December 17th: Exploring an 80% lower cost hosted Redis alternative - register

Pulling the Curtain Back on AWS's Price Cut for ElastiCache

Explore AWS's recent price cut for ElastiCache and why it doesn't address the underlying scaling and performance issues, plus why Dragonfly offers a more efficient alternative.

October 23, 2024

Pulling the Curtain Back on AWS's Price Cut for ElastiCache

Last week, AWS made headlines by announcing a 20% price reduction for ElastiCache users migrating from ElastiCache for Redis to ElastiCache for Valkey. At first glance, this seems like great news: who doesn't want to reduce their AWS bill? What this move does not address, however, is the scaling and performance issues that most customers running heavy workloads on ElastiCache experience. That's because Valkey does not offer any benefit or feature beyond Redis; it's just a strategic play by AWS to distance themselves from the Redis project.

Why the Price Cut?

To understand why AWS announced this price cut, it's first important to understand why AWS has been the leader in fostering and promoting the Valkey fork of Redis.

Following the Redis licensing change announced earlier this year, AWS lost influence over a project and brand that they had built a multi-billion dollar business on in ElastiCache. This introduced two separate but related problems for AWS. ElastiCache is tightly tied to the Redis brand, and now that AWS no longer has influence over that project, they'd like to sever that tie and focus on the Valkey brand. Second is support. AWS would like to avoid supporting old Redis versions.

Now that they have committed to a strategy that involves migrating from Redis-based workloads to Valkey-based workloads, they have to figure out a way to get tens of thousands of customers to actually make the switch.

ElastiCache for Redis and ElastiCache for Valkey are the same from a product and technology standpoint. There's no difference in performance, scalability, or features. If Valkey offered value beyond what Redis offers, then customers would have motivation to undertake this migration on their own. Valkey does not offer any additional value, however, so AWS is forced to offer this price cut in order to incentivize customers to migrate.

(Not) Addressing Customer Pain Points

Whether ElastiCache customers are using the Redis-based solution of the Valkey solution, they are stuck with the same production limits they have been struggling with for years. This move does nothing to address the significant amount of work developers are investing to overcome the performance and scalability limits with ElastiCache.

ElastiCache customers are still forced to:

  • Migrate to a cluster configuration even for small workloads.
  • Horizontally scale and manage many small instances.
  • Pay for CPUs they are unable to fully utilize.
  • Protect against connection storms.
  • Over provision for snapshotting and other edge cases.
  • Implement various workarounds in order to make ElastiCache scale for their particular workload.
  • Deal with network throughput limits.

Dragonfly's Commitment to Customer Value

We built Dragonfly because we, as engineers, have felt the pain in scaling Redis first-hand many times. We wanted a solution that removes the scaling and performance bottlenecks Redis users are all too familiar with. Which is why Dragonfly scales, both vertically and horizontally, to handle any size workload. It also delivers dramatically lower management overhead, more efficient data structures, and much more efficient memory utilization.

Our goal in building Dragonfly is to be far more efficient in utilizing the modern hardware that all major cloud providers make accessible, to be able to deliver much more for less. A r7g.large instance delivers 9.8GiB of maxmemory1 for $128/month2, which comes out to about $13/GiB/month, or $12.1/GB/month.3 The same workload running on Dragonfly Cloud costs $8/GB/month, making ElastiCache over 50% more expensive.

AWS Memory-Optimized Cache Nodes for Valkey (US East N. Virginia, October 2024)

Cache Node TypevCPUNode MemoryNetwork PerformancePrice Per Hour
cache.r7g.large213.07 GiBUp to 12.5 Gigabit$0.1752
cache.r7g.xlarge426.32 GiBUp to 12.5 Gigabit$0.3496

Conclusion

AWS's recent price cut for migrating to Valkey may seem enticing, and if you are a current ElastiCache customer, you should certainly consider taking advantage of it. However, remember that the value comes from a reduction in cost in exchange for a migration that strategically benefits AWS, not from access to an improved service. For users looking for true performance gains, scalability, and cost efficiency, it's worthwhile to check out Dragonfly. If you're going to invest in a migration anyway, you might as well implement a technology that will improve your user experience and scale with your business.


Footnotes

  1. This is the maximum amount of memory that can be used by the Valkey process, which can be retrieved using the INFO MEMORY command:

    elasticache.valkey.cache.r7g.large$> INFO MEMORY
    # Memory
    # ...
    maxmemory:10527885773  # Bytes
    maxmemory_human:9.80G  # GiB
    # ...
    
  2. $0.1752/hour x 730 hours/month = $128/month.

  3. 1 GiB (gibibyte) = 2^30 bytes, and 1 GB (gigabyte) = 10^9 bytes. Thus, 1 GiB is roughly 1.07374 GB, and $13/GiB/month ÷ 1.07374 = $12.1/GB/month.

Stay up to date on all things Dragonfly

Join our community for unparalleled support and insights

Join

Switch & save up to 80% 

Dragonfly is fully compatible with the Redis ecosystem and requires no code changes to implement. Instantly experience up to a 25X boost in performance and 80% reduction in cost