1 .. This work is licensed under a Creative Commons Attribution 4.0 International License.
2 .. http://creativecommons.org/licenses/by/4.0
4 Gap analysis in upstream projects
5 =================================
7 This section presents the findings of gaps on existing VIM platforms. The focus
8 was to identify gaps based on the features and requirements specified in Section
9 3.3. The analysis work determined gaps that are presented here.
11 VIM Northbound Interface
12 ------------------------
14 Immediate Notification
15 ^^^^^^^^^^^^^^^^^^^^^^
17 * Type: 'deficiency in performance'
22 - VIM has to notify unavailability of virtual resource (fault) to VIM user
24 - Notification should be passed in '1 second' after fault detected/notified
26 - Also, the following conditions/requirement have to be met:
28 - Only the owning user can receive notification of fault related to owned
33 - OpenStack Metering 'Ceilometer' can notify unavailability of virtual
34 resource (fault) to the owner of virtual resource based on alarm
35 configuration by the user.
37 - Ceilometer Alarm API:
38 http://docs.openstack.org/developer/ceilometer/webapi/v2.html#alarms
40 - Alarm notifications are triggered by alarm evaluator instead of
41 notification agents that might receive faults
43 - Ceilometer Architecture:
44 http://docs.openstack.org/developer/ceilometer/architecture.html#id1
46 - Evaluation interval should be equal to or larger than configured pipeline
47 interval for collection of underlying metrics.
49 - https://github.com/openstack/ceilometer/blob/stable/juno/ceilometer/alarm/service.py#L38-42
51 - The interval for collection has to be set large enough which depends on
52 the size of the deployment and the number of metrics to be collected.
53 - The interval may not be less than one second in even small deployments.
54 The default value is 60 seconds.
55 - Alternative: OpenStack has a message bus to publish system events.
56 The operator can allow the user to connect this, but there are no
57 functions to filter out other events that should not be passed to the user
58 or which were not requested by the user.
62 - Fault notifications cannot be received immediately by Ceilometer.
64 Maintenance Notification
65 ^^^^^^^^^^^^^^^^^^^^^^^^
72 - VIM has to notify unavailability of virtual resource triggered by NFVI
73 maintenance to VIM user.
74 - Also, the following conditions/requirements have to be met:
76 - VIM should accept maintenance message from administrator and mark target
77 physical resource "in maintenance".
78 - Only the owner of virtual resource hosted by target physical resource
79 can receive the notification that can trigger some process for
80 applications which are running on the virtual resource (e.g. cut off
86 - AWS (just for study)
88 - AWS provides API and CLI to view status of resource (VM) and to create
89 instance status and system status alarms to notify you when an instance
90 has a failed status check.
91 http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/monitoring-instances-status-check_sched.html
92 - AWS provides API and CLI to view scheduled events, such as a reboot or
93 retirement, for your instances. Also, those events will be notified
95 http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/monitoring-system-instance-status-check.html
99 - VIM user cannot receive maintenance notifications.
103 + https://blueprints.launchpad.net/nova/+spec/service-status-notification
105 VIM Southbound interface
106 ------------------------
108 Normalization of data collection models
109 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
116 - A normalized data format needs to be created to cope with the many data
117 models from different monitoring solutions.
121 - Data can be collected from many places (e.g. Zabbix, Nagios, Cacti,
122 Zenoss). Although each solution establishes its own data models, no common
123 data abstraction models exist in OpenStack.
127 - Normalized data format does not exist.
135 OpenStack offers a telemetry service, Ceilometer, for collecting measurements of
136 the utilization of physical and virtual resources [CEIL]_. Ceilometer can
137 collect a number of metrics across multiple OpenStack components and watch for
138 variations and trigger alarms based upon the collected data.
140 Scalability of fault aggregation
141 ________________________________
143 * Type: 'scalability issue'
148 - Be able to scale to a large deployment, where thousands of monitoring
149 events per second need to be analyzed.
153 - Performance issue when scaling to medium-sized deployments.
157 - Ceilometer seems to be unsuitable for monitoring medium and large scale
162 + Usage of Zabbix for fault aggregation [ZABB]_. Zabbix can support a much
163 higher number of fault events (up to 15 thousand events per second, but
164 obviously also has some upper bound:
165 http://blog.zabbix.com/scalable-zabbix-lessons-on-hitting-9400-nvps/2615/
167 + Decentralized/hierarchical deployment with multiple instances, where one
168 instance is only responsible for a small NFVI.
170 Monitoring of hardware and software
171 ___________________________________
173 * Type: 'missing (lack of functionality)'
178 - OpenStack (as VIM) should monitor various hardware and software in NFVI to
179 handle faults on them by Ceilometer.
180 - OpenStack may have monitoring functionality in itself and can be
181 integrated with third party monitoring tools.
182 - OpenStack need to be able to detect the faults listed in the Annex.
