diff --git a/graphx/src/main/scala/org/apache/spark/graphx/GraphOps.scala b/graphx/src/main/scala/org/apache/spark/graphx/GraphOps.scala
index 116d1ea70017..8c788950f379 100644
--- a/graphx/src/main/scala/org/apache/spark/graphx/GraphOps.scala
+++ b/graphx/src/main/scala/org/apache/spark/graphx/GraphOps.scala
@@ -336,6 +336,69 @@ class GraphOps[VD: ClassTag, ED: ClassTag](graph: Graph[VD, ED]) extends Seriali
     Pregel(graph, initialMsg, maxIterations, activeDirection)(vprog, sendMsg, mergeMsg)
   }
 
+  /**
+   * An additional functionality for [[Pregel]] of partially sending message.
+   *
+   * Execute a Pregel-like iterative vertex-parallel abstraction with current iterative number.
+   * Part of the vertexes(called `ActiveVertexes`) send messages to their neighbours
+   * in each iteration.
+   *
+   * In some cases, `ActiveVertexes` are the vertexes that their attributes do not change
+   * between the previous and current iteration, so they need not to send message.
+   * At first, user can set Int value(eg. `flag:Int`) with `vprog`'s first parameter `curIter`
+   * to vertex's attribute in function `vprog`.
+   * Then in `sendMsg`, compare the Int value (`flag`) of Vertex attribute with `curIter` of
+   * `sendMsg`'s first parameter.
+   * In this way, it can determine whether sending message in current iteration.
+   * Provide a `isTerminal` to determine end up the loop with Int value `curIter` and the number
+   * of message count number previous iterate.
+   *
+   * Provide a `isTerminal` to determine end up the loop with Int value `curIter` and the number
+   * of message count number previous iterate.
+   *
+   * @tparam A the Pregel message type
+   *
+   * @param originGraph the input graph.
+   *
+   * @param initialMsg the message each vertex will receive at the on
+   * the first iteration. default is [[None]]
+   *
+   * @param isTerminal checking whether can finish loop
+   * Parameter Int is the current iteration variable `curIter`
+   * Parameter Long is the aggregate message number of previous iteration
+   *
+   * @param tripletFields which fields should be included in the [[EdgeContext]] passed to the
+   * `sendMsg` function. If not all fields are needed, specifying this can improve performance.
+   * default is [[TripletFields.All]]
+   *
+   * @param vprog the user-defined vertex program which runs on each
+   * vertex and receives the inbound message and computes a new vertex
+   * value.  On the first iteration the vertex program is invoked on
+   * all vertices and is passed the default message.  On subsequent
+   * iterations the vertex program is only invoked on those vertices
+   * that receive messages.
+   *
+   * @param sendMsg a user supplied function that is applied to out
+   * edges of vertices that received messages in the current iteration
+   *
+   * @param mergeMsg a user supplied function that takes two incoming
+   * messages of type A and merges them into a single message of type A.
+   * ''This function must be commutative and associative and
+   * ideally the size of A should not increase.''
+   *
+   * @return the resulting graph at the end of the computation
+   */
+  def pregelOps[A: ClassTag](
+    originGraph: Graph[VD, ED],
+    initialMsg: Option[A],
+    isTerminal: (Int, Long) => Boolean,
+    tripletFields: TripletFields = TripletFields.All)
+    (vprog: (Int, VertexId, VD, A) => VD,
+      sendMsg: (Int, EdgeContext[VD, ED, A]) => Unit,
+      mergeMsg: (A, A) => A): Graph[VD, ED] = {
+    PregelOps(originGraph, initialMsg, isTerminal, tripletFields)(vprog, sendMsg, mergeMsg)
+  }
+
   /**
    * Run a dynamic version of PageRank returning a graph with vertex attributes containing the
    * PageRank and edge attributes containing the normalized edge weight.
diff --git a/graphx/src/main/scala/org/apache/spark/graphx/PregelOps.scala b/graphx/src/main/scala/org/apache/spark/graphx/PregelOps.scala
new file mode 100644
index 000000000000..f164d23ec58b
--- /dev/null
+++ b/graphx/src/main/scala/org/apache/spark/graphx/PregelOps.scala
@@ -0,0 +1,199 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.graphx
+
+import org.apache.spark.Logging
+
+import scala.reflect.ClassTag
+
+/**
+ * Contains additional functionality for [[Pregel]] of partially sending message.
