going to be used as part of RAID arrays.

The BarraCuda Pro strikes a nice balance across many metrics, but it is rated only for 300K load / unload cycles. It also doesn't have the RV sensors present in the rest of the drives (other than the Toshiba X300).

In considering the non-enterprise drives, we note that the 'Unrecoverable Read Errors' metric is 10x worse for the WD and Toshiba drives compared to the Seagate ones. The MTTF metric for the IronWolf Pro is slightly better than the other drives (at 1.2M vs. 1M hours).

One of the aspects not mentioned in the above table is that the WD Red drives are in the 5400 RPM class, while the rest are all 7200 RPM. From a raw performance perspective at equivalent capacity points, the WD Red might not win on benchmarks, but, it is most likely to be the most power efficient of the lot.

Pricing Matrix and Concluding Remarks

The matrix below shows the current pricing for each available capacity point in all the considered hard drive families.

HDD Pricing Matrix (as of November 6, 2018)
Drive Family 14TB 12TB 10TB 8TB 6TB 4TB
Seagate BarraCuda Pro $586 $430 $349 $270 $220 $162
Seagate IronWolf NAS $560 $400 $300 $265 $195 $119
Seagate IronWolf Pro NAS $600 $450 $352 $282 $210 $160
Seagate Exos Enterprise ($615) $375 $315 $246 $188 $160
Toshiba N300 - - $363 $239 $175 $120
Toshiba X300 - - $330 $230 $165 $130
WD Gold - $467 $365 $308 $230 $172
WD Red - - $316 $250 $190 $120
WD Red Pro - - $389 $325 $232 $185

The desktop storage market is a straight shoot-out between the Seagate BarraCuda Pro and the Toshiba X300. The latter does not compete at the 14TB and 12TB capacity points, but, consistently beats the Seagate's pricing at the 10TB and lower capacities. Unfortunately, based on our analysis so far, we think it is fair to give the Toshiba X300 a pass. For the $19 - $55 extra, we get a much higher workload rating, better reliability, and three extra years of warranty. In addition, the Seagate BarraCuda Pro also comes with a bundled data recovery service. So, this is a case where the benefits outweigh the cost, and our recommendation goes to the costlier of the two drives - the Seagate BarraCuda Pro.

Prior to commenting on the other possible use-cases, one thing is clear from the above pricing matrix - if you absolutely require 12TB or 14TB per disk, the Seagate drives are your only choice for purchase in the retail market. The 12TB WD Gold is available on Amazon, but priced much higher than even the equivalent Seagate Exos x12 drive. At those two capacity points, Seagate drives are pretty much the only option, and it is just a question of figuring out which of the drives makes the best sense for the application.

On the SOHO / SMB NAS front, the Seagate Exos series, despite its enterprise background, continues to make a strong case across multiple capacity points. The only places where the WD Red could edge out as a better choice are scenarios where the power consumption needs to be kept low. Fortunately, there is not much difference in the pricing between the WD Red and the Seagate Exos. The IronWolf NAS models deliver slightly better performance compared to the WD Red due to the 7200RPM nature, but, have correspondingly higher power consumption numbers. On the SMB / SME NAS front, the IronWolf Pro pretty much edges out the WD Red Pro in pricing across the board. This is despite the bundled data recovery service in the IronWolf Pro pricing.

Based on the above analysis, the recommendations for the NAS drives are clear - WD Red when performance is not as important as overall power consumption, and the Seagate Exos Enterprise drives otherwise. This is assuming that the user has adopted the 3-2-1 backup rule and doesn't foresee the need for a data recovery service (DRS). The IronWolf Pro NAS and the BarraCuda Pro both bundle the DRS. This needs to be taken into account while considering the pricing difference against other drives in the same capacity class.

POST A COMMENT

51 Comments

View All Comments

  • PeachNCream - Thursday, November 15, 2018 - link

    I use a 1TB Seagate laptop drive stuffed into a cheap USB external case for basically that purpose. Given that its a backup, I'm not worried much about it dying on me since it gets only occasional use and even if it does die, it'd have to die at the same time as I suffer from a failure of my live data storage in my PC which is rather unlikely short of some kind of natural disaster that wipes out both which could only be addressed by using cloud backups. Go cheap and don't sweat the details. Reply
  • eastcoast_pete - Thursday, November 15, 2018 - link

