《OpenStack企业应用之道.pdf》由会员分享，可在线阅读，更多相关《OpenStack企业应用之道.pdf（47页珍藏版）》请在三一文库上搜索。
1、Win The Enterprise: OpenStack 企业应用之道 Shane Wang/王庆 Intel Open Source Technology Center 英特尔开源技术中心 微信号: qq559382 Agenda About Us OpenStack as an Open-Source Project OpenStack User Survey in 2014 Win The Enterprise at OpenStack Foundation Survey from OpenStack Shanghai Meetup at Intel Win The Enterpris
2、e at Intel Summary ABOUT US Who are We? A Platinum member in OpenStack Foundation. A couple of teams in Intel who are working on or using OpenStack, including Data Center Group, Software and Services Group, Intel IT etc. Top 10 contributor since Grizzly (http:/ A few cores in Ceilometer, Gantt, Hori
3、zon, oslo, Swift etc We are dev (the biggest dev team in Intel) instead of ops Datacenter and Cloud Software (DCS) Team Open Source Technology Center (OTC) Software and Services Group (SSG), Intel Teams at: Shanghai (Zizhu) and Beijing (GTC) in China Santa Clara and Portland in US Why and What are w
4、e doing OpenStack? What are we doing in China? Nova, Neutron, Ceilometer, Ironic, TripleO, Congress, Horizon, Gantt, Cinder, Gl ance, Keystone etc. but Intel is doing more, including running OpenStack in Intel (which is called iLab) Meetup, summit, hackathon, corportation collaboration, tech support
5、 etc OpenStack设计与实现is on-the-way (to be published in May, 2015) Why does Intel do OpenStack? Make OpenStack a better choice for Enterprises Add Intel features to make OpenStack rich Optimize OpenStack on Intel platforms Build OpenStack ecosystem Etc. Stress on Datacenter Operations 1: Source: Intel
6、IT internal estimate; 2: 3: IDCs Digital Universe Study, sponsored by EMC, December 2012; 4: IDC Server Virtualization and The Cloud 2012 Network 2-3 weeks to provision new services1 Storage 40% data growth, 90% unstructured3 Server Average utilization 500 nodes OpenStack Version 4% 9% 17% 11% 37% 1
7、3% 9% Folsom or earlier Grizzly Havana Icehouse Juno Kilo No idea Enterprise Readiness 11% 83% 0%6% Not mature Mature Very mature others Tools for Deployments and Configurations 2 0 4 0 14 5 0 9 19 1111 0 2 4 6 8 10 12 14 16 18 20 AnsibleCFEngineChefCrowbardevstackFuelJujuPackstackPuppetSaltstackOth
8、ers WIN THE ENTERPRISE AT INTEL High Availability of Tenants VM Live Migration Hardening (speed & failure ratio) Nova Live Migration & Volume Migration https:/ https:/ https:/ https:/ Automatic Evacuation http:/ https:/ https:/ Host Maintenance Mode https:/ in-maintenance-mode https:/ High Availabil
9、ity of Services Always ON and respond during failure and massive stress HA guide is out of date http:/docs.openstack.org/high-availability-guide Requests are lost when Heat engine dies https:/ https:/ https:/ https:/ No VLANs support in one of Neutron HA solutions http:/specs.openstack.org/openstack
10、/neutron-specs/specs/kilo/neutron-ovs-dvr- vlan.html (COMPLETED) High Availability of Services No retry in Glance https:/ Unresolved state when Cinder service dies https:/ taskflow-support https:/ taskflow-support-api https:/ taskflow-support-scheduler https:/ taskflow-support-volume Cinder cant run
11、 in an Active/Active mode (owned by RH) Rolling Upgrades ComputeManagerConductAPIConductManagerDbAPI Objects Objects call methodcall/cast call method remotable remotable_classmethod object_action (copy an obj) object_class_action db new routine old routine ComputeManager ComputeManager, Comp uteAPI,
12、 etc local call method local call method Object Versioning! Driving Monitoring Standards Collect dynamic runtime metrics in OpenStack Nova with monitor plugins on compute nodes CPU utilization, memory usage, network bandwidth etc. Add physical host monitoring into OpenStack Ceilometer Hardware polls
13、ters, SNMP inspectors, and enhanced pipeline OpenStack Service TypeMetrics (e.g.) Nova Static capabilities CPU features hypervisor version Dynamic Resources free memory/disk vCPU # PCI devices # of NIC virtual functions Extensible runtime measurementsCPU time Ceilometer Resources creation/deletion V
