back to article Hyperconvergence: Just where is the technology going?

When I started in business IT back in 1989, the machine room housed an IBM System/38 and an IBM PC-AT. The latter was the Novell NetWare 2.0a server. The S/38 had its proprietary connections, and the PCs were connected by traditional Token Ring. In fact a couple of PCs had IBM 5250 adaptor cards and terminal emulators so they …

  1. F0ul

    You've missed a bit!

    There is a total disregard for the concept of virtualization here. By focusing on the hardware, you are missing out on the real revolution of the last few years. The splitting of processing from the hardware doing it.

    I have no intention of ever buying expensive hardware again, realizing that Azure does everything I need it to. Datacentres filled with blade servers are far more efficient than a single CPU, regardless of size. they get rid of idle time, and make it possible to centralize your hardware costs. They also make auditing and compliance a lot easier - assuming your compliance people understand the concept too!

    Technology steps from the past are very interesting as part of their time (remember the memory boards on 286's?) but are totally irrelevant today. Today its about some thing as a service, and I'm fine with that. I am far more interested in making sure that the application that makes money carries on working,

  2. Toltec

    Sooner than you think

    "although 100GbE is well and truly trotting toward us I reckon that we'll be bonding 10GbE connections for a few years yet"

    We are already running kit that has 100GbE line cards available, given we recently bonded four 10GbE on our internet feeds it will not be too long before the first 100GbE link appears here. True though in the sense that there will be plenty of connections where bonded 10Gb will be fine for quite some time in the same way that we still use many 1Gb pairs.

    1. DougS Silver badge

      Re: Sooner than you think

      Not only that, but 10GbaseT is a reality with full 100 meter length over cat6a and reduced length over plain cat6. They're even working on 40GBaseT...

  3. DougS Silver badge

    Limits of Moore's Law

    Moore's Law is quite a ways from single atom transistors, if we were able to get down to that it would have another couple decades to run. If we want to go much beyond the early 2020s we'll probably need to find a way (i.e. nanotechnology) to build stuff from the bottom up atom by atom instead of trying to deposit and etch ever finer features from the top down.

  4. HildyJ

    A geezer's view

    Bottlenecks - The biggest problem/opportunity seems to lie in the MB buses. Moving RAM to the CPU fixes one bus but the standards for mass storage buses has lagged behind the capabilities of SSDs.

    Moore's law - One can argue die size limits but keep in mind that we're still stuck in a 2D world. A 3D chip, possibly with each core as a layer or with RAM in added layers, would allow the law to outlast its creator.

    Vitualization - It's 3270s for the 21st century. People keep coming up with new variations but most keyboards remain attached to their own CPU, RAM, GPU, and some sort of mass storage.

  5. swillson_dssd

    Those pesky disks and convergence.....

    Great article, and you make the statement 'AS disks will continue to be the slow component, though, we'll look to setups that keep far more in on-board RAM than ever before.' Thats true to an extent but flash can be much closer to memory performance than disk performance. Adding RAM also forces more workload, because databases need to track more in-memory data. I have seen disks replaced by flash, and memory shrunk, to get better performance....Of course, DRAM has that annoying habit of losing data when power is switched off.......

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

Biting the hand that feeds IT © 1998–2019