MemberFebruary 2, 2024 at 11:12 pm
A multinational corporate and investment bank is regularly processing steady workloads of accruals, loan interests, and other critical financial calculations every night from 10 PM to 3 AM on their on-premises data center for their corporate clients. Once the process is done, the results are then uploaded to the Oracle General Ledger which means that the processing should not be delayed or interrupted. The CTO has decided to move its IT infrastructure to AWS to save costs. The company needs to reserve compute capacity in a specific Availability Zone to properly run their workloads.
As the Senior Solutions Architect, how can you implement a cost-effective architecture in AWS for their financial system?
1- Use Regional Reserved Instances to reserve capacity on a specific Availability Zone and lower down the operating cost through its billing discounts.
2- Use On-Demand Capacity Reservations, which provide compute capacity that is always available on the specified recurring schedule.
3- Use On-Demand EC2 instances which allows you to pay for the instances that you launch and use by the second. Reserve compute capacity in a specific Availability Zone to avoid any interruption.
4- Use Dedicated Hosts which provide a physical host that is fully dedicated to running your instances, and bring your existing per-socket, per-core, or per-VM software licenses to reduce costs.
WRT option 2: AFAIK there is no recurring schedule with capacity reservation. The company should write a script that creates a reservation starting every day at 10 PM and then cancel it at 3 AM the following day… or do that manually. Also, a capacity reservation does not come with a price discount, so this option is not going to be cheaper than option 3… I understand that 3 may be wrong because it implies the capacity reservation is running all the time, is that right?
AdministratorFebruary 6, 2024 at 3:02 pm
Thanks for your feedback.
The question focuses more on what would enable them to reserve capacity within a specific timeframe, rather than on the best way to implement it. As you mentioned, they can either automate the process by running a script on schedule or handle it manually.
You’ve raised a valid point regarding cost savings. Capacity reservation, by itself, does not come with discounts. However, pairing them with Savings Plans and Regional Reserved Instances can help reduce the costs associated with running the instances. On another note, since the company only needs the instances to run for 5 hours a day (every night from 10 PM to 3 AM), committing to a reserved instance or a savings plan for 1 year or 3 years won’t justify the costs. The company could simply terminate the instances and cancel reservation after every processing. We’ll make sure to clarify this in the explanation.
Please let us know if you need further clarification.
Carlo @ Tutorials Dojo
Log in to reply.