云管理架构基本特点
• Divide and conquer on Service Level
– IAM - Identity & Access management(Authentication/Authorization)
– IRM – resource management and policy enforcement on all IT Resource
– UOM - traditional F/P management of all IT resources
– UEM – Data/Event collection service( event based)
– DMS – relation management among resources/entity
– AAS - root cause/Impact analysis and remediation decision making (no action execution)
– USM –service offering lifecycle management, service instance provisioning
• Self-contained Service , no much dependency among services.
– Each service should/could expose Rest-API to access feature/function
– Each service should/could send notification on managed entity
– Each service don’t lock in specific implementation of depended functions,
• broker based internal design
• MSG/event based
– Each service could/should do its own authentication/authorization
– Each service has its own field model
• Design for change and Technical swapping
– Not lock-in on virtualization platform
– Not lock-in on NSM system/OM facility
– Not lock-in on configuration management system
– Support vmware / openstack / GalaX as virtualization but GalaX at first
– support ovf / huawei formation template but huawei template at first
– Changing with less pain
• Design for future and real world
– Provide integration point for following point : Ticket system, workflow, Authentication/Authorization, Lagency resource pool, Lagency OSS/BSS system
• Group common function as internal service
– infrastructure service : workflow service/Message service
– Function service : UOM/CIS
• Message based communication
• DMS as central relationship/dependency information center