diff --git a/setup.cfg b/setup.cfg
index 7dc6122e1..e21534b49 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -54,6 +54,7 @@ watcher_goals =
     workload_balancing = watcher.decision_engine.goal.goals:WorkloadBalancing
     airflow_optimization = watcher.decision_engine.goal.goals:AirflowOptimization
     noisy_neighbor = watcher.decision_engine.goal.goals:NoisyNeighborOptimization
+    saving_energy = watcher.decision_engine.goal.goals:SavingEnergy
 
 watcher_scoring_engines =
     dummy_scorer = watcher.decision_engine.scoring.dummy_scorer:DummyScorer
@@ -67,6 +68,7 @@ watcher_strategies =
     dummy_with_resize = watcher.decision_engine.strategy.strategies.dummy_with_resize:DummyWithResize
     basic = watcher.decision_engine.strategy.strategies.basic_consolidation:BasicConsolidation
     outlet_temperature = watcher.decision_engine.strategy.strategies.outlet_temp_control:OutletTempControl
+    saving_energy =  watcher.decision_engine.strategy.strategies.saving_energy:SavingEnergy
     vm_workload_consolidation = watcher.decision_engine.strategy.strategies.vm_workload_consolidation:VMWorkloadConsolidation
     workload_stabilization = watcher.decision_engine.strategy.strategies.workload_stabilization:WorkloadStabilization
     workload_balance = watcher.decision_engine.strategy.strategies.workload_balance:WorkloadBalance
diff --git a/watcher/decision_engine/goal/__init__.py b/watcher/decision_engine/goal/__init__.py
index 160788427..1cde9744c 100644
--- a/watcher/decision_engine/goal/__init__.py
+++ b/watcher/decision_engine/goal/__init__.py
@@ -22,7 +22,8 @@ ThermalOptimization = goals.ThermalOptimization
 Unclassified = goals.Unclassified
 WorkloadBalancing = goals.WorkloadBalancing
 NoisyNeighbor = goals.NoisyNeighborOptimization
+SavingEnergy = goals.SavingEnergy
 
 __all__ = ("Dummy", "ServerConsolidation", "ThermalOptimization",
            "Unclassified", "WorkloadBalancing",
-           "NoisyNeighborOptimization",)
+           "NoisyNeighborOptimization", "SavingEnergy")
diff --git a/watcher/decision_engine/goal/goals.py b/watcher/decision_engine/goal/goals.py
index e5be78fc4..58909dde5 100644
--- a/watcher/decision_engine/goal/goals.py
+++ b/watcher/decision_engine/goal/goals.py
@@ -192,3 +192,27 @@ class NoisyNeighborOptimization(base.Goal):
     def get_efficacy_specification(cls):
         """The efficacy spec for the current goal"""
         return specs.Unclassified()
+
+
+class SavingEnergy(base.Goal):
+    """SavingEnergy
+
+    This goal is used to reduce power consumption within a data center.
+    """
+
+    @classmethod
+    def get_name(cls):
+        return "saving_energy"
+
+    @classmethod
+    def get_display_name(cls):
+        return _("Saving Energy")
+
+    @classmethod
+    def get_translatable_display_name(cls):
+        return "Saving Energy"
+
+    @classmethod
+    def get_efficacy_specification(cls):
+        """The efficacy spec for the current goal"""
+        return specs.Unclassified()
diff --git a/watcher/decision_engine/strategy/strategies/__init__.py b/watcher/decision_engine/strategy/strategies/__init__.py
index c1a28217a..afefced2c 100644
--- a/watcher/decision_engine/strategy/strategies/__init__.py
+++ b/watcher/decision_engine/strategy/strategies/__init__.py
@@ -19,6 +19,7 @@ from watcher.decision_engine.strategy.strategies import dummy_strategy
 from watcher.decision_engine.strategy.strategies import dummy_with_scorer
 from watcher.decision_engine.strategy.strategies import noisy_neighbor
 from watcher.decision_engine.strategy.strategies import outlet_temp_control
+from watcher.decision_engine.strategy.strategies import saving_energy
 from watcher.decision_engine.strategy.strategies import uniform_airflow
 from watcher.decision_engine.strategy.strategies import \
     vm_workload_consolidation
@@ -29,6 +30,7 @@ BasicConsolidation = basic_consolidation.BasicConsolidation
 OutletTempControl = outlet_temp_control.OutletTempControl
 DummyStrategy = dummy_strategy.DummyStrategy
 DummyWithScorer = dummy_with_scorer.DummyWithScorer
+SavingEnergy = saving_energy.SavingEnergy