+ *
+ * Execute a Pregel-like iterative vertex-parallel abstraction with current iterative number.
+ * Part of the vertexes(called `ActiveVertexes`) send messages to their neighbours
+ * in each iteration.
+ *
+ * In some cases, `ActiveVertexes` are the vertexes that their attributes do not change
+ * between the previous and current iteration, so they need not to send message.
+ * At first, user can set Int value(eg. `flag:Int`) with `vprog`'s first parameter `curIter`
+ * to vertex's attribute in function `vprog`.
+ * Then in `sendMsg`, compare the Int value (`flag`) of Vertex attribute with `curIter` of
+ * `sendMsg`'s first parameter.
+ * In this way, it can determine whether sending message in current iteration.
+ *
+ * @example sample:
+ * {{{
+ *
+ *   // invoke
+ *  PregelOps[(Int, Int), Int, Map[Int, Int]](graph, isTerminal = isTerminal)(
+ *    vprog, sendMessage, mergeMessage)
+ *
+ *  //  set a `flag:Int` value of vertex attribute object
+ *  def vprog(curIter: Int, vid: VertexId, attr: (Int, Int),
+ *   messages: Map[Int, Int]): (Int, Int) = {
+ *   if (attr > 1024) {
+ *   //  logic code...
+ *   //  assign the curIter, the vertex can send message to its neighbors in sendMsg
+ *     (curIter, xxxx)
+ *   } else {
+ *     (0, xxxx)
+ *   }
+ *  }
+ *
+ *  def sendMessage(curIter: Int,
+ *    ctx: EdgeContext[(Int, Int), Int, Map[Int, Int]]): Unit = {
+ *    if (curIter == 0) {
+ *     ctx.sendToDst(Map(ctx.srcAttr._2 -> -1, ctx.srcAttr.xx -> 1))
+ *     ctx.sendToSrc(Map(ctx.dstAttr._2 -> -1, ctx.dstAttr.xx -> 1))
+ *    } else if (curIter == ctx.srcAttr._1) {
+ *     //  determine whether sending message
+ *     ctx.sendToDst(Map(ctx.srcAttr.preKCore -> -1, ctx.srcAttr.curKCore -> 1))
+ *     ctx.sendToSrc(Map(ctx.dstAttr.preKCore -> -1, ctx.dstAttr.curKCore -> 1))
+ *    }
+ *   }
+ *
+ *  def isTerminal(curIter: Int, messageCount: Long): Boolean = {
+ *   if (messageCount < 10 || curIter > 1000)  false else  true
+ *  }
+ *
+ *  // mergeMessage
+ *  def mergeMessage(source: Map[Int, Int], target: Map[Int, Int]): Map[Int, Int] = {
+ *   //  logic code...
+ *
+ *   target
+ *  }
+ *
+ * }}}
+ *
+ */
+object PregelOps extends Logging {
+
+  /**
+   * Implementing Part of the vertexes(we call them ActiveVertexes) send messages to their
+   * neighbours in each iteration.
+   *
+   * Provide a `isTerminal` to determine end up the loop with Int value `curIter` and the number
+   * of message count number previous iterate.
+   *
+   * @tparam VD the vertex data type
+   * @tparam ED the edge data type
+   * @tparam A the Pregel message type
+   *
+   * @param originGraph the input graph.
+   *
+   * @param initialMsg the message each vertex will receive at the on
+   * the first iteration. default is [[None]]
+   *
+   * @param isTerminal checking whether can finish loop
+   * Parameter Int is the current iteration variable `curIter`
+   * Parameter Long is the aggregate message number of previous iteration
+   *
+   * @param tripletFields which fields should be included in the [[EdgeContext]] passed to the
+   * `sendMsg` function. If not all fields are needed, specifying this can improve performance.
