As a company’s data assets grow, the need for cloud computing increases in tandem. For keeping pace with this growth, Snowflake stands above the rest. What makes Snowflake so special? This cloud-agnostic platform takes the best of traditional database technology and combines it with modern cloud computing to drive the agility and innovation companies need to remain competitive. It features on-the-fly scaling, flexible clustering options, and the capability to hold several petabytes of information. Not only that, its unique architecture offers more cost-efficiency than traditional data warehouses. This overview discusses what is Snowflake, its benefits, and how it fits into your ecosystem.

Table of Contents

  1. What Is Snowflake?
  2. What Is Snowflake Architecture?
  3. What Is Snowflake?: A Robust Data Warehouse With Flexible Storage Options 
  4. What Is Snowflake?: A Data Warehouse With Immense Benefits
  5. What Is Snowflake?: A Platform that Fits Anywhere in Your Ecosystem
  6. How Integrate.io Can Help

What Is Snowflake?

Snowflake is a cloud-based data warehouse that employs a subscription-based payment model. In Snowflake, storage and computers are charged independently. This allows companies to pay for only the resources they need while retaining the ability to scale quickly. The system uses a hot/cold storage technique where data that is accessed frequently is stored in what is known as a "hot" cache. These cached queries do not incur any additional costs to retrieve the data stored there.

What Is the Snowflake Architecture?

Traditional warehouses have all services (storage and computing) bundled as a package. The company pays for all resources as a bundle regardless of whether they are used. As a result, companies often end up paying for these unneeded services. The platform eliminates this issue by separating the architecture into three distinct layers. In that way, companies only pay for the items used at each layer. 

Storage Layer

The platform provides a highly scalable storage platform that supports both structured and unstructured information. This layer contains schemas, databases, and tables. Each table can store multiple petabytes of information. The tables in Snowflake are separated into micro-partitions. These partitions represent contiguous units of storage. 

Compute Layer

The compute layer is where query processing takes place. Queries execute on the information stored in the storage layer. Snowflake uses the concept of virtual warehouses for computing. A virtual warehouse is a virtual representation of a warehouse that contains all the required resources to operate independently. 

These virtual entities contain CPU, memory, and cache required to perform data processing on the underlying information. The benefit of this approach is that different environments can be created based on each department’s specific requirements. Each warehouse can work with one storage layer. However, each has its own independent compute cluster. Thus virtual warehouses don’t interact with each other.

Cloud Services

This layer controls services that manage tasks that operate in compute and storage layers. Examples of these services include:

  • Infrastructure and User Management
  • Authentication
  • Access Control
  • Data Sharing
  • Query Compilation and Optimization

What Is Snowflake?: A Robust Data Warehouse With Many Options 

Snowflake Can Be Used as a Data Warehouse and Data Lake

The platform provides versatility in storage mechanisms by providing the option to store information in a Data Lake or a Data Warehouse. 

data lake is a repository that holds information in its native format in a flat architecture. The information can include a mix of structured and semi-structured information. A data lake is ideal for companies that want to store a lot of information, but don’t need the extra step of transforming the information.

A data warehouse also stores a large volume of information. However, it differs from a data lake in that it holds information that has been transformed or enriched. As a result, the information stored in a warehouse allows for advanced querying and analytics. 

Flexible Clustering Options

For those with revenue models tied to high-availability systems, an underperforming technology stack can wreak havoc on a company’s bottom line. 

Clustering is a methodology that the platform uses to increase the performance of the technology stack. Clustering involves deploying multiple processors to handle the workload. 

Flexible Architecture

Snowflake provides a flexible architecture that can be configured for your company’s needs.

Shared-Disk Architecture

The shared disk architecture involves all computing nodes sharing the same disk. Every node has a private processor, but each process can access all disks in the storage layer. With so many performing database operations, each node needs to have an updated copy of all information. Clustering control software is required to monitor this environment to avoid information access issues. The challenge with this approach is that the amount of monitoring required to keep things running smoothly can degrade system performance. 

Shared-Nothing Architecture

In a shared-nothing architecture, each node has its private memory and disk space. The database tables are spread over multiple machines. When the cluster receives a query request, every node in the cluster executes it against the portion of the information they store.

Hybrid

The platform can be configured for a hybrid architecture that combines features from shared-nothing and shared-disk architectures. In the hybrid architecture, Snowflake uses a central repository for information that can be accessed from all compute nodes. The system uses what is called Massively Parallel Processing (MPP) clusters where each node stores a portion of the data set locally. This approach provides the information management simplicity of a shared-disk model plus the performance benefits of the shared-nothing model.

Cloud Agnostic

Snowflake does not run on its own cloud. Instead, it is available on all three cloud providers (AWS, Azure, and GCP). Companies can easily integrate the platform into their current cloud architecture in a way that makes sense for their business. The platform, however, is limited to the cloud only. It can not run in a private cloud or on-premises.

Separate Storage and Compute Provides Ease of Scalability

Snowflake offers a three-layered architecture that can be scaled independently. This avoids performance issues to prevent users from competing for resources. Companies can simply scale what is needed at each layer to meet traffic demand. 

With Integrate.io, you can consolidate information from various systems into Snowflake regardless of which cloud provider you use. You even port your Snowflake integration from one cloud provider to the other. 

What Is Snowflake: A Data Warehouse With Immense Benefits

Snowflake’s robust architecture features several benefits. Each of these taken alone or combined make it an ideal platform for companies of any size.

Centralizes Data

The platform centralizes information to make it easy to provide access to specific information assets without requiring complicated configurations or data transfers.

Performance

The system includes an automatic query optimization feature. You won’t need to spend time optimizing queries or tuning the system. The platform takes care of this for you. According to Snowflake, there are “no indexes, no need to figure out partitions and partition keys, no need to pre-shard any data for distribution, and no need to remember to update statistics.”

Cost-Efficient

By separating compute and storage, company’s can pay for only the features they need and use. Pricing is based on an on-demand system. With this approach, companies can pay for scalability at any layer rather than being forced to pay a bundled cost for resources they may not use. 

Robust Data Sharing

Data sharing helps break down silos by providing easy access for anyone that needs to see the information. Snowflake allows Data Sharing of selected database objects with other Snowflake accounts. Items that can be shared include:

  • Tables
  • Secure Views
  • Secure Materialized Views
  • Secure UDFs

No information is copied between accounts, thus shared information does not contribute to a company's monthly storage costs. All sharing takes place using Snowflake’s services layer and metadata store.

Integrate.io’s platform allows you to realize these benefits by providing integrations to centralize your data assets. 

What Is Snowflake?: A Platform that Fits Anywhere in Your Ecosystem

Not every company will use the platform in the same way. The beauty of Snowflake is that it is flexible enough to suit any implementation. Some companies may only use it to store a subset of information. Others may use it as the central hub of data that gets used throughout the enterprise.

How Integrate.io Can Help

Integrate.io’s low-code/no-code platform allows you to ingest data from your systems into Snowflake with little to no technical experience required. The tool features a huge catalog of pre-built integrations to many popular enterprise systems. It can help you extract data, transform it and store it in an environment where all stakeholders have access to the information. 

Integrate.io is your all-in-one data management tool for integrating all of your systems into Snowflake. Now’s the time to leverage your warehouse for deep analytics. Learn how Integrate.io integrates with Snowflake by signing up for a customized demo.