 VMWorkloadConsolidation = vm_workload_consolidation.VMWorkloadConsolidation
 WorkloadBalance = workload_balance.WorkloadBalance
 WorkloadStabilization = workload_stabilization.WorkloadStabilization
@@ -37,4 +39,5 @@ NoisyNeighbor = noisy_neighbor.NoisyNeighbor
 
 __all__ = ("BasicConsolidation", "OutletTempControl", "DummyStrategy",
            "DummyWithScorer", "VMWorkloadConsolidation", "WorkloadBalance",
-           "WorkloadStabilization", "UniformAirflow", "NoisyNeighbor")
+           "WorkloadStabilization", "UniformAirflow", "NoisyNeighbor",
+           "SavingEnergy")
diff --git a/watcher/decision_engine/strategy/strategies/base.py b/watcher/decision_engine/strategy/strategies/base.py
index 607f98a37..406baa75b 100644
--- a/watcher/decision_engine/strategy/strategies/base.py
+++ b/watcher/decision_engine/strategy/strategies/base.py
@@ -358,3 +358,11 @@ class NoisyNeighborBaseStrategy(BaseStrategy):
     @classmethod
     def get_goal_name(cls):
         return "noisy_neighbor"
+
+
+@six.add_metaclass(abc.ABCMeta)
+class SavingEnergyBaseStrategy(BaseStrategy):
+
+    @classmethod
+    def get_goal_name(cls):
+        return "saving_energy"
diff --git a/watcher/decision_engine/strategy/strategies/saving_energy.py b/watcher/decision_engine/strategy/strategies/saving_energy.py
new file mode 100644
index 000000000..f31b90182
--- /dev/null
+++ b/watcher/decision_engine/strategy/strategies/saving_energy.py
@@ -0,0 +1,199 @@
+# -*- encoding: utf-8 -*-
+# Copyright (c) 2017 ZTE Corporation
+#
+# Authors: licanwei <li.canwei2@zte.com.cn>
+#
+# Licensed 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.
+#
+
+import random
+
+from oslo_log import log
+
+from watcher._i18n import _
+from watcher.common import exception as wexc
+from watcher.decision_engine.strategy.strategies import base
+
+LOG = log.getLogger(__name__)
+
+
+class SavingEnergy(base.SavingEnergyBaseStrategy):
+
+    def __init__(self, config, osc=None):
+
+        super(SavingEnergy, self).__init__(config, osc)
+        self._ironic_client = None
+        self._nova_client = None
+
+        self.with_vms_node_pool = []
+        self.free_poweron_node_pool = []
+        self.free_poweroff_node_pool = []
+        self.free_used_percent = 0
+        self.min_free_hosts_num = 1
+
+    @property
+    def ironic_client(self):
+        if not self._ironic_client:
+            self._ironic_client = self.osc.ironic()
+        return self._ironic_client
+
+    @property
+    def nova_client(self):
+        if not self._nova_client:
+            self._nova_client = self.osc.nova()
+        return self._nova_client
+
+    @classmethod
+    def get_name(cls):
+        return "saving_energy"
+
+    @classmethod
+    def get_display_name(cls):
+        return _("Saving Energy Strategy")
+
+    @classmethod
+    def get_translatable_display_name(cls):
+        return "Saving Energy Strategy"
+
+    @classmethod
+    def get_schema(cls):
+        """return a schema of two input parameters
+
+        The standby nodes refer to those nodes unused
+        but still poweredon to deal with boom of new instances.
+        """
+        return {
+            "properties": {
+                "free_used_percent": {
+                    "description": ("a rational number, which describes the"
+                                    "quotient of"
+                                    " min_free_hosts_num/nodes_with_VMs_num"
+                                    "where nodes_with_VMs_num is the number"
+                                    "of nodes with VMs"),
+                    "type": "number",
+                    "default": 10.0
+                },
+                "min_free_hosts_num": {
+                    "description": ("minimum number of hosts without VMs"
+                                    "but still powered on"),
+                    "type": "number",
+                    "default": 1
+                },
+            },
+        }
+
+    def add_action_poweronoff_node(self, node_uuid, state):
+        """Add an action for node disability into the solution.
