11.2.3: Cloud-Native BI Solutions

Cloud-Native BI Solutions

As businesses increasingly move data and applications to the cloud, BI has followed suit. Cloud-native BI solutions are analytics platforms that are designed to run on cloud infrastructure and often offered as a managed service (Software-as-a-Service). These solutions leverage the scalability and connectivity of the cloud to provide BI with minimal on-premise infrastructure. They can often handle large, cloud-stored datasets and integrate with cloud data warehouses (like Snowflake, Google BigQuery, Amazon Redshift, Azure Synapse, etc.). The two examples given are Google Looker and Amazon QuickSight, which correspond to the Google Cloud and AWS ecosystems respectively. (Microsoft’s analogous offering would be Power BI in Azure, which we already covered under enterprise, and there are also others like Oracle Analytics Cloud, etc., but we’ll focus on the given examples.)

• Looker (Google Cloud Looker): Looker is a modern BI platform that gained popularity for its unique modeling layer and flexible architecture. It was an independent company until Google acquired it in 2019; now it is part of Google Cloud’s BI offerings. What sets Looker apart is its semantic modeling layer called LookML. Instead of creating ad-hoc calculations in each report, LookML allows data analysts to define consistent business logic (metrics, dimensions, relationships) in code (YAML-like syntax). These models then power Looker’s interactive dashboards and self-service exploration interface, ensuring that all users are using “one version of the truth” for metrics. Key features: API-first and cloud-native – Looker was built with web APIs and modern cloud databases in mind, meaning it queries data where it lives (it connects directly to databases like BigQuery, Snowflake, etc., and pushes down queries). It doesn’t require extracting data into its own storage; instead, it leverages the power of the underlying database (this approach is sometimes called “in-database” analytics). Looker’s front-end provides robust dashboards, scheduled reporting, and an ability to “explore” data ad hoc by pivoting and filtering within the bounds of the model. It also has strong embedded analytics capabilities: because of its API-first design, companies can use Looker’s APIs or iframes to embed charts and dashboards in other applications easily. Looker’s integration with Google Cloud means it can natively connect to BigQuery with high performance, and it’s now marketed as Google’s primary BI solution (complementing the more lightweight Google Data Studio, now renamed Looker Studio). For organizations already on Google Cloud or using cloud data warehouses, Looker provides a scalable, browser-based BI experience. It is a pure cloud product (there was an option to self-host on private cloud or on-prem, but most use Google-hosted) and is known for relatively quick deployment compared to legacy on-prem BI. Developers appreciate that they can manage LookML models in version control, bringing software development best practices to BI metrics definitions. Overall, Looker is seen as ideal for cloud-centric enterprises that want consistent governed metrics and have the technical resources to build a data model. It’s often used internally (for employees’ decision-making dashboards) but also can be extended to external customer-facing analytics since it supports multi-tenant embeds and has granular control over data access.
• Amazon QuickSight: QuickSight is Amazon Web Services’ native BI service. It is fully managed and serverless – AWS handles all the infrastructure, and it can automatically scale to thousands of users. QuickSight is designed to allow AWS customers to quickly create and share dashboards without needing to set up their own BI server. One notable aspect of QuickSight is its pricing model which can be usage-based (especially with QuickSight’s “Enterprise Edition” featuring session-based pricing). This can be cost-efficient for scenarios where many users view dashboards infrequently, as you only pay when they actively use it, unlike per-user per-month models. Key features: tight integration with AWS data sources (it can easily connect to AWS services like Redshift, RDS, Athena, S3, etc.), ML Insights (it has some built-in machine learning-driven insights like anomaly detection and forecasting provided by AWS AI services), and embedding and API support for integration into applications. QuickSight provides an in-browser editor to create visuals and dashboards, and also a natural language query feature called “QuickSight Q” where users can ask questions in plain language. Being serverless, it requires minimal administration – no servers to manage, and it takes care of scaling and high availability. AWS emphasizes QuickSight’s ability to deliver “unified data access” and put insights where users need them. For example, QuickSight can be embedded into internal portals or external software, enabling interactive analytics embedded directly into applications and workflows (e.g., a company could embed a QuickSight dashboard into their CRM system so salespeople see relevant analytics in-context). QuickSight is also notable for security features aligning with AWS standards (encryption, FedRAMP, HIPAA eligibility, etc., making it suitable for regulated industries leveraging AWS). Organizations that have much of their data infrastructure in AWS might choose QuickSight to minimize data movement – since QuickSight can sit in the same cloud environment as the data, latency can be low and data transfer costs minimized. It’s positioned as cost-effective, scalable BI for the cloud, especially useful if you have many occasional users (like a large community of external partners or customers who just view dashboards).
  

Figure \(\PageIndex{1}\): Cloud-Native BI in the Modern Data Stack. This architecture diagram shows how cloud-native BI tools like Looker and QuickSight connect directly to cloud data warehouses, querying data in place rather than extracting it, and delivering dashboards to users via browser-based interfaces. (CC BY 4.0; Felix Amoruwa) Access to a detailed description of Figure \(\PageIndex{1}\)

Other cloud BI offerings include Microsoft’s Power BI service (which is part of Power BI we discussed; it can be thought of as a cloud BI as well since it is SaaS on Azure) and Oracle Analytics Cloud, SAP Analytics Cloud, IBM Cognos on Cloud, as well as newer players like ThoughtSpot (Cloud) or Mode, Sigma Computing, etc., which are SaaS BI tools. There are also cloud-based versions of open-source or smaller tools (e.g., hosted Metabase, or services like Google Looker Studio – formerly Google Data Studio – which is a free lightweight cloud BI/reporting tool).

The advantages of cloud-native BI solutions generally are: - No/low infrastructure maintenance: The vendor manages uptime, scaling, patches, and so on. - Scalability: It’s easier to handle spikes in usage or growing data by leveraging cloud scaling (adding more nodes, etc.) which is handled behind the scenes. - Accessibility: Users just need a browser to access, and data can be integrated from various cloud sources easily, enabling remote and distributed teams to collaborate. - Frequent updates: SaaS BI tools often get new features continuously without the need for customers to do major upgrades.

Considerations when using cloud BI include ensuring data security (making sure your data sources are securely accessible to the cloud tool, dealing with network/VPC configurations, etc.), and sometimes data locality (for very sensitive data, some companies may still prefer on-prem tools, though cloud security has matured a lot).

From a cost perspective, cloud BI can switch from large upfront licenses to more of a subscription or usage-based model. This can be good or bad: it lowers initial cost but one must monitor ongoing costs, especially if queries on large datasets incur cloud compute charges. Tools like Looker and QuickSight are typically priced per user or per session, plus possibly based on compute usage.

In essence, cloud-native BI solutions like Looker and QuickSight are well-suited for organizations already leveraging cloud data platforms, who want a BI solution that integrates seamlessly with that environment and can be delivered to users with minimal IT friction. They represent the current trend of BI moving to the cloud, allowing faster deployment and easier scaling than traditional on-premise BI installations.