14、M network/subnet/port image Resources usage data CPU usage in VM memory usage in VM network usage in VM storage usage stats Hardware related CPU load Memory usage Network usage Disk usage Power consumption Intelligent Workload Scheduling VM Placement Utilization Based Scheduling Pluggable metric dat
15、a collecting framework New filters/weighers in OpenStack Nova Monitor metrics CPU load Power & Themal via IPMI VM Runtime Dynamic Resource Reschduling Reuse policy engine in OpenStack Congress Step towards SLA Hardware Compute, Network Storage elements Scheduler (Gantt) VM requests alarms New / Paus
16、e Stop / Destroy Migrate Workload History CPU, RAM Network Security policies SLAs Energy Cost Function Hardware Availability VMs, hosts, Migrate action VM Profiles VM Profiler Filter-List + Weights Heat, Horizon, Nova Tetris, Notifications Ceilometer Ceilometer, Monitors NagiOS Nova Neutro n Cinder
17、Intelligent Scheduler Tetris API Migrate_VM(vm, host) Migrate_VM(vm) Spawn_VM(vm-data) noDB Intel Enables OpenStack Deployments Make Enterprises and cloud service providers deploy cloud solution successfully Contributions across OpenStack projects plus tools released to Open Source Active contributo
18、r in OpenStack community Performance optimizations, validation and patches Intel IT Open Cloud with OpenStack Deliver Consumable Services Enable Automated Management of Cloud Intel Cloud Builders Reference Architecture Contributions Intel IT Open Cloud Validate Intel IT customization of OpenStack Do
19、cument best practices Share best practices with enterprises and services providers Others Persistent connections for calls New filter, new scheduler considering rack Improve database throughput and performance Ceilometer tuning, time-series monitoring Real-time monitoring and alarming Speed up messa
20、ge Q handling, e.g. enable ZeroMQ Dashboard localization Deployment tool customization Storage optimization, e.g. fast boot Docker support Auto scaling Rollback support for rolling upgrades Mobile OpenStack Management Authentication, Administration anywhere anytime, Alarming, so many others SUMMARY
21、Is OpenStack Production Ready for Enterprise? OpenStack Projects Open Source Framework of a loosely coupled components A set of key components to run and satisfy minimum needs from enterprises Interests from different companies Open source community Products customized based on OpenStack Mature and
22、stable enough Easy to deploy and use Customer needs satisfied Continuous technical support and service Call for Actions Test OpenStack with your customers and collect feedbacks Report bugs as more as possible when you find them in the deployment Help to fix bugs as many as possible Bring WTE feature
23、s to OpenStack upstream After fixing problems for your customers, try to contribute patches to WTE features in OpenStack Legal Information Copyright 2015 Intel Corporation. All rights reserved. Intel, the Intel logo, Xeon are trademarks of Intel Corporation in the U.S. and/or other countries. *Other
24、 names and brands may be claimed as the property of others. All products, computer systems, dates and figures specified are preliminary based on current expectations, and are subject to change without notice. The products described in this document may contain design defects or errors known as errat
25、a which may cause the product to deviate from published specifications. Current characterized errata are available on request. Contact your local Intel sales office or your distributor to obtain the latest specifications and before placing your product order. No computer system can provide absolute
26、security under all conditions. Built-in security features available on select Intel processors may require additional software, hardware, services and/or an Internet connection. Results may vary depending upon configuration. Consult your system manufacturer for more details. For more information, see http:/security- Thank You!