+   * default is [[TripletFields.All]]
+   *
+   * @param vprog the user-defined vertex program which runs on each
+   * vertex and receives the inbound message and computes a new vertex
+   * value.  On the first iteration the vertex program is invoked on
+   * all vertices and is passed the default message.  On subsequent
+   * iterations the vertex program is only invoked on those vertices
+   * that receive messages.
+   *
+   * @param sendMsg a user supplied function that is applied to out
+   * edges of vertices that received messages in the current iteration
+   *
+   * @param mergeMsg a user supplied function that takes two incoming
+   * messages of type A and merges them into a single message of type A.
+   * ''This function must be commutative and associative and
+   * ideally the size of A should not increase.''
+   *
+   * @return the resulting graph at the end of the computation
+   */
+  def apply[VD: ClassTag, ED: ClassTag, A: ClassTag]
+  (originGraph: Graph[VD, ED],
+    initialMsg: Option[A] = None,
+    isTerminal: (Int, Long) => Boolean = defaultTerminal,
+    tripletFields: TripletFields)
+    (vprog: (Int, VertexId, VD, A) => VD,
+      sendMsg: (Int, EdgeContext[VD, ED, A]) => Unit,
+      mergeMsg: (A, A) => A): Graph[VD, ED] = {
+
+    //  init iterate 0
+    val initIter = 0
+    var graph = initialMsg match {
+      case None => originGraph.cache()
+      case _ => originGraph.mapVertices((vid, vdata) => vprog(initIter, vid, vdata,
+        initialMsg.get)).cache()
+    }
+
+    // compute the messages
+    var messageRDD = graph.aggregateMessages(sendMsg(initIter, _: EdgeContext[VD, ED, A]), mergeMsg)
+    var activeMsgCount = messageRDD.count()
+
+    // Loop, from i = 1
+    var i = 1
+    while (activeMsgCount > 0 && isTerminal(i, activeMsgCount)) {
+      val ct = System.currentTimeMillis()
+      val curIter = i
+
+      // 1. Receive the messages. Vertices that didn't get any messages do not appear in newVerts.
+      val newVerts = graph.vertices.innerJoin(messageRDD)(
+        vprog(curIter, _: VertexId, _: VD, _: A)).cache()
+
+      // 2. Update the graph with the new vertices.
+      val preGraph: Graph[VD, ED] = graph
+      graph = graph.outerJoinVertices(newVerts) { (vid, old, newOpt) => newOpt.getOrElse(old)}
+      graph.cache()
+
+      val oldMessages = messageRDD
+      // 3. aggregate message
+      // Send new messages. Vertices that didn't get any messages don't appear in newVerts, so don't
+      // get to send messages. We must cache messages so it can be materialized on the next line,
+      // allowing us to uncache the previous iteration.
+      messageRDD = graph.aggregateMessages(sendMsg(curIter, _: EdgeContext[VD, ED, A]), mergeMsg,
+        tripletFields).cache()
+      // The call to count() materializes `messages`, `newVerts`, and the vertices of `g`. This
+      // hides oldMessages (depended on by newVerts), newVerts (depended on by messages), and the
+      // vertices of prevG (depended on by newVerts, oldMessages, and the vertices of g).
+      activeMsgCount = messageRDD.count()
+
+      // Unpersist the RDDs hidden by newly-materialized RDDs
+      oldMessages.unpersist(blocking = false)
+      newVerts.unpersist(blocking = false)
+      preGraph.unpersistVertices(blocking = false)
+      preGraph.edges.unpersist(blocking = false)
+      if (i == 1) {
+        originGraph.unpersistVertices(blocking = false)
+        originGraph.edges.unpersist(blocking = false)
+      }
+
+      i += 1
+
+      logInfo("{\"name\":\"pregel\", \"iterate\":" + i + ",\"cost\":"
+        + (System.currentTimeMillis() - ct) + "}")
+    }
+
+    graph
+  } // end of apply
+
+  /**
+   * default terminal function
+   * @return
+   */
+  private def defaultTerminal(curIter: Int, msgCount: Long): Boolean = true
+}