+
+        :param node: node uuid
+        :param state: node power state, power on or power off
+        :return: None
+        """
+        params = {'state': state}
+        self.solution.add_action(
+            action_type='change_node_power_state',
+            resource_id=node_uuid,
+            input_parameters=params)
+
+    def get_hosts_pool(self):
+        """Get three pools, with_vms_node_pool, free_poweron_node_pool,
+
+        free_poweroff_node_pool.
+
+        """
+
+        node_list = self.ironic_client.node.list()
+        for node in node_list:
+            node_uuid = (node.to_dict())['uuid']
+            node_info = self.ironic_client.node.get(node_uuid).to_dict()
+            hypervisor_id = node_info['extra'].get('compute_node_id', None)
+            if hypervisor_id is None:
+                LOG.warning(('Cannot find compute_node_id in extra '
+                             'of ironic node %s'), node_uuid)
+                continue
+            hypervisor_node = self.nova_client.hypervisors.get(hypervisor_id)
+            if hypervisor_node is None:
+                LOG.warning(('Cannot find hypervisor %s'), hypervisor_id)
+                continue
+            hypervisor_node = hypervisor_node.to_dict()
+            compute_service = hypervisor_node.get('service', None)
+            host_uuid = compute_service.get('host')
+            if not self.compute_model.get_node_by_uuid(host_uuid):
+                continue
+
+            if not (hypervisor_node.get('state') == 'up'):
+                """filter nodes that are not in 'up' state"""
+                continue
+            else:
+                if (hypervisor_node['running_vms'] == 0):
+                    if (node_info['power_state'] == 'power on'):
+                        self.free_poweron_node_pool.append(node_uuid)
+                    elif (node_info['power_state'] == 'power off'):
+                        self.free_poweroff_node_pool.append(node_uuid)
+                else:
+                    self.with_vms_node_pool.append(node_uuid)
+
+    def save_energy(self):
+
+        need_poweron = max(
+            (len(self.with_vms_node_pool) * self.free_used_percent / 100), (
+                self.min_free_hosts_num))
+        len_poweron = len(self.free_poweron_node_pool)
+        len_poweroff = len(self.free_poweroff_node_pool)
+        if len_poweron > need_poweron:
+            for node in random.sample(self.free_poweron_node_pool,
+                                      (len_poweron - need_poweron)):
+                self.add_action_poweronoff_node(node, 'off')
+                LOG.debug("power off %s", node)
+        elif len_poweron < need_poweron:
+            diff = need_poweron - len_poweron
+            for node in random.sample(self.free_poweroff_node_pool,
+                                      min(len_poweroff, diff)):
+                self.add_action_poweronoff_node(node, 'on')
+                LOG.debug("power on %s", node)
+
+    def pre_execute(self):
+        """Pre-execution phase
+
+        This can be used to fetch some pre-requisites or data.
+        """
+        LOG.info("Initializing Saving Energy Strategy")
+
+        if not self.compute_model:
+            raise wexc.ClusterStateNotDefined()
+
+        if self.compute_model.stale:
+            raise wexc.ClusterStateStale()
+
+        LOG.debug(self.compute_model.to_string())
+
+    def do_execute(self):
+        """Strategy execution phase
+
+        This phase is where you should put the main logic of your strategy.
+        """
+        self.free_used_percent = self.input_parameters.free_used_percent
+        self.min_free_hosts_num = self.input_parameters.min_free_hosts_num
+
+        self.get_hosts_pool()
+        self.save_energy()
+
+    def post_execute(self):
+        """Post-execution phase
+
+        This can be used to compute the global efficacy
+        """
+        self.solution.model = self.compute_model
+
+        LOG.debug(self.compute_model.to_string())
diff --git a/watcher/tests/decision_engine/strategy/strategies/test_saving_energy.py b/watcher/tests/decision_engine/strategy/strategies/test_saving_energy.py
new file mode 100644
index 000000000..df74fbda3
--- /dev/null
+++ b/watcher/tests/decision_engine/strategy/strategies/test_saving_energy.py
@@ -0,0 +1,210 @@
+# -*- encoding: utf-8 -*-
+# Copyright (c) 2017 ZTE
+#
+# Licensed 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.
