Skip to main content

Seagate Displays 60 TB SSD At Flash Memory Summit

Seagate is refreshing its entire line of enterprise SSDs, and it continues the trend by kicking density up a notch with the demonstration of a prototype 60 TB SSD at the Flash Memory Summit. Seagate indicated that the SSD is a technology demo, but the company based the overall design of the new SSD on proven and mature technology. A more refined version of the SSD will eventually come to market.

The 60 TB SSD comes in the 3.5" form factor, which is much larger than the standard 2.5" designs that we are accustomed to. Early enterprise SSDs came in the 2.5" form factor but eventually transitioned to 2.5" as flash continued to shrink, thus allowing SSD vendors to cram more capacity into smaller SSDs. The shrinking trend continues today as m.2 SSDs become more plentiful. Seagate is betting that the incredible density of the 3.5" SSD, which can enable up to 1PB of storage with only 17 SSDs, will be a hit for system builders.

Samsung debuted its 16TB PM1633 last year in a 2.5" form factor, but the increased real estate of the 3.5" form factor and several architectural advancements allows Seagate to offer more density than two PM1633a units combined. The Seagate SSD uses 3D TLC NAND to reach the heights of storage capacity, but the company tied the SSD to a 12 Gbps SAS interface. Many of the fastest SSDs on the market employ the PCIe connection in tandem with the NVMe interface, but Seagate geared the new SSD for read-intensive workloads, large storage arrays and active archive use cases, which means the SSD focuses on cheap capacity instead of high performance. Twelve Gbps SAS also provides a dual-port capability, which allows two servers to access the SSD to provide HA (High Availability) functionality. Some SAS SSDs also employ a wideport scheme to access both ports in tandem to provide more performance.

Seagate 60 TB 12 Gbps SSD
Capacity (TB)60
Form FactorSFF 3.5"
Interface / ArchitectureDual 12 Gbps SAS
Sequential Read (MB/s) 128KB1,500
Sequential Write (MB/s) 128KB1,000
Random Read (IOPS) Sustained 4KB QD32150,000
Power: Active Max Average (W)15

The SSD uses the same design as the Seagate 1200.2 SSDs we recently reviewed. A co-branded LSI/Seagate TT50761 controller works in concert with an array of eASIC ONFI bridges, which allows the SSD to employ scalable channel interleaving. This technique allows the SSD to address up to 1,280 flash die per controller. The ONFI bridge arrangement allows Seagate to carve out varying capacities by simply adjusting the number of bridges and NAND packages.

The performance specifications are somewhat conservative at this point, which is likely due to tying such a large amount of flash to a single SSD controller. Seagate noted that it does not plan to ship the SSD this year, at least not in its current form, but that a variation of this will likely come to market in 2017. Samsung's 16TB SSD recently appeared at retail outlets for more than $10,000, and the quadrupled capacity of the Seagate SSD implies that it will carry a much heftier price point.

  • littleleo
    I think you have a misprint:

    "Early enterprise SSDs came in the 2.5" form factor but eventually transitioned to 2.5" as flash continued to shrink"

    I think you mean "came in 3.5" form factor".
    Reply
  • thundervore
    The misprint is there to distract us from seeing that this is made by Seagate.
    Reply
  • hst101rox
    No pics of the drive? I want to see the heatsink since power consumption is 15 watts max
    Reply
  • drtweak
    I remember commenting on the 16TB SSD on facebook and why they don't just go to 3.5 to give them much more space to fill. Some people told me that is like going backwards. The thing is a LOT of servers out there are designed for 3.5 and not 2.5 and usually to switch requires a whole new chassis or even a whole new server ( I know I sell Dell workstations and servers as part of my IT job) and having a 3.5 SSD would give higher yield and be more compatible with current servers out there. Yes it is Seagate and not samsung but still the fact that they made a 60TB SSD in a 3.5 form factor could be a game change for the mass storage area. Yea pretty sure it will have like a 50K Price tag on it though lol
    Reply
  • zodiacfml
    Brute forcing density. WD can beat team in this regard.
    There is a very small niche for this type of storage density but they will surely get buyers willing to pay for this. I wonder if Samsung will compete or block Seagate to get this very small niche.
    Reply
  • jimmysmitty
    18410327 said:
    The misprint is there to distract us from seeing that this is made by Seagate.

    And that means what? It is using Microns 3D NAND and Micron/Intel has some of the best NAND in the market. Seagate is just putting it together.

    Either way I don't see that it is a bad thing as the more SSD players in the market the better and cheaper it will get for us.

    18411280 said:
    Brute forcing density. WD can beat team in this regard.
    There is a very small niche for this type of storage density but they will surely get buyers willing to pay for this. I wonder if Samsung will compete or block Seagate to get this very small niche.

    This is not brute forcing density this is a logical progression. 3.5" is a viable market and has much more room to put NAND chips than 2.5". It is the same as to why we can get a 10TB desktop drive but 4TB is the maximum in the laptop space and that is with a over sized 15mm drive while most laptops only suppot 9mm drives or 7mm drives.

    And how is storage density a niche market? A single good sized company has multiple backups spanning 7 or more years and can easily use TBs worth of storage space. Hell I think we use probably 100TB or so with all the servers and backups of emails and user data/archives.
    Reply
  • 3ogdy
    18410327 said:
    The misprint is there to distract us from seeing that this is made by Seagate.

    *Instantly throws up*
    Can...barely...speak...through...vomit. Yuck!
    Reply
  • problematiq
    18411221 said:
    I remember commenting on the 16TB SSD on facebook and why they don't just go to 3.5 to give them much more space to fill. Some people told me that is like going backwards. The thing is a LOT of servers out there are designed for 3.5 and not 2.5 and usually to switch requires a whole new chassis or even a whole new server ( I know I sell Dell workstations and servers as part of my IT job) and having a 3.5 SSD would give higher yield and be more compatible with current servers out there. Yes it is Seagate and not samsung but still the fact that they made a 60TB SSD in a 3.5 form factor could be a game change for the mass storage area. Yea pretty sure it will have like a 50K Price tag on it though lol

    When dealing with most small servers you are using 3.5" drives. most of your mid/high end servers and mid/high end sans use 2.5" due to the slight decrease in size of drives it opens up new configurations. eg. A 2U server with 3.5" bays could hold 8 3.5" drives with a max of 4TB per drive (We are talking SAS) so a max of 32TB of space. in the same 2U you can get 24 2.5" drives with a capacity of 1.8TB per drive (Again SAS because in a server environment that's not just hosting a domain actually doing work requires throughput) giving you 43.2TB of data. may not seem like much but when you are dealing with a corporate environment or a data center that's a HUGE difference.
    Reply
  • drtweak
    18411795 said:
    18411221 said:
    I remember commenting on the 16TB SSD on facebook and why they don't just go to 3.5 to give them much more space to fill. Some people told me that is like going backwards. The thing is a LOT of servers out there are designed for 3.5 and not 2.5 and usually to switch requires a whole new chassis or even a whole new server ( I know I sell Dell workstations and servers as part of my IT job) and having a 3.5 SSD would give higher yield and be more compatible with current servers out there. Yes it is Seagate and not samsung but still the fact that they made a 60TB SSD in a 3.5 form factor could be a game change for the mass storage area. Yea pretty sure it will have like a 50K Price tag on it though lol

    When dealing with most small servers you are using 3.5" drives. most of your mid/high end servers and mid/high end sans use 2.5" due to the slight decrease in size of drives it opens up new configurations. eg. A 2U server with 3.5" bays could hold 8 3.5" drives with a max of 4TB per drive (We are talking SAS) so a max of 32TB of space. in the same 2U you can get 24 2.5" drives with a capacity of 1.8TB per drive (Again SAS because in a server environment that's not just hosting a domain actually doing work requires throughput) giving you 43.2TB of data. may not seem like much but when you are dealing with a corporate environment or a data center that's a HUGE difference.


    You do have a point but for people who get new servers sure. But those who have existing 3.5 inch bays? Most of our severs are told to a client who does Security. So they will buy a high end Power Edge R730 which can do 8 3.5 or 16 2.5 (Not sure where you got 24 from. Maybe if they had nothing in the front of the server period) but the thing is SAS drives are up to 8TB today so it is more like 64 vs 43 TB. My client buy a R730 usually with 8 8TB and then sometimes gets a Dell Power vault to more than double that.

    Now the thing is depending on which brand you go with, the 8TB can cost only twice as much as a 1.8 2.5. If a 8TB cost more than 4 times the 2.5 then yes i could then understand going with 2.5. Also having more drives could mean better through put on some things depending on what kind of RAID you do.

