back to article Cisco dumps its Hyperflex hyperconverged infrastructure

Cisco has discontinued its HyperFlex hyperconverged infrastructure products. The IT goliath told The Register today "evolving customer needs and market dynamics" made it rethink its plans for the equipment, which combined a hybrid cloud software stack and Cisco's own Unified Computing System, or UCS, servers to offer a rival …

  1. Mayday
    Megaphone

    Hyperfux

    I've been working with Cisco for years. I have never even had a customer ask me about Hyperflex, what it does, or even have a use case for which I could suggest it to them. Truth be told I totally forgot about the platform until I saw this article.

  2. BigZeus

    Not surprised

    I have to say I am not at all surprised.

    I have currently had the unfortunate experience to try to upgrade a Hyperflex environment for a customer and it is certainly a very sensitive bucket of shit lol!

    Anything and everything has gone wrong with it and it's only running ESXi which I have been playing with for a very long time.

    It's taken every scrap of VMware knowledge I have to get as far as I have.

    VSAN makes Hyperflex look very third rate so ditching that for a partnership with Nutanix makes sense for Cisco to try to get a foothold into the HCI market.

    1. Simon 49

      Re: Not surprised

      Anything multi-vendor is a PIA come upgrade time. Straight vSAN with other VMware products on top is bearable provided you follow the correct order, Lifecycle Manager gets better each release. Flex / Rail etc. you have to wait longer for patches or upgrade guidance and when it comes it's touch and go whether it works or not.

      1. Mike Pellatt

        Re: Not surprised

        And Hyperflex is actually "multivendor", like so much Cisco stuff (I'm looking at you too, Firepower) with a central and critical component coming from M&A activity - in this case, Springpath.

        I'm screwed by it's use of NFSv3 which leads to VMs stalling if a snapshot is removed from a different host from the running one. Somewhat screwing Druva backups. Apparently fixed by NFSv4, but the Springpath VMs don't do that.

        Won't mourn it's passing

  3. jpreis

    Consistency is king

    Anyone remember back in 2013/14 when HCI became a thing? Cisco never really committed to it. They showed interest as a founding member of VCE (a short lived JV by VMWare, Cisco and EMC) but were already starting to take their marbles and go home by the time Miguel Dell dropped by Wal Mart and bought EMC back in 2016. I'm a little surprised by this, but I'm old now and my crystal ball needs cleaning.

    1. Nate Amsden

      Re: Consistency is king

      Maybe I'm wrong but I recall VCE as a Convergence platform not a Hyper convergence platform. Not that I've had any interest in CI or HCI, just regular servers+SAN storage is good enough for me. Haven't even touched a blade server since 2011(yes I know blades are out of style now), nothing but DL38x (with 3PAR SAN) in the years since for my critical stuff. HCI sounds neat on paper, and can make some things simpler I'm sure.

      My oldest active 3PAR is a 4-controller 7450(all flash), will enter it's 10th year of service in about 45 days, 0 software failures, and just 1 hardware failure(single SSD failed early this year, still had ~88% wear left on it). Still works fine, so no reason to retire it. Have 4 hour onsite hardware support from 3rd party.

      Company I work at now has a vSAN(technically VxRail) for development stuff and am not impressed(perhaps not vSAN technology's fault, but the performance is terrible despite being all flash). I get tons of metrics out of it but no smoking gun. Fortunately(perhaps) it's been running bad for years and people are used to it. It will be retired soon.

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

Other stories you might like