diff --git a/README.md b/README.md
index a50ca066..68725af9 100755
--- a/README.md
+++ b/README.md
@@ -1,3 +1,5 @@
+
+
# Sentinel: Sentinel of Your Application
[![Travis Build Status](https://travis-ci.org/alibaba/Sentinel.svg?branch=master)](https://travis-ci.org/alibaba/Sentinel)
diff --git a/sentinel-demo/sentinel-demo-dubbo/README.md b/sentinel-demo/sentinel-demo-dubbo/README.md
index d929c996..920afa33 100644
--- a/sentinel-demo/sentinel-demo-dubbo/README.md
+++ b/sentinel-demo/sentinel-demo-dubbo/README.md
@@ -11,6 +11,7 @@ Sentinel 提供了与 Dubbo 整合的模块 - Sentinel Dubbo Adapter,主要包
```
引入此依赖后,Dubbo 的服务接口和方法(包括调用端和服务端)就会成为 Sentinel 中的资源,在配置了规则后就可以自动享受到 Sentinel 的防护能力。
+
若不希望开启 Sentinel Dubbo Adapter 中的某个 Filter,可以手动关闭对应的 Filter,比如:
```java
@@ -63,16 +64,33 @@ Demo 2 演示了此限流场景,我们看一下这种模式的效果。假设
## Sentinel Dashboard
-Sentinel 还提供 API 用于获取实时的监控信息,对应文档见[此处](https://github.com/alibaba/Sentinel/wiki/%E5%AE%9E%E6%97%B6%E7%9B%91%E6%8E%A7)。为了便于使用,Sentinel 还提供了一个控制台(Dashboard)用于配置规则、查看监控、机器发现等功能。我们只需要按照 [Sentinel 控制台文档](https://github.com/alibaba/Sentinel/wiki/%E6%8E%A7%E5%88%B6%E5%8F%B0) 启动控制台,然后给对应的应用程序添加相应参数并启动即可。比如本文中 Service Provider 示例的启动参数:
+Sentinel 还提供 API 用于获取实时的监控信息,对应文档见[此处](https://github.com/alibaba/Sentinel/wiki/%E5%AE%9E%E6%97%B6%E7%9B%91%E6%8E%A7)。为了便于使用,Sentinel 还提供了一个控制台(Dashboard)用于配置规则、查看监控、机器发现等功能。
+
+接入 Dashboard 的步骤(**缺一不可**):
+
+1. 按照 [Sentinel 控制台文档](https://github.com/alibaba/Sentinel/wiki/%E6%8E%A7%E5%88%B6%E5%8F%B0) 启动控制台
+2. 应用引入 `sentinel-transport-simple-http` 依赖,以便控制台可以拉取对应应用的相关信息
+3. 给应用添加相关的启动参数,启动应用。需要配置的参数有:
+ - `-Dcsp.sentinel.api.port`:客户端的 port,用于上报相关信息
+ - `-Dcsp.sentinel.dashboard.server`:控制台的地址
+ - `-Dproject.name`:应用名称,会在控制台中显示
+
+注意某些环境下本地运行 Dubbo 服务还需要加上 `-Djava.net.preferIPv4Stack=true` 参数。比如 Service Provider 示例的启动参数:
```bash
-Djava.net.preferIPv4Stack=true -Dcsp.sentinel.api.port=8720 -Dcsp.sentinel.dashboard.server=localhost:8080 -Dproject.name=dubbo-provider-demo
```
-这样在启动 Service Provider 示例以后,就可以在 Sentinel 控制台中找到我们的服务了。可以很方便地在控制台中配置限流规则:
+Service Consumer 示例的启动参数:
+
+```bash
+-Djava.net.preferIPv4Stack=true -Dcsp.sentinel.api.port=8721 -Dcsp.sentinel.dashboard.server=localhost:8080 -Dproject.name=dubbo-consumer-demo
+```
+
+这样在启动 Service Provider 和 Service Consumer 示例以后,就可以在 Sentinel 控制台中找到我们的服务了。可以很方便地在控制台中配置限流规则:
![规则配置](http://dubbo.incubator.apache.org/img/blog/sentinel-dashboard-view-rules.png)
或者查看实时监控数据:
-![秒级实时监控](http://dubbo.incubator.apache.org/img/blog/sentinel-dashboard-metrics.png)
\ No newline at end of file
+![秒级实时监控](http://dubbo.incubator.apache.org/img/blog/sentinel-dashboard-metrics.png)
diff --git a/sentinel-extension/README.md b/sentinel-extension/README.md
index 5522d9a4..d1f7443a 100755
--- a/sentinel-extension/README.md
+++ b/sentinel-extension/README.md
@@ -1,9 +1,6 @@
# Sentinel Extension
-This is the parent of all extension points to Sentinel.
