ElastiCache Serverless Scaling Characteristics

Question

IHAC with a number of very bursty workloads.  They are currently challenged with scaling their playback service, as its responses can sometimes be upto ~512k in size.  They're looking at caching solutions, but would like something with an ondemand payment model as their round the clock utilisation is typically less than 10%

I'm struggling to find any details about the scaling characteristics of elasticache serverless, but would like to know how aggressively it can scale, and whether there are any considerations around scaling windows and request ramp up ?

It is most likely they would be looking at the Redis based service.

Answer

Hello,

In addition to the blog, here are a few more details about how ElastiCache Serverless scales. ElastiCache Serverless constantly monitors your memory, compute, and network utilization. It enables your application to achieve micro bursts in requests per second (RPS) or data stored by scaling up in place. While scaling up is fast, it is limited, so ElastiCache Serverless also scales out based on how fast your RPS or data  stored is increasing.

In summary, ElastiCache Serverless allows bursts in traffic enabling applications to increase RPS and data stored by 25-35% instantaneously, and can double RPS and memory storage every 10-12 minutes.

Thanks,
Abhay

Answer

Hi,

Elasticache Serverless was announced recently for re:Invent. So, few performance data already available. This blog post seems to be the most detailled information on this topic: https://aws.amazon.com/blogs/aws/amazon-elasticache-serverless-for-redis-and-memcached-now-generally-available/

The announcement also says *"instantly"* : *Amazon ElastiCache Serverless helps customers create highly available caches in under a minute and instantly scales vertically and horizontally to support customers’ most demanding applications, without needing to manage the infrastructure.* See https://press.aboutamazon.com/2023/11/aws-announces-three-serverless-innovations-to-help-customers-analyze-and-manage-data-at-any-scale

Best,

Didier

ElastiCache Serverless Scaling Characteristics

Contenuto pertinente