Snowflake - flexible and isolated workloads
June 06, 2019
Snowflake - flexible and isolated workloads
Score 9 out of 10
Vetted Review
Verified User
Overall Satisfaction with Snowflake
Capital One has many LOBs (lines of business). I have supported IAM and Commercial LOBs. They are using Snowflake as a data warehouse solution. The team is also using Redshift. The team noticed the concurrency limit with Redshift. Snowflake is a simple architecture with isolated compute capabilities. Redshift does not separate storage and computer capacities. Based on the workloads, the capacity is scaled up and scaled down using the option: Dataware. Right now, Snowflake is the heart-cake requirement for the company.
- The beauty of Snowflake is virtual warehouses. This provides an isolated workload and capacity (Virtual warehouse ). This allows separating or categorizing workloads and query processing according to our requirements.
- Snowflake provides a unique architecture that supports structured and semi-structured data. provides the ready-made solutions to process with the semi-structured data such as JSON, Avro, ORC, Parquet, and XML
- SQL kind queries make easy to write/develop the code. Provided features such as materialized views to optimize the query processing and loads
- Snowflake database is managed by snowflake team. This simplifies the DBA job from maintenance kinds of tasks such as patchworks and regular upgrades.
- Performance degradation happens when many users are on the system at the same time.
- Snowflake support is slow often.
- Snowflake provides Virtual warehouses for having performance. however, it may cost more.
- Quick installation/setup for a data warehouse solution.
- Make easy to handle/manage the various type of semi-structured data using the native solutions and provided new data modeling concepts such as schema-read data model and schema-write data model.
- Make easy and simplified the workload management using virtual warehouses and materialized views.
Redshift and Hive both have unique architecture. Both have their own cons. My guess is that Snowflake is made up by using the concepts of the two architecture concepts such as Amazon Redshift and Haddop, addressed the issues or gaps found in Redshift and Hadoop.