diff --git a/docs/en/setup/backend/backend-storage.md b/docs/en/setup/backend/backend-storage.md
index dac86fbbccff..f878a3b7890a 100644
--- a/docs/en/setup/backend/backend-storage.md
+++ b/docs/en/setup/backend/backend-storage.md
@@ -7,26 +7,27 @@ storage:
   selector: ${SW_STORAGE:banyandb}
 ```
 
-Natively supported storage:
-
 ## BanyanDB - Native APM Database
 - [BanyanDB](storages/banyandb.md)
 
-It has demonstrated significant potential in performance improvement. It indicates 5x less memory usage, 
+BanyanDB is a native-built SkyWalking database, which can completely focus on SkyWalking use cases.
+It has demonstrated significant potential for performance improvement and reduced resource usage requirements. It indicates 5x less memory usage, 
 1/5 disk IOPS, 1/4 disk throughput, and 30% less disk space, albeit with a slightly higher CPU trade-off, compared to Elasticsearch.
-It is designed and implemented natively for SkyWalking use cases.
 
 ## SQL database
 - [MySQL and its compatible databases](storages/mysql.md)
 - [PostgreSQL and its compatible databases](storages/postgresql.md)
 
-MySQL and PostgreSQL are recommended for production environments for medium scale deployments, especially for low trace
-and log sampling rate. Some of their compatible databases may support larger scale better, such as TiDB and AWS Aurora.
+MySQL and PostgreSQL are recommended for production environments for medium-scale deployments, especially for low trace
+and log sampling rates. Some of their compatible databases may support larger scale better, such as TiDB and AWS Aurora. 
+But the logs and traces performance could be significantly lower than BanyanDB and Elasticsearch, even more, their performance can't be improved linearly
+through scaling out nodes.
 
 ## Elasticsearch+
 
 - [OpenSearch](storages/elasticsearch.md#opensearch)
 - [ElasticSearch 7 and 8](storages/elasticsearch.md#elasticsearch)
 
-Elasticsearch and OpenSearch are recommended for production environments, specially for large scale deployments.
-OpenSearch derived from Elasticsearch 7.10.2 and iterate by its own roadmap.
+Elasticsearch and OpenSearch are recommended for production environments, especially for large-scale deployments.
+OpenSearch derived from Elasticsearch 7.10.2 and iterates by its own roadmap. But notice, Elasticsearch cluster resource costs are high,
+due to its high requirement for memory and replication requirement to keep cluster robustness.