Uploaded image for project: 'IGB'
  1. IGB
  2. IGBF-2337

Clarify: Reduce costs with EC2 setting change

    Details

    • Type: Task
    • Status: Closed (View Workflow)
    • Priority: Major
    • Resolution: Done
    • Affects Version/s: None
    • Fix Version/s: None
    • Labels:
      None

      Description

      A few months ago you found some information about how to reduce costs by changing a setting during EC2 creation. It had something to do with the length of time the EC2 would be available. By setting a projected end date for the EC2's existence, we could reduce costs.

      I don't recall the details. Can you explain in more detail here?

        Attachments

          Activity

          Hide
          cdias1 Chester Dias (Inactive) added a comment - - edited

          AWS offers in all 3 type of instances
          On-Demand
          Spot
          Reserved

          Each differs from the other with respect to the costing and management. For now what we are using everywhere is On-Demand with direct cost to usage model.

          spot instance work more like auctioned price....we set a bidding price and if the quote is achieved then the spot instance is ours...until a higher quote is active, in that case, we get a warning of 10 min and then spot instance ceases to exist

          Reserved instances - These were designed for an organization which want to have production systems that are long term(3-6 years). In On-Demand the cost becomes too high if the server involved is huge.
          Reserved Instances have less cost like 60% cost reduction (up to 72% off On-Demand)so this allows us to have servers of higher config to serve more content at a reduced cost.
          How it works I believe is once a server is provisioned for 3 years. AWS manages the infra in advance so in the background they know that this will be used for longer durations so the hardware provided is more dedicated with resources that need lesser maintenance.
          OnDemand keeps changing so more optimizations on hardware are needed. Although they are VM's they still run on fleets of distributed hardware servers.

          Show
          cdias1 Chester Dias (Inactive) added a comment - - edited AWS offers in all 3 type of instances On-Demand Spot Reserved Each differs from the other with respect to the costing and management. For now what we are using everywhere is On-Demand with direct cost to usage model. spot instance work more like auctioned price....we set a bidding price and if the quote is achieved then the spot instance is ours...until a higher quote is active, in that case, we get a warning of 10 min and then spot instance ceases to exist Reserved instances - These were designed for an organization which want to have production systems that are long term(3-6 years). In On-Demand the cost becomes too high if the server involved is huge. Reserved Instances have less cost like 60% cost reduction (up to 72% off On-Demand)so this allows us to have servers of higher config to serve more content at a reduced cost. How it works I believe is once a server is provisioned for 3 years. AWS manages the infra in advance so in the background they know that this will be used for longer durations so the hardware provided is more dedicated with resources that need lesser maintenance. OnDemand keeps changing so more optimizations on hardware are needed. Although they are VM's they still run on fleets of distributed hardware servers.
          Hide
          cdias1 Chester Dias (Inactive) added a comment -

          The above is just cost saving using Reserved instances for prod
          I feel for this project we can save cost on
          Jira server
          Nexus
          Prod server

          once prod is on Reserved we will have some cost reduction there which can be used for planning DR server
          Jira and Nexus are bigger servers we can get better cost reduction there

          Show
          cdias1 Chester Dias (Inactive) added a comment - The above is just cost saving using Reserved instances for prod I feel for this project we can save cost on Jira server Nexus Prod server once prod is on Reserved we will have some cost reduction there which can be used for planning DR server Jira and Nexus are bigger servers we can get better cost reduction there
          Hide
          cdias1 Chester Dias (Inactive) added a comment -

          Reserved instance has to separately be provisioned and then maybe disks can be attached from existing to newer systems.
          https://aws.amazon.com/ec2/pricing/reserved-instances/

          Show
          cdias1 Chester Dias (Inactive) added a comment - Reserved instance has to separately be provisioned and then maybe disks can be attached from existing to newer systems. https://aws.amazon.com/ec2/pricing/reserved-instances/
          Hide
          cdias1 Chester Dias (Inactive) added a comment -

          I have attached screenshot on how to do this

          Show
          cdias1 Chester Dias (Inactive) added a comment - I have attached screenshot on how to do this
          Hide
          cdias1 Chester Dias (Inactive) added a comment -

          Please let me know when you do this... I have never done this one ever before because they involve long term provisioning.
          Even server type too has a lot of benefits if it is t2, m4 or c4

          Show
          cdias1 Chester Dias (Inactive) added a comment - Please let me know when you do this... I have never done this one ever before because they involve long term provisioning. Even server type too has a lot of benefits if it is t2, m4 or c4
          Hide
          ann.loraine Ann Loraine added a comment -

          Thank you!
          Marking this as complete.

          Show
          ann.loraine Ann Loraine added a comment - Thank you! Marking this as complete.

            People

            • Assignee:
              cdias1 Chester Dias (Inactive)
              Reporter:
              ann.loraine Ann Loraine
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: