Snowflake is primarily used to handle the data warehousing part, for creating data modeling, and also keeping the raw data and creating reporting data so that it is further used for data analytics.

Whenever we receive files from API integration from the front end, such as Excel files, HR data, sales data, or marketing data, we load those files into Snowflake. These files may be in non-SQL formats, such as Parquet files or text files. From Snowflake, we transform those files into tables. Once the tables are transformed, we create dimension tables according to the business needs, either in a star schema or Snowflake schema. After creating that schema, we create curated data on top of it. Once the curated data is created, we create a view query which is further used as reporting data analytics. We load the raw data, transform it, and then use it further as reporting data analytics.

Snowflake is utilized as a hybrid cloud. We primarily use it as a private cloud, but some components are hybrid.

Snowflake has a very wide variety of data integration capabilities. For workflows and data sources, we can capture semi-structured or non-structured data such as Avro, Parquet files, JSON files, and text files, or we fetch data from API integrations, such as Salesforce API or SAP data. We try to inculcate and grab all the data and then load it into Snowflake. Snowflake works as a very good tool because it handles and creates micro-partitions automatically. Whenever we create SQL queries, it automatically divides and runs the whole query itself on a multi-cloud, multi-cluster, and multi-cloud enterprise workflow.

The top features about Snowflake are first that it is hybrid cloud native. Everything is on cloud, so we have very zero maintenance for computational or in order to maintain Snowflake. Second, it is a very simple tool to understand with an architecture dividing the data layer, storage layer, computational layer, and the UI layer. This unique feature of Snowflake gives us very low costs. It is very helpful for SQL analysis. In Snowflake, we keep the data in a very low-cost manner, and if we don't compute it, we can also move it to cloud storage. Third, it is very elastic in scaling. Whenever a large amount of SQL or computational work is required, the warehouse automatically scales in and out. This is one of the best features in Snowflake that it manages automatically and runs on ACID properties: Atomicity, consistency, isolation, and durability. Each query works on isolation, hits isolation, and Snowflake very much optimizes performance. It is very simple to use.

The elastic scaling helps significantly; for example, if we get a large amount of data after a leave period, it requires scaling out the warehouse to a larger size. If we have a bulky query, the warehouse size increases automatically, dividing the query into micro-partitions for quick execution. If we encounter unexpected data, such as special characters, Snowflake creates zero-copy cloning, aiding in development and testing. Such cloning does not require physical storage, hence reducing storage costs.

Snowflake has positively impacted us by making everything cloud-native, significantly reducing the systems and application running process. The maintenance costs are low because Snowflake maintains a high amount of data efficiently at a low cost. There is also a massive scalable data warehousing performance. Whenever we need to increase SQL data, the data warehouse scales up or down based on our needs and query requirements. If the query is not heavy, the costs remain low, and the differentiation between storage and computational layers makes Snowflake very cost-effective compared to other data warehousing tools.

Snowflake has contributed to significant cost savings. Previously, we paid for both storage and computational layers, where unused data accumulated costs unnecessarily. With Snowflake, the unique approach to managing storage and computational costs allows us to transfer less frequently accessed data to Glacier, further optimizing expenses. Snowflake's performance enhances business efficiency, allowing timely achievement of targets. The analytics and transformations we execute using Snowflake are notably faster, providing significant advantages. The scalability is remarkable; marketing teams can analyze data on time-sensitive trends swiftly.

Snowflake is already quite improved, but they have recently introduced AI features. AI integration would be beneficial for direct data capturing from systems such as SAP and Salesforce to Snowflake as raw data and allow for efficient data warehousing.

Snowflake is very good overall, but it could improve documentation for supporting different structures. Even though Snowflake has a strong ANSI SQL presence, there are some features that do not perform optimally, particularly for Change Data Capture (CDC).

I do not rate it a 10 because there are still areas for improvement, such as AI integration and raw data capturing. Once those enhancements are achieved, the dependency on ETL would decrease, as would the scheduling aspect to streamline major data workloads within Snowflake.

I have been working for more than six months in my current field, and it is about a year going to happen.

Snowflake is highly stable and performs well even with large data sets exceeding terabytes, maintaining stability throughout.

Snowflake's scalability is excellent, depending on dataset sizes and query handling. It can scale from one engine to up to 32 engines seamlessly, maintaining performance.

We interacted with customer support, and they were quite helpful. They listened to our issues and provided solutions, but we faced challenges acquiring documentation for various structures. We sought this documentation multiple times but faced difficulty in obtaining it.

Positive

Earlier, we were using on-premise data warehousing with Hadoop, but it was complex and high-maintenance. The setup was also prolonged and costly. Therefore, we transitioned to Snowflake, which allows for rapid setup.

For pricing, setup cost, and licensing, everything is managed smoothly. Regarding licensing, it is inexpensive. The setup cost is low, mainly due to AWS Marketplace; we only need to pay for serverless configurations. When it comes to cloud support, the setup cost is very cheap compared to other platforms, such as Oracle or PostgreSQL, which typically require higher costs. Snowflake's handling of computational resources and hardware is excellent, making it one of the most affordable data warehousing solutions on the market.

Before choosing Snowflake, we evaluated other options such as Redshift and Databricks. What we observed was Redshift's architecture seemed coupled, leading to slower scaling and limited concurrency issues requiring extensive maintenance. Redshift's clustering methods increased complexity in data management. In contrast, we found Snowflake offers fully managed services with micro-partitioning and independent scaling, making it a more desirable option.

One more feature is Time Travel, which maintains historical data. For transient tables, it is retained for seven days, but for permanent tables, it is stored for 90 days. If we want to keep historical data beyond that, we can extend it via Time Travel. If we accidentally delete some tables, we can retrieve that data within 90 days, or if that does not happen, we can reach out to Snowflake support team for retrieval known as Fail-safe travel. Additionally, Snowflake offers very secure data sharing, ensuring no data movement among different applications while collecting data into one place for transformation and loading into data warehousing. The storage, security, and compliance are all managed well within Snowflake, and we just need to maintain the security policies as per administration.

One thing is very low maintenance. The infrastructure has transitioned to cloud, which is very manageable. Snowflake gives us very high concurrency, and we can predict our costs effectively. The enterprise data analytics we perform—whether BI reports or analytics—everything is very helpful with Snowflake.

My advice is to leverage Snowflake for its scalability, efficient query handling, user role assignments, and direct integrations with platforms such as SAP and Salesforce. The decoupling of computational and storage architecture is a key feature, allowing us to pay based on actual usage. Features such as zero-copy cloning and Time Travel enable easy data recovery and improve overall data management. Additionally, automatic tuning enhances query performance and sharing securely aligns with stringent security policies.

I believe we have covered almost everything. To summarize, Snowflake offers scalable storage that adjusts automatically, high concurrency, and virtually no downtime. It continually enhances performance by improving cluster configurations. I rate this product an 8.5 out of 10.

I started working with Snowflake when I was with Fidelity Investments around 2016-2017. We used Snowflake on AWS cloud because Snowflake doesn’t have an on-premise offering. You need to use it with AWS, Azure, or Google Cloud. 

As a consultant now, I assist enterprise customers, though I don't have Snowflake deployments yet.

Snowflake is a data lake on the cloud where all processing happens in memory, resulting in very fast query responses. One key feature is the separation of compute and storage, which eliminates storage limitations. 

It also has tools for migrating data from legacy databases like Oracle. Its stability and efficiency enhance performance greatly. Tools in the AI/ML marketplace are readily available without needing development.

Cost reduction is one area I would like Snowflake to improve. The product is not very cheap, and a reduction in costs would be appreciated.

Snowflake is very stable, especially when used with AWS. It works best with AWS compared to Google Cloud and Azure.

Snowflake is very scalable and has a dedicated team constantly improving the product. There are no problems on the scalability side.

Snowflake's technical support is excellent. During my time at Fidelity, I received great support in migrating data to Snowflake, with quick responses and innovative solutions.

