diff --git a/examples/README.md b/examples/README.md
index fec6c442..355cc12f 100644
--- a/examples/README.md
+++ b/examples/README.md
@@ -20,6 +20,7 @@ Example: [Using Federated Learning Job in Surface Defect Detection Scenario](./f
 ### Shared Storage
 | Support Protocols  |Support Features| Examples |Release|
 | :-------------: | :-------------: |:-------------: | :-------------: |
-| S3 |Joint Inference/Incremental Learning| [Using Incremental Learning Job in Helmet Detection Scenario on S3](./storage/s3/README.md) | 0.2.0+|
+| S3 |Incremental Learning| [Using Incremental Learning Job in Helmet Detection Scenario on S3](storage/s3/incremental_learning/README.md) | 0.3.1+|
+| S3 |Joint Inference| [Using Joint Inference Service in Helmet Detection Scenario on S3](storage/s3/joint_inference/README.md) | 0.3.0+|
 
 
diff --git a/examples/storage/s3/README.md b/examples/storage/s3/incremental_learning/README.md
similarity index 58%
rename from examples/storage/s3/README.md
rename to examples/storage/s3/incremental_learning/README.md
index 324a3fbd..cb48b164 100644
--- a/examples/storage/s3/README.md
+++ b/examples/storage/s3/incremental_learning/README.md
@@ -1,10 +1,15 @@
 # Using Incremental Learning Job in Helmet Detection Scenario on S3
 
-This example based on the example: [Using Incremental Learning Job in Helmet Detection Scenario](/examples/incremental_learning/helmet_detection/README.md)
+This example is based on the example: [Using Incremental Learning Job in Helmet Detection Scenario](/examples/incremental_learning/helmet_detection/README.md).    
 
-### Create a secret with your S3 user credential.
+### Prepare Nodes
+Assume you have created a [KubeEdge](https://github.com/kubeedge/kubeedge) cluster that have two cloud nodes(e.g., `cloud-node1`, `cloud-node2`)
+and one edge node(e.g., `edge-node`).  
 
-```yaml
+### Create a secret with your S3 user credential.   
+
+```shell
+kubectl create -f - <<EOF
 apiVersion: v1
 kind: Secret
 metadata:
@@ -15,19 +20,15 @@ metadata:
 stringData: # use `stringData` for raw credential string or `data` for base64 encoded string
   ACCESS_KEY_ID: XXXX
   SECRET_ACCESS_KEY: XXXXXXXX
+EOF
 ```
 
-### Attach the created secret to the Model/Dataset/Job.
-`EDGE_NODE` and `CLOUD_NODE` are custom nodes, you can fill it which you actually run.  
-
-```
-EDGE_NODE="edge-node" 
-CLOUD_NODE="cloud-node"
-```
-
-* Attach the created secret to the Model.  
+### Prepare Model
+* Download [models](https://kubeedge.obs.cn-north-1.myhuaweicloud.com/examples/helmet-detection/model.tar.gz).  
+* Put the unzipped model file into the bucket of your cloud storage service.  
+* Attach the created secret to the Model and create Model.  
 
-```yaml
+```shell
 kubectl create -f - <<EOF
 apiVersion: sedna.io/v1alpha1
 kind: Model
@@ -40,7 +41,7 @@ metadata:
 EOF
 ```
 
-```yaml
+```shell
 kubectl $action -f - <<EOF
 apiVersion: sedna.io/v1alpha1
 kind: Model
@@ -50,12 +51,15 @@ spec:
   url: "s3://kubeedge/model/deploy_model/saved_model.pb"
   format: "pb"
   credentialName: mysecret
-EO
+EOF
 ```
 
-* Attach the created secret to the Dataset.  
+### Prepare Dataset
+* Download [dataset](https://kubeedge.obs.cn-north-1.myhuaweicloud.com/examples/helmet-detection/dataset.tar.gz).  
+* Put the unzipped dataset file into the bucket of your cloud storage service.  
+* Attach the created secret to Dataset and create Dataset.  
 
-```yaml
+```shell
 kubectl $action -f - <<EOF
 apiVersion: sedna.io/v1alpha1
 kind: Dataset
@@ -64,13 +68,42 @@ metadata:
 spec:
   url: "s3://kubeedge/data/helmet_detection/train_data/train_data.txt"
   format: "txt"
-  nodeName: $CLOUD_NODE
+  nodeName: cloud-node1
   credentialName: mysecret
 EOF
+```
 
+### Prepare Image
+This example uses the image:  
+
+```shell
+kubeedge/sedna-example-incremental-learning-helmet-detection:v0.3.1
 ```
-* Attach the created secret to the Job(IncrementalLearningJob).
-```yaml
+
+This image is generated by the script [build_images.sh](/examples/build_image.sh), used for creating training, eval and inference worker.  
+
+### Prepare Job
+* Inference/Train/Eval worker now can be deployed by nodeName and [nodeSelector](https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#nodeselector) on multiple nodes.  
+
+* Make sure to follow the local dir which exists in edge side.    
+
+```shell
+mkdir -p /incremental_learning/video/
+```
+
+* Download [video](https://kubeedge.obs.cn-north-1.myhuaweicloud.com/examples/helmet-detection/video.tar.gz), unzip video.tar.gz, and put it into `/incremental_learning/video/`.    
+
+```
+cd /incremental_learning/video/
+wget https://kubeedge.obs.cn-north-1.myhuaweicloud.com/examples/helmet-detection/video.tar.gz
+tar -zxvf video.tar.gz
+```
+
+* Attach the created secret to the Job and create Job.  
+
+```shell
+IMAGE=kubeedge/sedna-example-incremental-learning-helmet-detection:v0.3.1  
+
 kubectl create -f - <<EOF
 apiVersion: sedna.io/v1alpha1
 kind: IncrementalLearningJob
@@ -85,9 +118,9 @@ spec:
   trainSpec:
     template:
       spec:
-        nodeName: $CLOUD_NODE
+        nodeName: cloud-node1
         containers:
-          - image: kubeedge/sedna-example-incremental-learning-helmet-detection:v0.1.0
+          - image: $IMAGE
             name:  train-worker
             imagePullPolicy: IfNotPresent
             args: ["train.py"]
@@ -116,9 +149,9 @@ spec:
   evalSpec:
     template:
       spec:
-        nodeName: $CLOUD_NODE
+        nodeName: cloud-node2
         containers:
-          - image: kubeedge/sedna-example-incremental-learning-helmet-detection:v0.1.0
+          - image: $IMAGE
             name:  eval-worker
             imagePullPolicy: IfNotPresent
             args: ["eval.py"]
@@ -144,9 +177,9 @@ spec:
           value: "0.9"
     template:
       spec:
-        nodeName: $EDGE_NODE
+        nodeName: edge-node
         containers:
-          - image: kubeedge/sedna-example-incremental-learning-helmet-detection:v0.1.0
+          - image: $IMAGE
             name:  infer-worker
             imagePullPolicy: IfNotPresent
             args: ["inference.py"]
@@ -169,11 +202,12 @@ spec:
           - name: localvideo
             hostPath:
               path: /incremental_learning/video/
-              type: Directory
+              type: DirectoryOrCreate
           - name: hedir
             hostPath:
               path:  /incremental_learning/he/
-              type: Directory
-  outputDir: "/incremental_learning/output"
+              type: DirectoryOrCreate
+  credentialName: mysecret
+  outputDir: "s3://kubeedge/incremental_learning/output"
 EOF
 ```
diff --git a/examples/storage/s3/joint_inference/README.md b/examples/storage/s3/joint_inference/README.md
new file mode 100644
index 00000000..8fda1602
--- /dev/null
+++ b/examples/storage/s3/joint_inference/README.md
@@ -0,0 +1,149 @@
+# Using Joint Inference Service in Helmet Detection Scenario on S3
+
+This example is based on the example: [Using Joint Inference Service in Helmet Detection Scenario](/examples/joint_inference/helmet_detection_inference/README.md).  
+
+### Prepare Nodes
+Assume you have created a [KubeEdge](https://github.com/kubeedge/kubeedge) cluster that have one cloud node(e.g., `cloud-node`)
+and one edge node(e.g., `edge-node`).  
+
+### Create a secret with your S3 user credential.        
