Written by GoodData Author |
Architecture for Large-Scale Analytics Deployments
GoodData recently announced a partnership agreement with Snowflake. This article will help you understand how the GoodData and Snowflake technologies complement each other, especially for large-scale analytic deployments of the kind that GoodData is known for.
Why is the GoodData-Snowflake partnership so interesting?
While Snowflake is known as the data warehouse built for the cloud, GoodData is known for analytics without limits, and the combination makes it practical to deploy embedded analytics at massive scale while keeping cloud processing costs under control.
When people think about analytics, they usually envision a business analyst or a data scientist, or perhaps a sophisticated business user manipulating data: slicing and dicing and creating beautiful visualizations. These scenarios tend to occur sporadically over the course of a business day and tend to inform specific business questions or decisions.
But there is another use case for analytics (including machine learning) that is growing like wildfire: analytics at the point of work. This is the credit decision, the decision of which advertisement to put in front of a Web store shopper, the mobile service employee deciding whether to repair or replace, or the traveler in bad weather making a flight choice on her mobile phone. These use cases are very high-volume, transactional, embedded, and mission-critical.
And that’s where GoodData comes in: If you are building an insights-driven software product or deploying a very large internal application with embedded analytics, and your data is in Snowflake, GoodData can productionalize analytics and machine learning to millions of users. This means all of your users: front-line employees, supervisors, vendors, partners, and customers.
When should you consider Snowflake+GoodData?
Let’s imagine a scenario where your company has created the next hot multi-tenant SaaS application. Your platform includes both Web portals and intelligent mobile apps for both front-line workers (B2B) and consumers (B2C, think: Uber.) You are at 1000 customers and growing. The thing is, each of your customers has, on average, 1000 customers of their own. That’s 1 million users!
Applications with embedded computational and AI-driven insights at this scale require IT capabilities that are beyond those typically needed to support traditional analytics use cases.
Here are a few capabilities to consider:
- Data and Metadata distribution for data isolation and parallel computation
- Localization and customization for different locales and business needs
- Security, including GDPR, HIPAA, and other regulations
- Change management: the ability to control the rollout of new features
- Manageability: provisioning resources, configuring. and monitoring via APIs.
- Total cost of ownership: preventing runaway processing costs
- Future proofing: being able to incorporate new technologies
The main thing to recognize is the difference between data scale and application scale.
Snowflake provides the data scale: all of your data delivery with agility, flexibility, and elasticity. GoodData provides the application scale: the ability to manage the ongoing delivery of analytics and machine learning to globally deployed applications while maintaining security, governance, and price-performance.
The partnership with Snowflake allows GoodData to extend its capabilities to customers that use Snowflake as their data warehouse.
If you are interested in learning more, join our early access program today.
Written by GoodData Author |