As Service Providers utilize OpenStack to move into the Network Function Virtualization (NFV) world, it is evident that the cloud can not operate in a vacuum. To make NFV a reality, there will be Management and Orchestration (MANO) capabilities added to the OpenStack environment. Then there will be further automation and integration with existing BSS/OSS systems.
Not only will it be vital for the different environments to be integrated but it will be necessary for them to work together in order to understand the end to end picture and how events outside their area of responsibility can have impacts on the services within their purview.
Correlation of events across all environments will be key to successful life cycle management of VNFs. Lack of good correlation will result in inappropriate actions being taken by systems and people to address issues, leading to potentially longer service outages.
Monitoring & Correlation within the cloud (VIM)
1 - Has state of the infrastructure but no information on the services that utilize the infrastructure.
2 - Can recover a VNF but what about its configuration?
3 - What happens when there are issues outside the cloud that impact the service the VNF provides?
Monitoring & Correlation within MANO (NFVO/VNFM)
1- Has state of the services but does it truly know the infrastructure to correlate infrastructure events into service outages and recovery.
2 - Can recover and configure a VNF but can it really determine when the VNF is having issues versus underlay network issue?
3 - Does it understand how the customer connects to the cloud and can it isolate issues to prevent taking inappropriate actions?
Monitoring & Correlation in Legacy Assurance Systems (NOCs)
1 - Integration of the environments
2 - Will these systems understand the virtual world enough to determine when a customers service has been disrupted?
3 - Will these systems be able to have a complete end to end view of the customer’s service?
Life Cycle Management - will it be clear who own what components or will multiple orchestrators and systems attempt to “Heal” the same components?
