Snowflake's mission and vision is not only to solve the problems of the current data warehouse, but also to share data inside and outside the company. And I want to contribute to the new world and competitiveness that can be seen by doing so.
Unlike the data warehouse so far, it is a data warehouse that was developed from scratch on the assumption that it will be used on the cloud. By making the best use of the cloud environment, scalability with a high degree of freedom, and a unique architecture design that separates the compute node (execution processing) and storage ray (storage device), etc. Execution processing etc. are resolved.
Strengths of Snowflake Data Cloud
· Focusing on cloud storage (S3, Google Strage, Azure Storage), the data to be stored is encrypted
· The gear mark around the cloud storage becomes a virtual warehouse (compute node), and you can prepare the required number for each application.
· Compute nodes can be prepared from XS (1 server / 8vCPU) to 4XL (128 servers / 1024cCPU)
· Supports business separation by separating a cluster of compute nodes according to the application
· In other words, it is possible to deal with the problem that the dashboard display becomes slow because the ETL processing is not completed.
· There are three main billing targets
1. Storage: Show price through as is without markup
2. Compute node: Billing in seconds depends on the size of the node. If the node size is large, the processing capacity will be doubled and the processing time will be halved. As a result, higher-spec nodes can be used at the same cost.
3. Multi-cluster: Can autoscale according to load (up to 10 units)
· Clone
1. Test data can be prepared without affecting the main body data
2. Operations on clones do not affect production data
· time travel
1. Data can be retrieved back to the last 90 days, so there is no need for backup
2. You can go back in time by looking at the time stamp (however, you cannot go back 100 days, so you need to back up in 90-day units).
· Data sharing
1. If the data sharing partner uses Snowflake, you can use the partner's compute node.
2. Pass pointers instead of copies so you can share securely
· Cross cloud
1. Even if you create accounts in multiple regions, you can integrate them later.
2. Replication can also be done across regions
3. In addition, cross-cloud replication is also possible between AWS-GCP-Azure
Snowflake Data Exchange
· One step ahead of data sharing that we are about to proceed!
· A data sharing service that allows any Snowflake customer to become a data provider
· In the near future, we plan to provide "Private Data Exchange" to realize data sharing between group companies.
With the contents so far, I hope you understand that there is a big merit in terms of maintenance with unique functions that have not been found in conventional data warehouse software. You can find the best India Snowflake consultants for connecting your company to the world’s data. However, it cannot be adopted as a data warehouse just because it is easy to operate and maintain.
Post a Comment