-
-Examples of what makes sense as a extension submodule are:
-
-* alternate implementations rules config
+Sentinel extension modules provide additional extension points
+and functions.
diff --git a/sentinel-extension/sentinel-datasource-nacos/README.md b/sentinel-extension/sentinel-datasource-nacos/README.md
new file mode 100644
index 00000000..6b16fe28
--- /dev/null
+++ b/sentinel-extension/sentinel-datasource-nacos/README.md
@@ -0,0 +1,27 @@
+# Sentinel DataSource Nacos
+
+Sentinel DataSource Nacos provides integration with [Nacos](http://nacos.io) so that Nacos
+can be the dynamic rule data source of Sentinel.
+
+To use Sentinel DataSource Nacos, you should add the following dependency:
+
+```xml
+
+ com.alibaba.csp
+ sentinel-datasource-nacos
+ x.y.z
+
+```
+
+Then you can create an `NacosDataSource` and register to rule managers.
+For instance:
+
+```java
+// remoteAddress is the address of Nacos
+// groupId and dataId are concepts of Nacos
+DataSource> flowRuleDataSource = new NacosDataSource<>(remoteAddress, groupId, dataId,
+ source -> JSON.parseObject(source, new TypeReference>() {}));
+FlowRuleManager.register2Property(flowRuleDataSource.getProperty());
+```
+
+We've also provided an example: [sentinel-demo-nacos-datasource](https://github.com/alibaba/Sentinel/tree/master/sentinel-demo/sentinel-demo-nacos-datasource).
\ No newline at end of file
diff --git a/sentinel-extension/sentinel-datasource-zookeeper/README.md b/sentinel-extension/sentinel-datasource-zookeeper/README.md
new file mode 100644
index 00000000..735b8a92
--- /dev/null
+++ b/sentinel-extension/sentinel-datasource-zookeeper/README.md
@@ -0,0 +1,27 @@
+# Sentinel DataSource ZooKeeper
+
+Sentinel DataSource ZooKeeper provides integration with ZooKeeper so that ZooKeeper
+can be the dynamic rule data source of Sentinel. The data source uses push model (listener).
+
+To use Sentinel DataSource ZooKeeper, you should add the following dependency:
+
+```xml
+
+ com.alibaba.csp
+ sentinel-datasource-zookeeper
+ x.y.z
+
+```
+
+Then you can create an `ZookeeperDataSource` and register to rule managers.
+For instance:
+
+```java
+// `path` is the data path in ZooKeeper
+DataSource> flowRuleDataSource = new ZookeeperDataSource<>(remoteAddress, path, source -> JSON.parseObject(source, new TypeReference>() {}));
+FlowRuleManager.register2Property(flowRuleDataSource.getProperty());
+```
+
+> Note: It's not recommended to add a large amount of rules to a single path (has limitation, also leads to bad performance).
+
+We've also provided an example: [sentinel-demo-zookeeper-datasource](https://github.com/alibaba/Sentinel/tree/master/sentinel-demo/sentinel-demo-zookeeper-datasource).
\ No newline at end of file