Reduce your AWS costs by 40% today
Building event-driven systems on AWS requires infrastructure to setup, and even an entry-level configuration can cost a lot of money for small teams and startups.
How we help teams save money
Infrastructure costs on AWS can be a big problem when developing event-driven applications. We can reduce them by at least 40%
Event-driven applications typically require a database (RDS, DynamoDB, DocumentDB or similar) a message broker (Kafka, SNS/SQS, Kinesis) and potentially a search engine (OpenSearch/ElasticSearch). Serialized is built from the ground up for event-driven systems and removes the need for many infrastructure components and plumbing.
Comparison example for startups
Small team building an event-driven application
In this example we'll look at a startup building an event-driven application. The company has a development
team of 5 people and the application is a microservice-based system running on AWS with Kubernetes. For data storage
the team uses DocumentDB as a source of truth and Kafka as a message broker to transmit events between services. The
frontend also has search support which currently is backed by an OpenSearch cluster.
Below is a breakdown of the monthly costs for the data infrastructure costs for this setup and a comparison to what it
looks like when using Serialized instead.
AWS
All cluster configurations are sized using a lower-end instances type and a smaller number of nodes. The product have
availability requirements so using a single node is not an option. Below is a summary of this setup.
Serialized
With Serialized, the team will write data to Aggregates to get the consistency
requirements that the DocumentDB database
provides. Since the data is written as events and transparently replicated to projections with search support, there is
no need
for a Kafka cluster or OpenSearch cluster. The team can rely on a single Serialized project to build the same features.
DocumentDB cluster
Managed Kafka cluster
OpenSearch cluster
Core Subscription
Project cost for 1 project
Team cost for 5 team members
Comparison example for larger organizations
Larger organization with multiple teams building a complex event-driven application
In this example we'll look at a slightly larger organization building a more complex event-driven system. The
organization has 8 teams and uses different storage technologies. Most of the teams use RDS (PostgreSQL) for data
storage and MSK (Kafka) is used as a message broker to transmit events between services (and between the teams).
Each of the teams have a separate AWS account for production and testing, respectively. Below is a rough breakdown of
the monthly costs for the data infrastructure costs for this setup and a comparison to what it
looks like when using Serialized instead.
AWS setup
The production environments are sized using a mid-to-high instances type and a smaller (but more than one) number of
nodes for availability. The testing environments are sized using a low-end instance type and single nodes if possible.
Serialized setup
With Serialized, the team will write data to Aggregates to get the consistency
requirements that the RDS database provides and use Projections to replace the query logic from the SQL database. Since the data is written as events and
transparently replicated to projections there is no need for a Kafka cluster. The team can use a Serialized project
per team and environment to build the same features.
8 production RDS clusters
8 testing RDS clusters
1 production Kafka cluster
1 testing Kafka cluster
8 Core Subscriptions
64 team members
16 projects
Want to reduce your AWS data costs?
Contact us today to learn more about how Serialized can help you reduce your AWS data costs. We'll be happy to help you and your team get started and answer any questions you might have.
No need to change cloud provider
We're also on AWS!
You don't have to change your cloud provider or worry about where your data is stored since Serialized runs on AWS. Your processing components and application code can run in your AWS account and use our platform as an event-oriented data platform - without leaving AWS.