    @Ganesh: Thanks for an article on this under-covered subject. Spinning rust is still my storage of choice for large files (media) that I like to hang on to for a while, and much cheaper than SSDs in GB/$. However, I do have a few questions: You (and some manufacturers) use MTTF, some others MTBF for HDs, same thing IMO). I'd suggest a short, 1-2 sentences overview over how these times are arrived at, and whether all manufacturers use the same methodology. I was/am skeptical about MBTF and MTTF data, as they represent (number of disks x hours to first failure of any disk) or similar, which doesn't tell me anything about the fate of the other drives. That being said, a 5-fold difference is potentially meaningful.
    Second, some comments on the track record of these disks and manufacturers would be much appreciated. Backblaze used to publish some really useful actual use survival data, but I'm not sure if they kept that great tradition up. Maybe they did, maybe some other large-volume users make data available. I always trusted those real-life data a lot more than MBTF or MTTF numbers.
    Third,
    Reply
  • eastcoast_pete - Thursday, November 15, 2018 - link

    continued:..Third, while a longer warranty is nice, the pain of having to restore TB of data from backups often outweighs the monetary benefit of a longer warranty. Lastly, a comment on how easy or difficult it is to have the warranty honored from each respective manufacturer would be appreciated here, too. Reply
  • Hul8 - Thursday, November 15, 2018 - link

    I believe the MTTF values are calculated statistically from a large pool of tested drives. The metric tells the time it takes for 50% of the files in a large pool to have failed.

    Likely being a binomial distribution, the failures will concentrate around the time stated in MTTF, with only outliers failing long before or long after.
    Reply
  • Hul8 - Thursday, November 15, 2018 - link

    *50% of the disks.

    [I really hope AnandTech will get around to allowing editing at some point. At least allow it for 10 minutes after posting to fix typos and wrong words.]
    Reply
  • eastcoast_pete - Thursday, November 15, 2018 - link

    On that I couldn't agree more! Reply
  • eastcoast_pete - Thursday, November 15, 2018 - link

    Actually, the MTTF is a projection based on assumptions and initial failure data. A key variable is what kind of distribution is assumed (normal [Gaussian], Weibull, Weibull-Bayer, some distribution they made up...), and (missing from the table) what the confidence interval (upper and lower borders) is for each MTTF. The 50% of all drives having non-recoverable failures is not possible, as 1 million hours MTTF equals over 114 years. That is why I rather go by the data of actual users like Backblaze that have thousands of drives running under identical conditions and just report total and annualized failure rates by model. See https://www.backblaze.com/blog/hard-drive-stats-fo...
    Lastly, I really wish Google, facebook, AWS and Microsoft would publish their hard drive stats; they buy HDs by the hundreds of thousands a year.
    Reply
  • JoeyJoJo123 - Thursday, November 15, 2018 - link

    Backblaze does still report on hard drive fail rates, however, they're pretty conservative about which ones they start investing capital into, so they don't have as much variety and selection that their reports from years past had. These days they report on less than a dozen specific models, all of which are fairly large disk sizes, and these numbers really only generally show which brands tend to be most reliable, with HGST drives still being very good overall, although as with all things individual products should be taken into consideration and while the HGST drives they use tend to be very reliable, the same might not hold true for lower storage capacity drives. Reply
  • melgross - Thursday, November 15, 2018 - link

    What I find interesting is that both Facebook and Google, both of whom run millions of drives of all kinds, find, year after year, that consumer drives seem to do just as well as enterprise drives in reliability. And I’ve seen a number of studies of very large numbers of drives that show that it doesn’t seem to matter which drives you buy relating to performance either. That is, whatever the drives are aimed at, the performance doesn’t seem to reflect those aims, as long as the physical numbers are the same, mostly reflecting drive speed. Drives for raid, NAS and desktop, etc. all seem to perform about the same. A small difference here and there doesn’t really matter. Reply
  • Hul8 - Thursday, November 15, 2018 - link

    The choice between desktop and NAS doesn't so much depend on reliability but firmware optimizations.

    If you run a desktop drive in RAID and the drive encounters a read error, it will try and try again to read for many minutes. Getting no response from the drive, the RAID controller will assume the drive dead (because it expects properly behaving NAS drives that return errors quickly), and drops the drive from the array. This means the user has to manyally add the drive back and resilver the array.

    Facebook and Google don't use RAID, but probably some sort of objective storage, so the firmware doesn't matter to them.
    Reply

Log in

Don't have an account? Sign up now