Analytics Normalization Efforts (In progress)

Dublin and Beyond

Deployment Automation Workflow:

• Init

1. Deploy Normalization Microservice as part of ONAP Central
• On Demand
1. Deploy Edge Analytics App 
2. Create Kafka topic “edge-analytics-central-<id>” in DMaaP (Kafka) broker for publishing alarm from normalization microservice in ONAP format
3. Program DMaaP (Kafka) client in normalization microservice to publish on “edge-analytics-central-<id>”
4. Program DMaaP (Kafka) client in Policy to subscribe to “edge-analytics-central-<id>”

 Operational Workflow:

• Support the following alerts
• Host Aggregate/Cluster CPU usage exceeded a programmed threshold in the edge cloud
• Host Aggregate/Cluster Memory usage exceeded a programmed threshold in the edge cloud
• Host Aggregate/Cluster CPU & Memory usage exceeded a programmed thresholds in the edge cloud
• ONAP Central normalization microservice receives alerts from Edge Analytics App through REST interface
• ONAP Central normalization microservice publishes the above alerts on “edge-analytics-central-<id>” DMaaP (kafka) topic in ONAP format mapping from VM to relevant VNF by querying A&AI
• Policy receives alerts on “edge-analytics-central-<id>” DMaaP (kafka) topic in ONAP format (leverage the current format between DCAE TCA App and Policy -- https://wiki.onap.org/display/DW/TCA+application)
• Policy to perform VNF scale out

 Assumptions:

• Edge Cloud supports OpenStack and K8S
• Edge Analytics App is manually configured – configuration automation involves APP-C and out of scope for this effort
• ONAP Central Normalization microservice is a separate repo in multi-cloud

Dublin Scope

Goal

• Develop basic service assurance microservice supporting ONAP Threshold Crossing Alert (TCA) format 
  • TCA application
    • TCA format explained in "Table 1: Format of the Output Alert from the cdap-tca-hi-lo Microservice"
  • Policy R3 Casablanca - Control Loop Operational Policy 
• Supported service assurance policies on VMware Service Assurance Application
  • Host CPU usage exceeded a programmed threshold in the edge cloud
  • Host Memory usage exceeded a programmed threshold in the edge cloud
  • Host CPU & Memory usage exceeded a programmed thresholds in the edge cloud
  • Alert reporting format
    • Will report VMs which are part of the affected host
• Hardcoding of various parameters acceptable

Stretch Goals

• VMware service assurance product integration testing
• ONAP Lab Integration 
• Create Kafka topic “edge-analytics-central-<id>” in DMaaP (Kafka) broker for publishing alert from "service-assurance" normalization microservice in ONAP TCA format 

Key Dublin Work Items

• Helm Chart and Dockerfile for "service-assurance" Normalization Microservice  - Ramki - mostly complete
• VMware  Service Assurance Product Analytics API investigation for TCA w/ future proofing for Smarts - Girish/Ramki - in progress
• Initial code development for "service-assurance" Normalization Microservice - expected to be 1 week - this code will go into demo repository
  
  • Receive TCA Alert from VMware Service Assurance product Kafka interface on pre-specified VMware Service Assurance Application Kafka topic 
  • Convert TCA Alert to ONAP format – (leverage the current format between DCAE TCA App and Policy -- TCA application)
    • Map VM in Alert message to relevant VNF by querying A&AI
  • Publish TCA Alert to ONAP on pre-specified ONAP DMaaP (Kafka) topic 

VMware Service Assurance Product References

1) https://xtremio.me/2012/08/09/emc-smarts-part-i-overview-architecture/