Brit outfit rolls own virtual server appliances

Rack and roll


A British consulting company that specializes in server virtualization was so frustrated by the mismatch between general purpose x64-based servers and what the popular hypervisors from VMware and Citrix Systems required that it has rejigged itself into a hardware vendor pushing what it calls virtual machine appliances.

Virtual Machine Company, based in Cambridge, was founded two years ago by Steve Barnett and Nick Hutton to do virtual machine installations at the financial giants in the City of London.

Barnett, Hudson, and a number of the early employees at the company had worked in the high performance computing and telecom fields, so they had experience with systems and modifying them to run very specific workloads. They had no intention of getting into the hardware business.

"In our consulting engagements, we discovered that we could never quite get the right mix of cores, memory slots, RAM speeds, and other features that would let the hypervisors run optimally," Barnett tells El Reg.

One big issue was that on many general purpose server designs, as you add more memory sticks or fatter memory sticks, the speed of the memory has to be geared down to keep the system from choking. So VMC built a few prototype machines and proved that you could tune up a box specifically for a given hypervisor and deliver more performance than you could with a generic box from Hewlett-Packard or Dell.

Since the initial Virtual Server Appliances were launched a little more than a year ago, VMC has gone to motherboard and system maker Super Micro to create its own custom motherboards that offer fat memory capacities using dual-ranked 8GB DDR3 memory sticks but also allow for the memory to run at 1.33GHz even in large banks and even running at the lower 1.35 volts (which saves on heat). The memory is ECC and has chipkill scrubbing as well.

VMC 1200 Series Virty Server

The VMC 1200 Series virty server appliance

The VMC appliances are based on the Opteron 6100 processors from Advanced Micro Devices, and use its SR5690/SP5100 chipset to glue processors to the peripherals in the system.

The company has standardized on the twelve-core Opteron 6164HE processor, which has an ACP rating of 65 watts and which spins at 1.7GHz. To VMC's way of thinking, the extra heat created by the faster 80 and 105 watt parts are not justified by the performance those extra clocks bring, not when the real issue in a server running virtual machines is memory, disk I/O, and network contention.

That is why the VMC machines don't use disk drives, but rather come with a single 100GB flash-based SSD from Intel, which stores the virtual machine hypervisor and other systems-level software needed in the appliance.

The SSD has a very high I/O bandwidth – more than what a slew of disks could deliver, if you could jam them into a 1U or 2U rack-mounted server, which you can't. The SSDs also generate a lot less heat than disk drives, and disks are not necessary in the appliances because for most customers installing private clouds, machines are installed in parts or even in threes for n+1 redundancy.

Next page: Waste of bent metal

Narrower topics


Other stories you might like

  • DRAM prices to drop 3-8% due to Ukraine war, inflation
    Wait, we’ll explain

    As the world continues to grapple with unrelenting inflation for many products and services, the trend of rising prices is expected to have the opposite impact on memory chips for PCs, servers, smartphones, graphics processors, and other devices.

    Taiwanese research firm TrendForce said Monday that DRAM pricing for commercial buyers is forecast to drop around three to eight percent across those markets in the third quarter compared to the previous three months. Even prices for DDR5 modules in the PC market could drop as much as five percent from July to September.

    This could result in DRAM buyers, such as system vendors and distributors, reducing prices for end users if they hope to stimulate demand in markets like PC and smartphones where sales have waned. We suppose they could try to profit on the decreased memory prices, but with many people tightening their budgets, we hope this won't be the case.

    Continue reading
  • US seeks exascale systems 10 times faster than current state-of-the-art computers
    China claims to have 10 in the pipeline and may pull ahead in HPC arms race

    The US Department of Energy is looking to vendors that will help build supercomputers up to 10 times faster than the recently inaugurated Frontier exascale system to come on stream between 2025 and 2030, and even more powerful systems than that for the 2030s.

    These details were disclosed in a request for information (RFI) issued by the DoE for computing hardware and software vendors, system integrators and others to "assist the DoE national laboratories (labs) to plan, design, commission, and acquire the next generation of supercomputing systems in the 2025 to 2030 time frame."

    Vendors have until the end of July to respond.

    Continue reading
  • US weather forecasters triple supercomputing oomph with latest machines
    NOAA makes it rain for General Dynamics IT, HPE, AMD

    Predicting the weather is a notoriously tricky enterprise, but that’s never held back America's National Oceanic and Atmospheric Administration (NOAA).

    After more than two years of development, the agency brought a pair of supercomputers online this week that it says are three times as powerful as the machines they replace, enabling more accurate forecast models.

    Developed and maintained by General Dynamics Information Technology under an eight-year contract, the Cactus and Dogwood supers — named after the fauna native to the machines' homes in Phoenix, Arizona, and Manassas, Virginia, respectively — will support larger, higher-resolution models than previously possible.

    Continue reading

Biting the hand that feeds IT © 1998–2022