+
+```shell
+kubectl create -f - <<EOF
+apiVersion: v1
+kind: Secret
+metadata:
+  name: mysecret
+  annotations:
+    s3-endpoint: s3.amazonaws.com # replace with your s3 endpoint e.g minio-service.kubeflow:9000 
+    s3-usehttps: "1" # by default 1, if testing with minio you can set to 0
+stringData: # use `stringData` for raw credential string or `data` for base64 encoded string
+  ACCESS_KEY_ID: XXXX
+  SECRET_ACCESS_KEY: XXXXXXXX
+EOF
+```
+
+### Prepare Model  
+* Download [little model](https://kubeedge.obs.cn-north-1.myhuaweicloud.com/examples/helmet-detection-inference/little-model.tar.gz)
+  and [big model](https://kubeedge.obs.cn-north-1.myhuaweicloud.com/examples/helmet-detection-inference/big-model.tar.gz).  
+* Put the unzipped model file into the bucket of your cloud storage service.  
+* Attach the created secret to the Model and create Model.  
+
+```shell
+kubectl create -f - <<EOF
+apiVersion: sedna.io/v1alpha1
+kind: Model
+metadata:
+  name: big-model
+  spec:
+    url : "s3://kubeedge/model/big-model/yolov3_darknet.pb"
+    format: "pb"
+    credentialName: mysecret
+EOF
+```
+
+```shell
+kubectl $action -f - <<EOF
+apiVersion: sedna.io/v1alpha1
+kind: Model
+metadata:
+  name: little-model
+spec:
+  url: "s3://kubeedge/model/little-model/yolov3_resnet18.pb"
+  format: "pb"
+  credentialName: mysecret
+EOF
+```
+
+### Prepare Images
+This example uses these images:  
+1. little model inference worker: ```kubeedge/sedna-example-joint-inference-helmet-detection-little:v0.3.0```
+2. big model inference worker: ```kubeedge/sedna-example-joint-inference-helmet-detection-big:v0.3.0```
+
+These images are generated by the script [build_images.sh](/examples/build_image.sh).  
+
+### Prepare Job 
+* Make preparation in edge node  
+
+```
+mkdir -p /joint_inference/output
+```
+
+* Attach the created secret to the Job and create Job.    
+
+```shell
+LITTLE_MODEL_IMAGE=kubeedge/sedna-example-joint-inference-helmet-detection-little:v0.3.0
+BIG_MODEL_IMAGE=kubeedge/sedna-example-joint-inference-helmet-detection-big:v0.3.0
+
+kubectl create -f - <<EOF
+apiVersion: sedna.io/v1alpha1
+kind: JointInferenceService
+metadata:
+  name: helmet-detection-inference-example
+  namespace: default
+spec:
+  edgeWorker:
+    model:
+      name: "helmet-detection-inference-little-model"
+    hardExampleMining:
+      name: "IBT"
+      parameters:
+        - key: "threshold_img"
+          value: "0.9"
+        - key: "threshold_box"
+          value: "0.9"
+    template:
+      spec:
+        nodeName: edge-node
+        containers:
+          - image: $LITTLE_MODEL_IMAGE
+            imagePullPolicy: IfNotPresent
+            name:  little-model
+            env:  # user defined environments
+              - name: input_shape
+                value: "416,736"
+              - name: "video_url"
+                value: "rtsp://localhost/video"
+              - name: "all_examples_inference_output"
+                value: "/data/output"
+              - name: "hard_example_cloud_inference_output"
+                value: "/data/hard_example_cloud_inference_output"
+              - name: "hard_example_edge_inference_output"
+                value: "/data/hard_example_edge_inference_output"
+            resources:  # user defined resources
+              requests:
+                memory: 64M
+                cpu: 100m
+              limits:
+                memory: 2Gi
+            volumeMounts:
+              - name: outputdir
+                mountPath: /data/
+        volumes:   # user defined volumes
+          - name: outputdir
+            hostPath:
+              # user must create the directory in host
+              path: /joint_inference/output
+              type: DirectoryorCreate
+
+  cloudWorker:
+    model:
+      name: "helmet-detection-inference-big-model"
+    template:
+      spec:
+        nodeName: cloud-node
+        containers:
+          - image: $BIG_MODEL_IMAGE
+            name:  big-model
+            imagePullPolicy: IfNotPresent
+            env:  # user defined environments
+              - name: "input_shape"
+                value: "544,544"
+            resources:  # user defined resources
+              requests:
+                memory: 2Gi
+EOF
+```
+
+### Mock Video Stream for Inference in Edge Side
+Refer to [here](https://github.com/kubeedge/sedna/tree/main/examples/joint_inference/helmet_detection_inference#mock-video-stream-for-inference-in-edge-side).