Amazon Redshift: easy, simple, fast
June 08, 2019

Amazon Redshift: easy, simple, fast

Jacob Biguvu | TrustRadius Reviewer
Score 8 out of 10
Vetted Review
Verified User

Overall Satisfaction with Amazon Redshift

Capital One has many LOBs (line of businesses). I have supported IAM and Commercial LOB. They are using Redshift as a data warehouse solution. Oracle is not a Data Warehouse solution but was being used when the application was on on-premises. When they wanted to migrate all data to the Cloud, they chose Redshift as a solution to move the data from Oracle. Oracle is not a data warehouse solution. Redshift has been found as a good solution because of its unique features such as its MPP architecture, columnar architecture, and storage capacity.
  • User-experience. The user wants something quick to view the output, rather than spending too much time
  • preparing a code prior to seeing the output. Redshift provides SQL type queries. This makes any user happy and comfortable.
  • Architecture is very straightforward and simple to understand, such as MPP architecture, Encryption, and Columnar database design. We can easily address issues and help others to understand.
  • Scalability. We can scale-up and scale-down based on our workloads.
  • Performance tuning and database optimization can be done using the system tables and advisors. These solutions are similar to the solution available for Oracle SQL Server. It makes it easy to do the optimization for queries and databases.
  • The concurrency and scale up based on it could be improved. It would be good if it scale-up and scale-down the memory/CPU capacity automatically based on workload.
  • Often we experience slow on queries and dashboards. Self-tuning option in WLM does help.
  • Optimizing the areas such as Vacuum and reorganize the column data (sorting over time) automatically.
  • As long as you have the perfect designer and architect, you will receive the expected benefit from Amazon Redshift.
Snowflake supports semi-structured data types and provided solutions to manage/process the semi-structured data. It supported sharing data between the different accounts and makes it easy in the scale and scale down process. Snowflake doesn't limit users on the database.
All technologies have their own pros and cons, and Redshift has pros and cons. For example, in their architecture, the distribution/query process happens on the compute nodes where data is stored. The documentation suggested identifying the best distribution keys and sort keys based on the use of cases before loading the data. The dimensional model also suggests certain recommendations. If we follow these steps properly, certainly we can mitigate performance degradation. This would need data modeling knowledge as well as designing knowledge. A con is that design and modeling skills are required for a better result.

My understanding is that if we take proper measurements at the design phase and decision phase, everything would be up to our satisfaction. Hence I rated 8.
Amazon Redshift is good for when you need a Data Warehouse solution or a user-experience such as SQL kind queries. It is also good if you have limited budget constraints. It's not suggested when someone has DML queries such as INSERT, UPDATE, DELETE.