×
Community Blog Maintaining Availability With Auto Scaling – Part 4

Maintaining Availability With Auto Scaling – Part 4

Part 4 of this 4-part series discusses the prominent usage scenarios associated with Alibaba Cloud Auto Scaling.

By Shantanu Kaushik

In Parts 1-3, we discussed Alibaba Cloud Auto Scaling and how it can be configured using multiple modes for a suitable scaling operation. We have established that Alibaba Cloud Auto Scaling is a very capable service that can be a vital tool while handling massive business loads. It can offer extensive availability and business intelligence without any service interruptions.

In this article, we will discuss the prominent usage scenarios associated with Alibaba Cloud Auto Scaling.

Usage Scenarios

  • No Backup or Maintenance Costs

Alibaba Cloud Auto Scaling creates and releases ECS instances automatically based on the requirements of your business and applications. You do not need to maintain any backup resources, and you only need to reserve compute resources for daily business traffic. This depends on your SLA with Alibaba Cloud.

The scaling policies associated with your custom set can be pre-configured. When the load increases, Alibaba Cloud Auto Scaling creates and adds ECS instances automatically to the whitelist of the ApsaraDB for RDS instance and the SLB backend server group.

When the load decreases, Alibaba Cloud Auto Scaling removes ECS instances automatically from the SLB backend server group and from the whitelist of the ApsaraDB for RDS instance to release the ECS instances.

  • Flexibility and Intelligence

Alibaba Cloud Auto Scaling provides a variety of scaling modes. You can select a combination of multiple scaling modes based on various changes with your business or policies.

  • Cost-Effective

Alibaba Cloud Auto Scaling is free and supports the combination of pay-as-you-go and preemptible instances associated with the ECS instances you are billed for while scaling out. ECS instances are only purchased if the current set of ECS instances are insufficient to maintain the presented loads.

The pay-as-you-go model applies to the ECS instances created to ensure service availability.

Scenario 1: Building a Scalable Web Application

In this scenario, we will discuss the inclusion of Auto Scaling while we build a scalable web application. This application can respond automatically to increases and decreases of any business loads and manage daily business and traffic surges.

We will use an e-commerce platform set to launch a promotional campaign during holidays and other prominent dates. They have decided to launch multiple shopping festivals to attract users on these dates. This e-commerce organization consults their O&M team and uses historical data with DataWorks to map an enhanced projection based on data analytics using machine learning and AI models.

This report provides a specific resource usage projection to help organizations formulate policies. However, when unexpected traffic surges are encountered during the peak hours of the sale, the O&M personnel must create ECS instances manually to counter the load. This can time-consuming, counterproductive, and may affect the application availability.

The Alibaba Cloud Solution

Alibaba Cloud Auto Scaling can help this organization to deploy its service without worrying about scaling operations.

Auto Scaling will scale the compute resources automatically based on increases and decreases in user load. There is no need for prediction and manual intervention, ensuring the availability of your application. Alibaba Cloud Auto Scaling has delivered thousands of ECS instances within minutes during big promotions, such as the Double 11 Global Shopping Festival, and responded to traffic surges automatically for continued and uninterrupted service and availability.

Result

Alibaba Cloud Auto Scaling keeps some of the ECS instances protected to ensure service availability during daily business operations. These ECS instances in the protected state cannot be removed from the scaling group or reconfigured in a Server Load Balancer instance or ApsaraDB for RDS whitelist.

The scaling group keeps the average CPU utilization of ECS instances at about 50% to maintain high availability. When the average CPU utilization exceeds 50%, Alibaba Cloud Auto Scaling automatically creates ECS instances to balance loads.

Similarly, when the average CPU utilization drops below 50%, Alibaba Cloud Auto Scaling automatically releases ECS instances to reduce costs. Here, the number of ECS instances remains greater than or equal to the specified minimum number of instances. Then, less than or equal to the maximum number of instances are defined.

Scenario 2: Online Learning Application – Cost Reduction

There are a lot of resource-hogging scenarios, but we chose to use an online education platform for the usage scenarios due to the rigorous changes and revolution in online education over the past year.

Many educational institutions are providing online classes to their students. Online learning platforms have to endure a lot of traffic load and require a highly robust infrastructure to handle the workloads. Many online learning institutions experience traffic peaks from 9:00 AM until 3:00 PM every day.

However, during other times, the business traffic is significantly lower. Alibaba Cloud Auto Scaling can help the platform deliver reliable services during peak hours by scaling the number of ECS instances automatically and reducing the ECS instances when the load is lower.

The resources are wasted without Auto Scaling. Scaling was conducted manually and was not enough to accommodate the changing traffic conditions. Furthermore, when the platform experiences unexpected traffic surges, the ECS instances must be created manually to ensure service availability.

The Alibaba Cloud Solution

Alibaba Cloud Auto Scaling provides extensive cost-cutting using the pay-as-you-go method for any ECS instances created while scaling out. During peak hours, Alibaba Cloud Auto Scaling creates ECS instances automatically based on the average CPU utilization to implement smooth scaling and availability while saving costs.

Wrapping Up

Alibaba Cloud Auto Scaling can help any business that experiences erratic traffic loads to maintain service availability and reduce resource orchestration costs.

Upcoming Articles

  1. Multi-Tier Approach With VPC – Part 1
  2. Multi-Tier Approach With VPC – Part 2
  3. Multi-Tier Approach With VPC – Part 3
0 0 0
Share on

Alibaba Clouder

2,599 posts | 762 followers

You may also like

Comments