diff --git a/docs/mrs/umn/mrs_03_1133.html b/docs/mrs/umn/mrs_03_1133.html
index 4fbab04dd..d4fdda1ef 100644
--- a/docs/mrs/umn/mrs_03_1133.html
+++ b/docs/mrs/umn/mrs_03_1133.html
@@ -1,7 +1,7 @@
 <a name="mrs_03_1133"></a><a name="mrs_03_1133"></a>
 
 <h1 class="topictitle1">What Are the Solutions for processing 1 Billion Data Records?</h1>
-<div id="body1591271912871"><ul id="mrs_03_1133__ul364661692317"><li id="mrs_03_1133__li106816182266">GaussDB (for MySQL) is recommended for scenarios, such as data updates, online transaction processing (OLTP), and complex analysis of 1 billion data records.</li><li id="mrs_03_1133__li201102032313">Impala and Kudu in MRS also meet this requirement. Impala and Kudu can load all join tables to the memory in the join operation.</li></ul>
+<div id="body1591271912871"><ul id="mrs_03_1133__ul364661692317"><li id="mrs_03_1133__li106816182266">TaurusDB is recommended for scenarios, such as data updates, online transaction processing (OLTP), and complex analysis of 1 billion data records.</li><li id="mrs_03_1133__li201102032313">Impala and Kudu in MRS also meet this requirement. Impala and Kudu can load all join tables to the memory in the join operation.</li></ul>
 </div>
 <div>
 <div class="familylinks">