This document outlines some of the key differences between AWS Redshift and Snowflake Products.
AWS Redshift – Amazon Redshift is a fast, fully managed, petabyte-scale data warehouse that makes it simple and cost-effective to analyze all your data using your existing business intelligence tools.
SnowFlake – The Snowflake data warehouse is a cloud-based tool that supplies companies with flexible and scalable storage while simultaneously hosting solutions for BI.
|Cloud Support||AWS Only||Cloud Agnostic|
|Data Lake Integration||Deep Integration with Datalake Storage – S3 with out data movement||Data has to move from Customer account to Snowflake account and load into Snowflake file format.|
|Data Format||Supports multiple data formats – text, CSV, Avro,Parquet, JSON, Semi-Structured||Supports multiple data formats – CSV, Parquet, JSON, XML, Semi-Structured|
|Security||Data resides in Customer account||Data resides in Snowflake account|
|Pricing||Suitable for longer commitments with Reserved Instances pricing||Suitable for adhoc workloads. Costs will slowly add up as the usage increases.|
|Maintenance||Automated Maintenance – Vaccum, Backups||Automated. Minimal Maintenance|
|Scalibility||Concurrency Scaling with new RA3 nodes||Supports Scaling . Instant and Unlimited.|
|Performance||Highly Performant with more cutomized Optimization.||Highly Performant with auto optimization.|
|High Availablity||Multi AZ Support with Cross-AZ Cluster recovery||Multi AZ Support Available.|
|Disaster Recovery||Cross region Snaphot replication for multi region recovery.||Support Cross Region Cluster Refresh. DR can be across multi clouds.|
|End User Experience||Easier for Business users||More customization and rich feature set for technical users|
Redshift is suitable if your organization is deeply integrated with AWS Data Analytics Stack. Suitable for heavy usage and longer periods with three year Reserved Instances Commitment.
Snowflake is suitable if your organization requires adhoc data warehouse usage. Snowflake has automated pause and start of the environment and will only be charged during usage and can scale up quickly.