Announcement

Collapse
No announcement yet.

Amazon EC2 Cloud Comparison Performance Benchmarks

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Amazon EC2 Cloud Comparison Performance Benchmarks

    Phoronix: Amazon EC2 Cloud Comparison Performance Benchmarks

    As it's been nearly one year since we last put out some reference benchmarks of Amazon's EC2 Cloud, or compared the different operating system's in Amazon's cloud, it's time to kick off a new round of cloud performance benchmarks. Coming out today are new benchmarks of Amazon's Elastic Compute Cloud when benchmarking various on-demand instance types.

    http://www.phoronix.com/vr.php?view=19125

  • #2
    Would be nice to see some bare metal hardware included in the list for comparison with dedicated servers.

    Comment


    • #3
      Originally posted by krasnoglaz View Post
      Would be nice to see some bare metal hardware included in the list for comparison with dedicated servers.
      Agree, but at least those number are still nice to do your own comparison, if you don't mind testing yourself.

      Comment


      • #4
        Originally posted by krasnoglaz View Post
        Would be nice to see some bare metal hardware included in the list for comparison with dedicated servers.
        I'd also be interested in seeing popular server distros (ubuntu, debian, fedora, scientific/centos), and newer kernels benchmarked against this [RHEL-based] "Amazon Linux". I ran a rails development server (small) on Amazon briefly about 4-5 years ago, using an ubuntu ami. It was pretty pokey as I recall. I'm also intrigued by the announced (unreleased) "OSv".
        Last edited by dsmithhfx; 09-20-2013, 08:40 AM.

        Comment


        • #5
          Regions and AZ's

          I would love to see a comparison of the different regions (east / west) and availability zones (a,b,c,d).

          In any case, keep up the great work!


          --shamer

          Comment


          • #6
            Originally posted by shamer View Post
            I would love to see a comparison of the different regions (east / west) and availability zones (a,b,c,d).

            In any case, keep up the great work!


            --shamer
            Likely not to happen as too costly with too many instances and likely not enough traffic/interest to make up for it.
            Michael Larabel
            http://www.michaellarabel.com/

            Comment


            • #7
              Originally posted by krasnoglaz View Post
              Would be nice to see some bare metal hardware included in the list for comparison with dedicated servers.
              I have some data to release next week.
              Michael Larabel
              http://www.michaellarabel.com/

              Comment


              • #8
                Any interest in identical benchmarks performed in digitalocean, linode, etc as well? I've got some free time and I might be able to do it over the next week or so.

                Comment


                • #9
                  Originally posted by PelPix View Post
                  Any interest in identical benchmarks performed in digitalocean, linode, etc as well? I've got some free time and I might be able to do it over the next week or so.
                  I have interest but haven't been able to do them yet due to cost.
                  Michael Larabel
                  http://www.michaellarabel.com/

                  Comment


                  • #10
                    How long do the benches take? 2, 3, 4 hours? I can't see that costing more than $50 total to test all the tiers on an ultralight service like digitalocean. If it's more like 10 or 12, I can understand. Linode may also be significantly more expensive because of their larger and more useful toolset though.

                    I think linode and AWS are ten times as expensive as digitalocean, as far as I can remember, but you get what you pay for with more load distribution flexibility and additional services.

                    Let me do the math and see if I have the money either.

                    Edit:

                    I forgot. Linode doesn't have per-hour billing, does it? That limits things a lot, so I guess it's just down to DigitalOcean and AWS unless we want to spend a truckload of cash.
                    Last edited by PelPix; 09-20-2013, 01:10 PM.

                    Comment


                    • #11
                      [QUOTE=PelPix;358577]How long do the benches take? 2, 3, 4 hours? I can't see that costing more than $50 total to test all the tiers on an ultralight service like digitalocean. If it's more like 10 or 12, I can understand. Linode may also be significantly more expensive because of their larger and more useful toolset though.

                      I think linode and AWS are ten times as expensive as digitalocean, as far as I can remember, but you get what you pay for with more load distribution flexibility and additional services.

                      Let me do the math and see if I have the money either.

                      Edit:

                      I forgot. Linode doesn't have per-hour billing, does it? That limits things a lot, so I guess it's just down to DigitalOcean and AWS unless we want to spend a truckload of cash.[/QUOTE

                      Im usually lucky if a enterprise/virt article makes even 50$ in ad views, not to mention my time required, etc.
                      Michael Larabel
                      http://www.michaellarabel.com/

                      Comment

                      Working...
                      X