Elastic Load Balancer | Handling Heavy Load on On-Premises Infrastructure

Best Practices for Handling Heavy Load on On-Premises Infrastructure

Question

I have certain applications On-Premises that experience times within a year where infrastructure takes a heavier load impact (e.g., Christmas, Thanksgiving, etc.) than other times in the year.

You do not want to decommission the on-premises infrastructure.

What is the easiest and most cost-effective way in which I can handle this load?

Answers

Explanations

Click on the arrows to vote for the correct answer

A. B. C. D.

Answer: D.

Option A is incorrect.

On looking at the scenario, we see that the variable load is only for a specific duration of time.

So moving the infrastructure entirely to a Public Cloud will not be the best solution to gain maximum benefit out of the elastic nature of a Public Cloud.

Option B is incorrect.

A Private Cloud will be more beneficial where there is a consistent load rather than a variable load.

It will be good to have a Private Cloud hosting the applications for economies of cost and agility rather than elasticity, which can be best obtained using a Public Cloud.

Option C is incorrect.

Reserved Instances are usually best chosen where there is consistent usage and predictable load for a certain duration of time (1 - 3 Years)

Scheduled Reserved Instances have the ability to reserve capacity for a predictable recurring schedule that may be in a day, week, or month.

The advantage here is w.r.t the costs that I incur for Reserved Instances that can be managed without paying everything upfront rather than elasticity.

Option D is CORRECT.

This is the best way to reduce the costs of purchasing Hardware and getting the benefits of the elasticity and On-Demand pricing provided by a Public Cloud environment.

On these specific occasions, I can demand a burst capacity by going to the AWS Public cloud.

That will certainly help me maintain the performance of my applications at heavy load.

Once the load reduces, I can then terminate the instances that are no longer used to save costs.

References:

https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/burstable-performance-instances.html https://youtu.be/OzVpynd_2GM

Option A: By moving all my infrastructure to AWS Cloud and using On-Demand capacity Moving all the infrastructure to AWS cloud and using On-Demand capacity could be an option, but it may not be the easiest and most cost-effective way. On-Demand instances are the most expensive instances in AWS, and they are not suitable for long-term use. They are suitable for short-term, irregular workloads.

Option B: By creating a Private Cloud environment in my On-Premises data center that will provide me with the required elasticity Creating a Private Cloud environment in your On-Premises data center could provide the required elasticity. However, creating a private cloud environment requires a significant upfront investment, including hardware, software, and licensing costs, which may not be cost-effective. Also, maintaining a private cloud environment may be complex, and it may require a skilled IT team.

Option C: By using Scheduled Reserved Instances to match capacity reservation for the load Using Scheduled Reserved Instances to match capacity reservation for the load could be a cost-effective option. Scheduled Reserved Instances allow you to reserve capacity for specific periods, which could match the expected heavy load period. By using Scheduled Reserved Instances, you could save up to 70% on your infrastructure costs compared to On-Demand instances.

Option D: By provisioning Burst Capacity on the AWS Cloud for the duration of the load Provisioning Burst Capacity on the AWS Cloud for the duration of the load could be an option. AWS offers burstable instance types that provide a baseline level of CPU performance with the ability to burst CPU usage. However, bursting requires credits, and if you exhaust the credits, the instance's CPU performance will be throttled. Also, if the workload is sustained, bursting may not be a suitable option.

In conclusion, the most cost-effective and easiest way to handle the heavy load periods while keeping the On-Premises infrastructure is to use Scheduled Reserved Instances.