186 - For each deployment of OpenStack, an operator has responsibility to
187 configure monitoring tools with relevant scripts or plugins in order to
188 monitor hardware and software.
189 - OpenStack Ceilometer does not monitor hardware and software to capture
194 - Ceilometer is not able to detect and handle all faults listed in the Annex.
196 * Related blueprints / workarounds
198 - Use other dedicated monitoring tools like Zabbix or Monasca
203 OpenStack Nova [NOVA]_ is a mature and widely known and used component in
204 OpenStack cloud deployments. It is the main part of an
205 "infrastructure-as-a-service" system providing a cloud computing fabric
206 controller, supporting a wide diversity of virtualization and container
209 Nova has proven throughout these past years to be highly available and
210 fault-tolerant. Featuring its own API, it also provides a compatibility API with
213 Correct states when compute host is down
214 ________________________________________
216 * Type: 'missing (lack of functionality)'
221 - There needs to be API to change VM power_State in case host has failed.
222 - There needs to be API to change nova-compute state.
223 - There could be single API to change different VM states for all VMs
224 belonging to specific host.
225 - As external system monitoring the infra calls these APIs change can be
227 - Correlation actions can be faster and automated as states are reliable.
228 - User will be able to read states from OpenStack and trust they are
233 - When a VM goes down due to a host HW, host OS or hypervisor failure,
234 nothing happens in OpenStack. The VMs of a crashed host/hypervisor are
235 reported to be live and OK through the OpenStack API.
236 - nova-compute state might change too slowly or the state is not reliable
237 if expecting also VMs to be down. This leads to ability to schedule VMs
238 to a failed host and slowness blocks evacuation.
242 - OpenStack does not change its states fast and reliably enough.
243 - There is API missing to have external system to change states and to
244 trust the states are then reliable (external system has fenced failed
246 - User cannot read all the states from OpenStack nor trust they are right.
250 + https://blueprints.launchpad.net/nova/+spec/mark-host-down
251 + https://blueprints.launchpad.net/python-novaclient/+spec/support-force-down-service
253 Evacuate VMs in Maintenance mode
254 ________________________________
261 - When maintenance mode for a compute host is set, trigger VM evacuation to
262 available compute nodes before bringing the host down for maintenance.
266 - If setting a compute node to a maintenance mode, OpenStack only schedules
267 evacuation of all VMs to available compute nodes if in-maintenance compute
268 node runs the XenAPI and VMware ESX hypervisors. Other hypervisors (e.g.
269 KVM) are not supported and, hence, guest VMs will likely stop running due
270 to maintenance actions administrator may perform (e.g. hardware upgrades,
275 - Nova libvirt hypervisor driver does not implement automatic guest VMs
276 evacuation when compute nodes are set to maintenance mode (``$ nova
277 host-update --maintenance enable <hostname>``).
282 Monasca is an open-source monitoring-as-a-service (MONaaS) solution that
283 integrates with OpenStack. Even though it is still in its early days, it is the
284 interest of the community that the platform be multi-tenant, highly scalable,
285 performant and fault-tolerant. It provides a streaming alarm engine, a
286 notification engine, and a northbound REST API users can use to interact with
287 Monasca. Hundreds of thousands of metrics per second can be processed
294 * Type: 'missing (lack of functionality)'
299 - Detect the failure and perform a root cause analysis to filter out other
300 alarms that may be triggered due to their cascading relation.
304 - A mechanism to detect root causes of failures is not available.
308 - Certain failures can trigger many alarms due to their dependency on the
309 underlying root cause of failure. Knowing the root cause can help filter
310 out unnecessary and overwhelming alarms.
312 * Related blueprints / workarounds
314 + Monasca as of now lacks this feature, although the community is aware and
315 working toward supporting it.
320 * Type: 'missing (lack of functionality)'
325 - It should support monitoring sensor data retrieval, for instance, from
330 - Monasca does not monitor sensor data
334 - Sensor monitoring is very important. It provides operators status
335 on the state of the physical infrastructure (e.g. temperature, fans).
337 * Related blueprints / workarounds
339 + Monasca can be configured to use third-party monitoring solutions (e.g.
340 Nagios, Cacti) for retrieving additional data.
342 Hardware monitoring tools
343 -------------------------
348 Zabbix is an open-source solution for monitoring availability and performance of
349 infrastructure components (i.e. servers and network devices), as well as
350 applications [ZABB]_. It can be customized for use with OpenStack. It is a
351 mature tool and has been proven to be able to scale to large systems with
354 Delay in execution of actions
355 _____________________________
358 * Type: 'deficiency in performance'
363 - After detecting a fault, the monitoring tool should immediately execute
364 the appropriate action, e.g. inform the manager through the NB I/F
368 - A delay of around 10 seconds was measured in two independent testbed
373 - Cause of the delay needs to be identified and fixed
376 vim: set tabstop=4 expandtab textwidth=80: