back to article OpenIO pulls up ARM controller SOCs: Kinetic's Marvellous... can anybody do it?

OpenIO, the French object-storage startup, is progressing its work with Kinetic drives. The idea is that object storage stores data on server nodes, and those nodes could be individual Kinetic disk drives. Such drives have a micro-server, so to speak, running on them, providing direct Ethernet access and an object-style Get: …

  1. Pascal

    Looking for a problem to solve...

    "For us, Kinetic storage still has elements of clever engineering technology looking for an end-user problem to solve."

    That's what I was thinking immediately / wanted to comment until it turned out to be the last paragraph of the article. Seems more like a toy for proof of concepts, the management issues you'd get at scales with that would be terrible, and you'd need so much effort to manually handle redundant storage in case of failures, and so on.

    Maybe as a cheap, proof of concept object storage system for dev work...

    I mean, it seems clever, but it also seems to have no real large scale practical application.

    1. banalyzer

      Re: Looking for a problem to solve...

      This looks like they are aiming at SME or large home installations. Pointless though it might be in the majority of cases in some it will fit, in others it's the bragging rights.

      My household is up to nearly 6TB but I doubt that's enough justify an object store.

      1. theOtherJT Silver badge

        Re: Looking for a problem to solve...

        It might work for us here - Instead of buying expensive servers full of expensive SAS RAID controllers which you have to keep populating with expensive SAS disks, and even more expensive flash caches, you could buy a cheap ethernet enclosure - basically a glorified switch with integrated drive caddies - and run something like CEPH where you just keep adding more disks as required.

        The magic only works if the kinetic enclosure + disks is cheaper than the dedicated SAN server machine of course.

        I'd really like to see them build these things with POE RJ45 sockets on the back so they could just be slapped straight onto an existing switch - but that's probably not actually all that useful once you're outside of a testing environment because the cable management would get annoying :/

      2. Anonymous Coward
        Anonymous Coward

        Re: Looking for a problem to solve...

        6TB? Do you live on your own? :)

        Last time I checked I was over 20TB at home and I'm not even trying to break records.

        I have friends with 60TB+ at home... Must have a lot of pr0n.... :/

  2. jms222

    I am biased but _everything_ should be on ethernet including disks and displays so we can get rid of SAS, Fibre Channel, HDMI, DVI, SATA and possibly even PCIe. which are at some level (packets over SERDES) all the same* and just have different connectors to be a pain in the arse.

    Whether object as opposed to block storage makes sense is another thing.

    * Save the possible analogue bits on DVI

    1. Anonymous Coward
      Anonymous Coward

      What would we plug our banks of ethernet adaptors into without pcie?

      Also, mightn't latency somewhat mar your bizarre nirvana?.. scrap all but FC perhaps?.. would probably bring the price down a bit to boot.. :)

    2. Tom 38

      I get what you're saying, but PCIe = 7.8 Gbps per lane, or 126 Gbps for a 16x link.

  3. Alistair
    Coat

    SOC running a disk.

    Lord.

    ethernet please, although I don't care if its on copper or optical.

    connect em all to the network directly, slap clustering software on the SOC. Use IPMI/RESTful.

    do it smart and use IPV6 local domain for the disk to disk comms and clustering control, put the objectstore REST api on ipv4 for the clients.

    Wheeeee. 4 or 5 C&C/Index servers. Who needs big storage systems.

    *grin*

    <I'm tired, punchy and feeling adventurous. Take away my sudo su - now!!!>

  4. Storage Guru

    SOC Running Directly on Disk

    The Marvell concept is interesting, but is a single-source, proprietary solution that would be difficult to find adoption as dual-sourcing is certainly required to have adoption move forward.

    For Ceph, the better solution is Ceph running directly on the disk drive. Seagate Kinetic API can run with Ceph but also the SOC on the Kinetic drive could run Ceph OSDs on the HDD as well.

    WD Labs previewed Ceph running on a 504 Ceph Node cluster in a single rack >4PB during OpenStack conference. Attaining speeds of ~ 90 MB/s per HDD! With only HDDs and TOR switches in the rack (no servers involved). The TCO/savings seems obvious.

    That takes care of the latency and performance issue. As stated in the other comment, SOC running on the HDD would be the most cost-effective method due to volume as well. Flash drives could do similar - with 64-bit SOCs on each Flash/SSD/NVMe drive! You could then mix and match solid and rotating media in the rack to suit use cases and load.

    The SAS/SATA connector is superior over RJ45 since it allows higher density EBOD Ethernet enclosures to utilize the passive SAS backplane to conduct power and Ethernet signals with the industry standard HDD connector.

    The Micro-Server HDDs could now investigate hyper-converged at the drive level!! MapR or other analytics running on the drive w Ceph OSDs? many hyper-converged use cases could be easily imagined! This has the potential of being a cool, disruptive movement.

  5. cloudguy

    Kinetic HDDs...not much to show so far

    Well, it was exciting to see the video of Mr. James Hughes from Seagate present and demo a Seagate Kinetic HDD at the RICON West Conference (Basho Riak) in San Francisco in October, 2013. Every OBS software vendor had a comment about Kinetic...a few were interested and willing to investigate it, while others were not convinced that Kinetic worked any better than what they were already doing with their OBS clusters.

    So, here we are almost three years later and no production quality deployments of Kinetic at scale. Maybe OpenIO is on to something, but it looks more like an engineering project right now. Maybe it will be "insanely great" and maybe it won't.

    In the meantime, every OBS software vendor wants to deliver their OBS software to solve current customer storage and data management problems at scale. No customer is going to wait another couple of years to see if Kinetic works as originally conceived by Mr. Hughes at Seagate. In fact, most customers using OBS software don't really care much about the hardware that does the storage. What they care about more is the cost of the storage and the ecosystem of solutions they can choose from to solve their data storage and management problems. From this perspective it is all about S3 and using OBS software on commodity hardware, and deploying it with the fewest headaches possible. A tricked-out Kinetic HDD without the OBS application software is useless. It is all about the OBS software.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon