[aptxkid]

Personally I think it’s a great response and very well written. I didn’t jump on the congrats-Databricks wagon when the result first came out because of the weird front page comparison against snowflake. Both companies are doing great work. Focusing on building a better product for your customer is much more meaningful than making your competitor look bad.

[tyingq]

It is well written, but there's some sleight of hand here and there too. Like using your lowest tier product to demonstrate price/performance against a competitor's highest tier. The Snowflake lowest tier doesn't have failover, for example...or compliance features.

[buttaphingas]

This is incorrect. Every edition of Snowflake is deployed across multiple availability zones with automatic failover in the case of failure or AZ outage. This is included in the price and requires no configuration by the customer. Cross-cloud/region failover requires the top edition and a few lines of SQL to configure (plus cloud egress costs for data replication).

The higher editions of Snowflake include features like materialised views, dynamic data masking, BYOK, PCI & HIPAA compliance etc., non of which are required for the benchmark.

[tyingq]

I'm getting it from Snowflake's own page:

https://www.snowflake.com/pricing/

Amongst other things, listed under the enterprise tier, and not lower tiers, is "Database failover and failback for business continuity".

"The higher editions of Snowflake include features like materialised views, dynamic data masking, BYOK, PCI & HIPAA compliance etc., non of which are required for the benchmark."

Yeah, but they are referencing a price/performance comparison to a Databricks tier that DOES have those things. That's the point. Update your own numbers with a lower tier, but don't update the competitor tier too?

[buttaphingas]

The "failover and failback for business continuity" is specifically for cross-region/cloud, i.e. this is something you explicitly have to do. tbh I've never used it, as I guess this would be only for very large accounts. But all editions have automatic failover between AZs out-of-the-box.

[Edit] Highly Available would be a better description per region, as that's out of the box with no configuration. e.g. if a node dies, your cluster will automatically heal and resubmit your query. If there's an entire AZ outage, your query should be resubmitted in another AZ. I think this is why failover/back is called out separately, as that's not automatic, incurs additional costs etc. Here's a link with an explanation: www.snowflake.com/blog/how-to-make-data-protection-and-high-availability-for-analytics-fast-and-easy

I didn't know DB did MVs, masking etc., so yes, that makes sense. Maybe a better idea would be to have a minimum offering comparison, and then a maximum offering comparison (with multi-AZ failover, masking feature costs etc. included) - the reality for a customer would be somewhere between those extremes.

[uvdn7]

Exactly. That’s why I think public benchmark war is just a waste of time. There will ALWAYS be some subtle differences between the two platforms that results will never be apple to apple.