    It just THERE is room for a 3.5 inch SSD, and I think if someone just did was seagate just did then it might catch on. I mean to have 240TB of raw space in a 8 bay servers? Comon. Standard 2.5 7-9mm can't even come close.

    But that is just my 2 cents
    Reply
  • jimmysmitty
    18411795 said:
    18411221 said:
    I remember commenting on the 16TB SSD on facebook and why they don't just go to 3.5 to give them much more space to fill. Some people told me that is like going backwards. The thing is a LOT of servers out there are designed for 3.5 and not 2.5 and usually to switch requires a whole new chassis or even a whole new server ( I know I sell Dell workstations and servers as part of my IT job) and having a 3.5 SSD would give higher yield and be more compatible with current servers out there. Yes it is Seagate and not samsung but still the fact that they made a 60TB SSD in a 3.5 form factor could be a game change for the mass storage area. Yea pretty sure it will have like a 50K Price tag on it though lol

    When dealing with most small servers you are using 3.5" drives. most of your mid/high end servers and mid/high end sans use 2.5" due to the slight decrease in size of drives it opens up new configurations. eg. A 2U server with 3.5" bays could hold 8 3.5" drives with a max of 4TB per drive (We are talking SAS) so a max of 32TB of space. in the same 2U you can get 24 2.5" drives with a capacity of 1.8TB per drive (Again SAS because in a server environment that's not just hosting a domain actually doing work requires throughput) giving you 43.2TB of data. may not seem like much but when you are dealing with a corporate environment or a data center that's a HUGE difference.

    18411878 said:
    18411795 said:
    18411221 said:
    I remember commenting on the 16TB SSD on facebook and why they don't just go to 3.5 to give them much more space to fill. Some people told me that is like going backwards. The thing is a LOT of servers out there are designed for 3.5 and not 2.5 and usually to switch requires a whole new chassis or even a whole new server ( I know I sell Dell workstations and servers as part of my IT job) and having a 3.5 SSD would give higher yield and be more compatible with current servers out there. Yes it is Seagate and not samsung but still the fact that they made a 60TB SSD in a 3.5 form factor could be a game change for the mass storage area. Yea pretty sure it will have like a 50K Price tag on it though lol

    When dealing with most small servers you are using 3.5" drives. most of your mid/high end servers and mid/high end sans use 2.5" due to the slight decrease in size of drives it opens up new configurations. eg. A 2U server with 3.5" bays could hold 8 3.5" drives with a max of 4TB per drive (We are talking SAS) so a max of 32TB of space. in the same 2U you can get 24 2.5" drives with a capacity of 1.8TB per drive (Again SAS because in a server environment that's not just hosting a domain actually doing work requires throughput) giving you 43.2TB of data. may not seem like much but when you are dealing with a corporate environment or a data center that's a HUGE difference.


    You do have a point but for people who get new servers sure. But those who have existing 3.5 inch bays? Most of our severs are told to a client who does Security. So they will buy a high end Power Edge R730 which can do 8 3.5 or 16 2.5 (Not sure where you got 24 from. Maybe if they had nothing in the front of the server period) but the thing is SAS drives are up to 8TB today so it is more like 64 vs 43 TB. My client buy a R730 usually with 8 8TB and then sometimes gets a Dell Power vault to more than double that.

    Now the thing is depending on which brand you go with, the 8TB can cost only twice as much as a 1.8 2.5. If a 8TB cost more than 4 times the 2.5 then yes i could then understand going with 2.5. Also having more drives could mean better through put on some things depending on what kind of RAID you do.

    It just THERE is room for a 3.5 inch SSD, and I think if someone just did was seagate just did then it might catch on. I mean to have 240TB of raw space in a 8 bay servers? Comon. Standard 2.5 7-9mm can't even come close.

    But that is just my 2 cents

    http://www.serversupply.com/products/part_search/pid_lookup.asp?pid=149283&gclid=CJbkv96itc4CFY-AaQoduBULWA

    I think he was talking about a SAN more than a server. We have a SAN that is 3U with 16 3.5" bays in it.

    I agree that 2.5" cannot come close to this. A 16 bay SAN with 60TBs each is INSANE. 960TBs in a single SAN. I think our SAN isn't even close to that. Even with the 10TB HDDs that are available that would be only 160TB.
    Reply