What is our primary use case?
SkySQL serves as our transactional layer where we store metadata for all our applications before moving data into our Snowflake database. We receive transactional data from various applications, store it first in SkySQL, then transform it, load it into an S3 bucket, and finally into Snowflake for reporting purposes.
We chose SkySQL because Snowflake is used only for reporting and does not have a transactional layer. Previously, we used a SQL Server database before migrating to SkySQL. We load transactional data first because SkySQL is fully managed in the cloud, eliminating concerns about scalability or performance. We often receive high volumes of transactions from applications, which made SkySQL an ideal choice. After loading all transactional data into SkySQL, we convert that data into aggregate formats for further reporting.
What is most valuable?
Since we do not have to manage servers and memory as they are completely managed by the cloud, this has been extremely beneficial. Whenever we receive high volumes of transactions, the database auto-scales and improves performance automatically. This is particularly helpful because we experience variable transaction loads, and the auto-scaling and down-scaling capabilities provide cost optimization benefits.
As a data engineer, the most valuable aspect is not having to worry about scalability or performance. I can focus on the actual data itself: whether it is dirty or clean, and what transformations I need to apply. I work on the actual data rather than on infrastructure tasks related to servers.
All our applications are integrated with SkySQL only, and the integration capabilities are powerful. SkySQL can be integrated with a wide range of databases, cloud providers, and supports multiple frameworks, including NCP. Regarding security, it is highly secure because we conduct penetration testing before pushing data from applications to SkySQL.
What needs improvement?
Currently, we have not identified significant improvement areas for SkySQL. However, enhanced AI functionality would be beneficial. We use AI in Snowflake but not in SkySQL. If SkySQL could write SQL queries automatically using AI or improve SQL query performance, this would be valuable for future use cases.
When comparing SkySQL to other databases such as Databricks and Snowflake, these platforms offer more AI-related functionality and data lineage capabilities, allowing visibility of the complete data journey from source to target. SkySQL currently lacks data lineage functionality. Since we transform data using stored procedures and views, implementing data lineage functionality would help us track the complete journey of data from source through all intermediate transformation steps to the final layer. I rate SkySQL an eight out of ten.
What do I think about the stability of the solution?
SkySQL has been stable in my experience with no downtime or issues.
What do I think about the scalability of the solution?
SkySQL's scalability is impressive because it auto-scales automatically. Whenever we receive high volumes of transactions from applications, it auto-scales memory and server utilization without any manual intervention. When transaction loads decrease, the server and memory down-scale automatically. This high scalability means that when we onboard new applications that load data into SkySQL, we do not have to worry about scalability as it is completely handled by the cloud provider.
How are customer service and support?
As a data engineer, I have not personally reached out to customer support. However, during the implementation phase, our infrastructure team contacted their customer support for initial setup and received comprehensive support and full responses from them. I would rate customer support as nine out of ten.
Which solution did I use previously and why did I switch?
Before SkySQL, we used a SQL Server database. With that database, we required a permanent infrastructure team because whenever we faced performance or memory allocation issues, we had to reach out to the infrastructure team and wait for their availability to address server-related issues. After adopting SkySQL, this process is completely automated. We never face performance issues, and backup and scalability are all managed automatically, which has significantly improved our SLA.
How was the initial setup?
SkySQL was purchased through the AWS Marketplace and is completely managed by our infrastructure team.
What was our ROI?
We have seen a clear return on investment through team size reduction. Previously, we had a four-member team where two members managed the infrastructure side, handling server management, performance, and server optimization, while two people worked as data engineers on the actual data. After adopting SkySQL, the two infrastructure team members were no longer needed because the platform is managed by the cloud. Currently, we have only a two-member team of data engineers working on actual data. We have reduced our team size by fifty percent.
What's my experience with pricing, setup cost, and licensing?
My experience with pricing, setup cost, and licensing is that it is a one-time cost that provides great flexibility after setup. However, in our project, this was completely managed by our infrastructure team, and we as data engineers were not involved with the licensing or infrastructure aspects.
Which other solutions did I evaluate?
Before choosing SkySQL, we evaluated other options, including the MySQL database. However, MySQL presented the same challenges requiring a separate infrastructure team. We adopted SkySQL because it functions like MySQL but is completely managed in the cloud.
What other advice do I have?
Reducing our team size has significantly affected efficiency and costs for our organization. This cost efficiency benefit resulted from replacing two infrastructure team members since the infrastructure work is now automated, eliminating the need for a dedicated infrastructure team. This is very helpful for cost optimization and provides complete monitoring capabilities. You can view server health and optimization from a central point, allowing visibility into when you receive higher data volumes and when volumes are lower.
Regarding SkySQL's AI capabilities, we are using the AI functionality to write SQL queries. Whenever we face issues with SQL queries that join multiple tables, we use the AI capability to fix errors automatically. This saves considerable time for data engineers.
The accuracy and reliability of SkySQL's AI output is quite consistent. We conducted a proof of concept comparing manually written SQL queries with queries generated by SkySQL's AI capabilities, and both approaches yielded the same results. This demonstrated that the AI functionality is highly accurate.
My advice to others evaluating SkySQL is that if you are primarily working on analytics, business intelligence, or reporting, other options such as Databricks or Snowflake may provide higher functionality related to analytics and GenAI. However, if you are working with a transactional database or seeking a platform to store transactional data, SkySQL is worth evaluating. I have rated this product an eight out of ten overall.