Snowflake Cloud Data Warehouse
Snowflake is fast becoming the defacto cloud data warehouse. Its rich features, instant accessibility, scalability, and data sharing make it a natural choice for cloud-first environments. This became apparent from numerous conversations I had with attendees at the recent Gartner Data and Analytics conference. The anecdotal evidence is clear: most people I spoke to were from organizations that had selected Snowflake as their cloud data warehouse.
However – without exception and usually within the same breath – people expressed concern about Snowflake pricing and pricing transparency, particularly when handling 24/7 workloads at scale. Snowflake’s ability to grow and shrink to fit the customer’s needs can soon become a runaway cost scenario as customers pay premium on-demand pricing for always-on workloads.
One way of lowering Snowflake costs is to offload complex, always-on workloads to a more cost-effective platform such as the Yellowbrick Data Warehouse. Use Snowflake for its great data-sharing capabilities and Yellowbrick for its ability to meet time-critical reporting SLAs, such as regulatory compliance. This “better together” story boosts overall performance while lowering overall costs, ultimately giving a competitive advantage over those companies that choose Snowflake alone.
Yellowbrick has been built as a high-quality, enterprise-grade database supporting highly concurrent, ad-hoc queries by thousands of users across complex schemas and changing data. It supports such mixed workloads with strong transactional consistency and the high availability required for critical business applications. Significant performance gains come through the efficient use of resources, which equates to lower costs.
This architecture was perfected on-premises, but early in 2022, we launched a multi-cloud version that achieves these same performance gains. While a late entrant to the cloud, the technology is robust and proven at some of the largest companies on the planet. The cloud version is quickly gaining market traction and is already in use at key Yellowbrick customers.
Snowflake Pricing Model
This table summarizes application characteristics that are a good fit for Yellowbrick.
|Application Characteristics||Yellowbrick Capability|
|Rapid growth||– Predictable, low-pricing model|
– Single node scalability
|Short processing windows (with costly SLAs)||– High performance|
– Real-time streaming data ingest and
high-performance query execution
|Hybrid cloud deployment||Hybrid cloud deployment|
|Security needs to be aligned with other application components||In your VPC – full control over network path and firewalls|
|Large user population||– High concurrency support|
– Workload management – predictable latency for operational workloads
– Penalty box “bad actor” queries that monopolize resources
|Complex mixed workloads||Workload management – fairly balance and share resources based on policy|
|Need for consistent performance and QoS / SLA||Workload management – fairly balance and share resources based on policy|
|Sensitive datasets or data items||In your VPC – no shared metadata and full local network control: no data leaves the customer’s domain of control|
|Streaming data / low latency||Support for Kafka streaming|
|Need for single-row inserts||Support for high performance, low latency single row inserts|
|24×7 operations (or peak operations > 50%)||Subscription-based pricing|
|SQL-based stored procedure logic||SQL-based stored procedure support|
|Multi-tenant applications serving multiple customers/business units||– Support for secure multi-tenancy through multiple databases, Virtual Compute Clusters, and DW Instances|
– No limitations on multi-tenancy
The inherent management simplicity of both Snowflake and Yellowbrick lend themselves to a best-of-breed approach. However, Yellowbrick won’t burden you with a complex or expensive pricing model. For example, Yellowbrick pricing is predictable – you don’t need to worry about buying credits, or if those credits will get canceled if you don’t renew. With Yellowbrick, choose the pricing plan you want: on-demand or subscription. For subscription, simply decide what size data warehouse you want (vCPUs) and pay a ﬂat subscription over a 1- or 3-year term. You can consume as much of Yellowbrick as you like, 24×7 and 365 days of the year, with our advanced workload management utilizing resources and preventing cost overruns. Additionally, Yellowbrick runs in your cloud account, so you pay your infrastructure costs with no additional markup.
In summary, Snowflake customers are becoming increasingly concerned about spiralling costs. Offloading complex workloads from Snowflake and using Yellowbrick for heavy lifting can lower costs and protect your Snowflake investment.