diff --git a/scheduler/scheduler.cc b/scheduler/scheduler.cc
new file mode 100644
index 0000000..e4f93e4
--- /dev/null
+++ b/scheduler/scheduler.cc
@@ -0,0 +1,184 @@
+/* -*- Mode: C++; c-file-style: "gnu"; indent-tabs-mode:nil -*- */
+/*
+ * Copyright (c) 2013 University of California, Los Angeles
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation;
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ * Author: Zhenkai Zhu <zhenkai@cs.ucla.edu>
+ *         Alexander Afanasyev <alexander.afanasyev@ucla.edu>
+ */
+
+#include "scheduler.h"
+#include <utility>
+
+using namespace std;
+
+#define EVLOOP_NO_EXIT_ON_EMPTY 0x04
+
+// IntervalGeneratorPtr
+// IntervalGenerator:: Null;
+
+void errorCallback(int err)
+{
+  cout << "Fatal error: " << err << endl;
+}
+
+Scheduler::Scheduler()
+  : m_running(false)
+{
+  event_set_fatal_callback(errorCallback);
+  evthread_use_pthreads();
+  m_base = event_base_new();
+}
+
+Scheduler::~Scheduler()
+{
+  event_base_free(m_base);
+}
+
+void
+Scheduler::eventLoop()
+{
+  while(true)
+  {
+    if (event_base_loop(m_base, EVLOOP_NO_EXIT_ON_EMPTY) < 0)
+    {
+      cout << "scheduler loop break error" << endl;
+    }
+    ReadLock lock(m_mutex);
+    if (!m_running)
+    {
+      cout << "scheduler loop break normal" << endl;
+      break;
+    }
+  }
+}
+
+void
+Scheduler::start()
+{
+  WriteLock lock(m_mutex);
+  if (!m_running)
+  {
+    m_thread = boost::thread(&Scheduler::eventLoop, this);
+    m_running = true;
+  }
+}
+
+void
+Scheduler::shutdown()
+{
+  WriteLock lock(m_mutex);
+  if (m_running)
+  {
+    event_base_loopbreak(m_base);
+    m_thread.join();
+    m_running = false;
+  }
+}
+
+bool
+Scheduler::addTask(const TaskPtr &task)
+{
+  TaskPtr newTask = task;
+
+  if (addToMap(newTask))
+  {
+    newTask->reset();
+    int res = evtimer_add(newTask->ev(), newTask->tv());
+    if (res < 0)
+    {
+      cout << "evtimer_add failed for " << task->tag() << endl;
+    }
+    return true;
+  }
+  else
+  {
+    cout << "fail to add task: " << task->tag() << endl;
+  }
+
+  return false;
+}
+
+void
+Scheduler::rescheduleTask(const Task::Tag &tag)
+{
+  ReadLock lock(m_mutex);
+  TaskMapIt it = m_taskMap.find(tag);
+  if (it != m_taskMap.end())
+  {
+    TaskPtr task = it->second;
+    task->reset();
+    int res = evtimer_add(task->ev(), task->tv());
+    if (res < 0)
+    {
+      cout << "evtimer_add failed for " << task->tag() << endl;
+    }
+  }
+}
+
+bool
+Scheduler::addToMap(const TaskPtr &task)
+{
+  WriteLock lock(m_mutex);
+  if (m_taskMap.find(task->tag()) == m_taskMap.end())
+  {
+    m_taskMap.insert(make_pair(task->tag(), task));
+    return true;
+  }
+  return false;
+}
+
+void
+Scheduler::deleteTask(const Task::Tag &tag)
+{
+  WriteLock lock(m_mutex);
+  TaskMapIt it = m_taskMap.find(tag);
+  if (it != m_taskMap.end())
+  {
+    TaskPtr task = it->second;
+    evtimer_del(task->ev());
+    m_taskMap.erase(it);
+  }
+}
+
+void
+Scheduler::deleteTask(const Task::TaskMatcher &matcher)
+{
+  WriteLock lock(m_mutex);
+  TaskMapIt it = m_taskMap.begin();
+  while(it != m_taskMap.end())
+  {
+    TaskPtr task = it->second;
+    if (matcher(task))
+    {
+      evtimer_del(task->ev());
+      // Use post increment; map.erase invalidate the iterator that is beening erased,
+      // but does not invalidate other iterators. This seems to be the convention to
+      // erase something from C++ STL map while traversing.
+      m_taskMap.erase(it++);
+    }
+    else
+    {
+      ++it;
+    }
+  }
+}
+
+int
+Scheduler::size()
+{
+  ReadLock lock(m_mutex);
+  return m_taskMap.size();
+}