Update README.md

buremba · Aug 5, 2024 · c297acb · c297acb
1 parent 56d3cd3
commit c297acb
Showing 1 changed file with 8 additions and 6 deletions.
diff --git a/README.md b/README.md
@@ -74,7 +74,7 @@ For Docker:
 docker run buremba/universql snowflake --account eq06461.eu-west-2.aws
 ```
 
-# CLI Usage
+# CLI
 
 ```
 > universql snowflake --help
@@ -116,7 +116,7 @@ Options:
 
 [![Interactive CLI](resources/how_to_use.png)](https://benchmark.clickhouse.com/#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)
 
-## Access to Storage
+# Access to Storage
 
 ### Polaris
 
@@ -155,16 +155,18 @@ By default, UniverSQL uses [your default Azure tenant](https://learn.microsoft.c
 
 # Limitations
 
-UniverSQL is not a replacement for Snowflake, instead it's a complementary tool that allows you to run queries locally on your computer.
+## Be aware of trade-offs
+
+UniverSQL is designed to be a complementary tool that helps you utilize local hardware when using Snowflake, for small datasets that you're working on locally.
 
 ### Cost
 
 The virtual warehouse concept is great for running large queries on large datasets but usually X-Small works OK for running ad-hoc queries on small datasets that has < 2B rows.
 X-Small warehouse costs $2/hour and is [likely](https://stackoverflow.com/questions/58973007/what-are-the-specifications-of-a-snowflake-server) using [m5.2xlarge](https://instances.vantage.sh/aws/ec2/m5.2xlarge).
-If your query runs on X-Small warehouse, it will to run on your computer if you have 8GB - 32GB memory.
+If your query runs on X-Small warehouse, it will likely run to run your computer if you have 8GB - 32GB memory.
 
 If you're the one executing the Snowflake queries and your computer has enough resources, the compute is free for you. 
-You will only need to pay for the egress network costs from your cloud provider but it's going to be much cheaper than running a warehouse on Snowflake.
+You will only **need to pay for the egress network costs** from your cloud provider but it's going to be much cheaper than running a warehouse on Snowflake.
 
 ### Performance
 
@@ -174,7 +176,7 @@ UniverSQL uses DuckDB as the local compute engine, which is a columnar database
 
 ### Latency
 
-Since UniverSQL runs the queries locally, the latency relies on your network bandwidth. Your local disk is used for caching the data in data lake and the cache is populated lazily as you query the data and persisted across multiple runs.
+The latency relies on your network bandwidth. Your local disk is used for caching the data in data lake and the cache is populated lazily as you query the data and persisted across multiple runs.
 Cold runs will likely to be slower than running the query on Snowflake as the data needs to be downloaded from the data lake with UniverSQL whereas Snowflake runs the compute in the same cloud region.
 
 The subsequent queries (hot run) on the same table will be served from the cache. If the Iceberg table is updated, only the new data is downloaded from the data lake.