Positive

The initial setup was rated six out of ten due to the time required for migrating existing data to Snowflake. Configuration and data migration are major steps involved.

Snowflake's pricing is on the higher side, rated as eight out of ten. If there were ways to reduce costs, it would be a positive improvement.

Snowflake is a great solution if you have substantial data volume. For those considering Snowflake, be prepared for the necessary initial investment in time and resources. 

I rate the overall solution nine out of ten.

We use Snowflake as a database to store all the data we stream from the source system.

The best thing about Snowflake is its flexibility in changing warehouse sizes or computational power. If I want to change the computation with other tools like Netezza, I would have to add a hard disk to it. With Snowflake, being on the AWS side, changing warehouses provides me with faster execution of my queries.

The real-time streaming feature is limited with Snowflake and could be improved.

Currently, Snowflake doesn't support unstructured data. With Snowflake, you need to be very particular about the type of data in your source systems. It has to be in a proper structure. You cannot push data to Snowflake in any possible way.

I have been using Snowflake for a couple of years.

Snowflake is a stable solution.

Around 400 users are using the solution in our organization.

We previously used IBM Netezza. We switched to Snowflake in 2020 because it provided us control over its scalability and costing model.

The solution’s initial setup is comparatively easier.

Snowflake is a cloud-based, scalable solution that provides strong data security. Handling all the data is much easier in Snowflake, and it has a very nice interface to control user access. The administration of Snowflake is also comparatively easier.

Snowflake is a cost-effective solution.

Choosing Snowflake completely depends on the quantum of data your organization has and the requirements. Snowflake is suitable for someone looking for a scalable and cost-effective solution that provides quick analysis.

Overall, I rate Snowflake a nine out of ten.

The solution is primarily used for data warehousing. 

It's a good tool. It is quite easy to manage. 

There are a lot of features. 

If you accidentally delete something, you can time travel and retrieve it. 

Within half an hour or even 15 minutes, you can set up a new data warehouse, and you can clone a table or whatever you like. 

It's easy to provide access to clients if they need to.

The solution is stable. 

It can scale well.

There is good online documentation via the community, and you can learn the solution on your own. 

These days, they are pushing users towards the GUI or graphical version. However, I am more familiar with the classic version. I'd like to continue to work with it using the older approach. That's just a personal choice. I prefer the database views like you would get on an SQL Server or other databases. 

I've been using the solution for two and a half years or so. I haven't used it for too long. 

The solution is stable and reliable. There are no bugs or glitches, and it doesn't crash or freeze. I've never had an issue when you connect it to your infrastructure. I'd rate the stability nine out of ten. 

We have 250 or more users on the solution currently. They are predominantly IT specialists, including engineers and developers. It's for those on the application side of development. The solution is scalable. I'd rate the scalability nine out of ten.

I've never directly dealt with technical support. We do have a direct contact. There are other people who might have. There might have been a version issue at some point. However, it wasn't something I dealt with. I've only interacted with the Snowflake community pages. 

We are currently moving everything from Oracle to Snowflake. 

We've used a combination of other tools in the past. We've used Microsoft's Stack, for example. We were using a SQL database in the past. 

We're migrating from Oracle right now. We have two people that handle the maintenance of the solution. A company may need two to four engineers to manage maintenance 24/7.

It's not difficult to deploy. It's pretty straightforward. It took us only a couple of minutes in order to get everything up and running. 

We pay a licensing fee to use the solution. It's paid monthly. My understanding is that it is not that expensive. We might have a specific extra cost surrounding running it on a private cloud. I'd rate the cost as six out of ten. It's moderately priced compared to others on the market. That said, I don't directly manage the account and don't directly deal with pricing. We have a contract of three to five years. 

I'm an end-user. 

Everything is on the cloud, and therefore I'm always using the latest version of the solution. It updates itself regularly. 

It fulfills our needs, and it's easy to pick up by looking at some reference guides. I'm still getting used to the GUI. I might find it even easier to use if I go through more formal training. Right now, I've simply learned it on my own. Overall, I'd rate the solution eight out of ten.