+#
+
+import mock
+
+from watcher.common import clients
+from watcher.common import utils
+from watcher.decision_engine.strategy import strategies
+from watcher.tests import base
+from watcher.tests.decision_engine.model import faker_cluster_and_metrics
+
+
+class TestSavingEnergy(base.TestCase):
+
+    def setUp(self):
+        super(TestSavingEnergy, self).setUp()
+
+        mock_node1 = mock.Mock()
+        mock_node2 = mock.Mock()
+        mock_node1.to_dict.return_value = {
+            'uuid': '922d4762-0bc5-4b30-9cb9-48ab644dd861'}
+        mock_node2.to_dict.return_value = {
+            'uuid': '922d4762-0bc5-4b30-9cb9-48ab644dd862'}
+        self.fake_nodes = [mock_node1, mock_node2]
+
+        # fake cluster
+        self.fake_cluster = faker_cluster_and_metrics.FakerModelCollector()
+
+        p_model = mock.patch.object(
+            strategies.SavingEnergy, "compute_model",
+            new_callable=mock.PropertyMock)
+        self.m_model = p_model.start()
+        self.addCleanup(p_model.stop)
+
+        p_ironic = mock.patch.object(
+            clients.OpenStackClients, 'ironic')
+        self.m_ironic = p_ironic.start()
+        self.addCleanup(p_ironic.stop)
+
+        p_nova = mock.patch.object(
+            clients.OpenStackClients, 'nova')
+        self.m_nova = p_nova.start()
+        self.addCleanup(p_nova.stop)
+
+        p_model = mock.patch.object(
+            strategies.SavingEnergy, "compute_model",
+            new_callable=mock.PropertyMock)
+        self.m_model = p_model.start()
+        self.addCleanup(p_model.stop)
+
+        p_audit_scope = mock.patch.object(
+            strategies.SavingEnergy, "audit_scope",
+            new_callable=mock.PropertyMock
+        )
+        self.m_audit_scope = p_audit_scope.start()
+        self.addCleanup(p_audit_scope.stop)
+
+        self.m_audit_scope.return_value = mock.Mock()
+        self.m_ironic.node.list.return_value = self.fake_nodes
+
+        self.strategy = strategies.SavingEnergy(
+            config=mock.Mock())
+        self.strategy.input_parameters = utils.Struct()
+        self.strategy.input_parameters.update(
+            {'free_used_percent': 10.0,
+             'min_free_hosts_num': 1})
+        self.strategy.free_used_percent = 10.0
+        self.strategy.min_free_hosts_num = 1
+        self.strategy._ironic_client = self.m_ironic
+        self.strategy._nova_client = self.m_nova
+
+    def test_get_hosts_pool_with_vms_node_pool(self):
+        mock_node1 = mock.Mock()
+        mock_node2 = mock.Mock()
+        mock_node1.to_dict.return_value = {
+            'extra': {'compute_node_id': 1},
+            'power_state': 'power on'}
+        mock_node2.to_dict.return_value = {
+            'extra': {'compute_node_id': 2},
+            'power_state': 'power off'}
+        self.m_ironic.node.get.side_effect = [mock_node1, mock_node2]
+
+        mock_hyper1 = mock.Mock()
+        mock_hyper2 = mock.Mock()
+        mock_hyper1.to_dict.return_value = {
+            'running_vms': 2, 'service': {'host': 'Node_0'}, 'state': 'up'}
+        mock_hyper2.to_dict.return_value = {
+            'running_vms': 2, 'service': {'host': 'Node_1'}, 'state': 'up'}
+        self.m_nova.hypervisors.get.side_effect = [mock_hyper1, mock_hyper2]
+
+        model = self.fake_cluster.generate_scenario_1()
+        self.m_model.return_value = model
+        self.strategy.get_hosts_pool()
+
+        self.assertEqual(len(self.strategy.with_vms_node_pool), 2)
+        self.assertEqual(len(self.strategy.free_poweron_node_pool), 0)
+        self.assertEqual(len(self.strategy.free_poweroff_node_pool), 0)
+
+    def test_get_hosts_pool_free_poweron_node_pool(self):
+        mock_node1 = mock.Mock()
+        mock_node2 = mock.Mock()
+        mock_node1.to_dict.return_value = {
+            'extra': {'compute_node_id': 1},
+            'power_state': 'power on'}
+        mock_node2.to_dict.return_value = {
+            'extra': {'compute_node_id': 2},
