101–110 of 111 posts
"Posting benchmark results is bad because it quickly becomes a race to the wrong solution. But somebody showed us sucking on a benchmark, so here's our benchmark results showing we're better."
Earlier quoted context omitted.
Delta lake is not meaningfully more "open" than whatever Snowflake (or BigQuery and Redshift) are doing. It does not require any less "moving data around" With all these, the data sits on cloud storage and compute is done by cloud machines - the difference between Databricks and the others is that with Databricks, you can take a look at that bucket. But you're not going to be able to do much with that data without pa…
Delta lake is very much open. You can install delta lake and run it yourself. It's a transaction layer running over parquet files. You can go to the delta.io GitHub and install binaries yourself. Snowflake cannot be run independently of their cloud. The rest of this is some vague claims of Databricks being unreliable techbros blah blah which is just emotionally charged hot air rather than being based on anything. RE…
This is the sort of FUD testing that gets thrown back and forth between companies of all kinds. If you're in networking, it's throughput, latency or fairness. If you're in graphics its your shaders or polygons or hashes. If you're in CPUs its your clock speed. If its cameras, it's megapixels (but nobody talks about lens or real measures of clarity) If you're in silicon it's your die size (None of that has mattered fo…
Exactly. Not sure about Netflix special, but there are experts that have dedicated their professional careers to creating fair benchmarks. Snowflake should just participate in the official TPC benchmark. Disclaimer: Databricks cofounder who authored the original blog post.
Earlier quoted context omitted.
Exactly. Not sure about Netflix special, but there are experts that have dedicated their professional careers to creating fair benchmarks. Snowflake should just participate in the official TPC benchmark. Disclaimer: Databricks cofounder who authored the original blog post.
The benchmark itself is kinda useless, so I don't see why they should. If you look at tpc-h for years, you had exasol as a top dog, but in the real world that meant nothing for them.
DB1.Databricks generated the TPC-DS datasets from TPC-DS kit before time started. Databricks starts time then generated all queries. Then Databricks loaded from CSV to Delta format (also some delta tables were partitioned delta tables by date) and also computed statistics. Then all of the queries are executed 1-99 for TPCDS 100TB
SF1. Databricks generated the TPC-DS datasets from TPC-DS kit before time started. Databricks starts time then generated all queries. Then load from S3 to Snowflake tables by - (i'm not sure about these next parts) - creating external stages and then "copy into" statements I guess? Or maybe just using copy into from an s3 bucket, that part doesnt matter much. But its not clear did they also allow target tables to be partitioned/clustering keys at all? Then all of the queries are executed 1-99 for TPCDS 100TB
Its just hard to say exactly what "They were not allowed to apply any optimizations that would require deep understanding of the dataset or queries (as done in the Snowflake pre-baked dataset, with additional clustering columns)" means exactly. Like what does that exactly mean. At a glance though, this looks very impressive for Databricks, but just want to be sure before I submit to an opinion.
Earlier quoted context omitted.
Delta lake is not meaningfully more "open" than whatever Snowflake (or BigQuery and Redshift) are doing. It does not require any less "moving data around" With all these, the data sits on cloud storage and compute is done by cloud machines - the difference between Databricks and the others is that with Databricks, you can take a look at that bucket. But you're not going to be able to do much with that data without pa…
Delta lake is very much open. You can install delta lake and run it yourself. It's a transaction layer running over parquet files. You can go to the delta.io GitHub and install binaries yourself. Snowflake cannot be run independently of their cloud. The rest of this is some vague claims of Databricks being unreliable techbros blah blah which is just emotionally charged hot air rather than being based on anything. RE…
Earlier quoted context omitted.
Delta lake is very much open. You can install delta lake and run it yourself. It's a transaction layer running over parquet files. You can go to the delta.io GitHub and install binaries yourself. Snowflake cannot be run independently of their cloud. The rest of this is some vague claims of Databricks being unreliable techbros blah blah which is just emotionally charged hot air rather than being based on anything. RE…
Sort of. You can stop using Databricks service, and keep using Delta lake. But Databricks code is not open. Delta Lake is not equivalent to Databricks delta. The value prop is that customers, if they choose to not retain databricks service, can migrate off databricks and still use the open source version of delta lake, which again, is not as good as databricks delta.
Take all the problems you have had with data warehousing and throw them in a proprietary cloud. That is Snowflake. They are the best today. Databricks started with the cloud datalake, sitting natively on parquet and using cloud native tools, fully open. Recently they added SQL to help democratize the data in the data lake versus moving it back and forth into a proprietary data warehouse. The selling point in Databric…
Whose result can be trusted is beside the point - I actually believe both experiments were likely conducted in good faith but with incomplete context. But that’s beside the point. The point is there’s no good reason to start a benchmark war to begin with.
Earlier quoted context omitted.
Sort of. You can stop using Databricks service, and keep using Delta lake. But Databricks code is not open. Delta Lake is not equivalent to Databricks delta. The value prop is that customers, if they choose to not retain databricks service, can migrate off databricks and still use the open source version of delta lake, which again, is not as good as databricks delta.
Ok you've got me there it's not 100% the exact same code Databricks are using there are some optimisations (that normally do end up downstream anyway). But I think it's getting a bit philosophical to say it's not open when you can run a delta lake "on-prem" and shuffle data between databricks and your own setup with few/no changes. Now Databricks SQL product afaik is not open and that's a proprietary C++ engine compa…