+            'power_state': 'power on'}
+        self.m_ironic.node.get.side_effect = [mock_node1, mock_node2]
+
+        mock_hyper1 = mock.Mock()
+        mock_hyper2 = mock.Mock()
+        mock_hyper1.to_dict.return_value = {
+            'running_vms': 0, 'service': {'host': 'Node_0'}, 'state': 'up'}
+        mock_hyper2.to_dict.return_value = {
+            'running_vms': 0, 'service': {'host': 'Node_1'}, 'state': 'up'}
+        self.m_nova.hypervisors.get.side_effect = [mock_hyper1, mock_hyper2]
+
+        model = self.fake_cluster.generate_scenario_1()
+        self.m_model.return_value = model
+        self.strategy.get_hosts_pool()
+
+        self.assertEqual(len(self.strategy.with_vms_node_pool), 0)
+        self.assertEqual(len(self.strategy.free_poweron_node_pool), 2)
+        self.assertEqual(len(self.strategy.free_poweroff_node_pool), 0)
+
+    def test_get_hosts_pool_free_poweroff_node_pool(self):
+        mock_node1 = mock.Mock()
+        mock_node2 = mock.Mock()
+        mock_node1.to_dict.return_value = {
+            'extra': {'compute_node_id': 1},
+            'power_state': 'power off'}
+        mock_node2.to_dict.return_value = {
+            'extra': {'compute_node_id': 2},
+            'power_state': 'power off'}
+        self.m_ironic.node.get.side_effect = [mock_node1, mock_node2]
+
+        mock_hyper1 = mock.Mock()
+        mock_hyper2 = mock.Mock()
+        mock_hyper1.to_dict.return_value = {
+            'running_vms': 0, 'service': {'host': 'Node_0'}, 'state': 'up'}
+        mock_hyper2.to_dict.return_value = {
+            'running_vms': 0, 'service': {'host': 'Node_1'}, 'state': 'up'}
+        self.m_nova.hypervisors.get.side_effect = [mock_hyper1, mock_hyper2]
+
+        model = self.fake_cluster.generate_scenario_1()
+        self.m_model.return_value = model
+        self.strategy.get_hosts_pool()
+
+        self.assertEqual(len(self.strategy.with_vms_node_pool), 0)
+        self.assertEqual(len(self.strategy.free_poweron_node_pool), 0)
+        self.assertEqual(len(self.strategy.free_poweroff_node_pool), 2)
+
+    def test_save_energy_poweron(self):
+        self.strategy.free_poweroff_node_pool = [
+            '922d4762-0bc5-4b30-9cb9-48ab644dd861',
+            '922d4762-0bc5-4b30-9cb9-48ab644dd862'
+            ]
+        self.strategy.save_energy()
+        self.assertEqual(len(self.strategy.solution.actions), 1)
+        action = self.strategy.solution.actions[0]
+        self.assertEqual(action.get('input_parameters').get('state'), 'on')
+
+    def test_save_energy_poweroff(self):
+        self.strategy.free_poweron_node_pool = [
+            '922d4762-0bc5-4b30-9cb9-48ab644dd861',
+            '922d4762-0bc5-4b30-9cb9-48ab644dd862'
+            ]
+        self.strategy.save_energy()
+        self.assertEqual(len(self.strategy.solution.actions), 1)
+        action = self.strategy.solution.actions[0]
+        self.assertEqual(action.get('input_parameters').get('state'), 'off')
+
+    def test_execute(self):
+        mock_node1 = mock.Mock()
+        mock_node2 = mock.Mock()
+        mock_node1.to_dict.return_value = {
+            'extra': {'compute_node_id': 1},
+            'power_state': 'power on'}
+        mock_node2.to_dict.return_value = {
+            'extra': {'compute_node_id': 2},
+            'power_state': 'power on'}
+        self.m_ironic.node.get.side_effect = [mock_node1, mock_node2]
+
+        mock_hyper1 = mock.Mock()
+        mock_hyper2 = mock.Mock()
+        mock_hyper1.to_dict.return_value = {
+            'running_vms': 0, 'service': {'host': 'Node_0'}, 'state': 'up'}
+        mock_hyper2.to_dict.return_value = {
+            'running_vms': 0, 'service': {'host': 'Node_1'}, 'state': 'up'}
+        self.m_nova.hypervisors.get.side_effect = [mock_hyper1, mock_hyper2]
+
+        model = self.fake_cluster.generate_scenario_1()
+        self.m_model.return_value = model
+
+        solution = self.strategy.execute()
+        self.assertEqual(len(solution.actions), 1)