Creating a Metadata Model

Understanding metadata and data models

Metadata is enriching information that can be associated to sessions to provide more context, filtering, and aggregation capabilities. See Understanding Metadata for a deeper dive into what it is and its uses within Cisco Provider Connectivity Assurance (formerly Skylight performance analytics). This article will help get you started with building your own metadata model.

Steps

1. Define user needs
2. Define business metadata
3. Define technical metadata
4. Investigate where this information can be sourced

Step 1 - Define user needs and set the context

Before diving into brainstorming all of the metadata you could add, take inventory of who your users are and what they want to accomplish with the tool.

• Who are the ultimate users ?
• What are the needs?
• Who are the personas?
• What will the metadata be used for?
• Which dashboards and filtering are needed?
• System be usage - proactive or reactive?

Work together with users from different groups within your organization to flush these use cases out up front.

For example:

• Regional metrics reporting may be needed
• In regions there may be sub areas such as MTX sites or cities
• Technical staff may want to see equipment types and technologies 2G, 3G, 4G, and 5G
• There may be different access technologies - microwave, fixed, broadbandm leased, ect.
• How about subtended links. Cell to cell microwave from the core.
• Data rates, link capacity, class of service, etc..

Step 2 - Technical Metadata

• Location

E-Line Service Offering

Step 3 - Business Metadata

Ethernet Business Service
A service entity that is defined by business ethernet services SLA's and technical implementation requirements under a contractual SLA agreement with a CUSTOMER that is a of TYPE Business Customer.

• Technical Metadata may descibe the structure, format, technical aspects, etc.
• Business Metadata describes the business context, service associations, context, etc
• The Data represents actual value for the metadata.

Reserved Metadata

Geo Coordinates

Used to display heatmaps on relevant map visuals monitored objects. The reserved pairs are:

What is included in the Metadata (Category):

• source_lat,
• target_lat,
• source_lon,
• target_lon,

Each category can then be populated with the objects starting and ending longitude and latitude.

You can see it against the session:

Examples

Ethernet Services Example

Point to Point Example

Hops are metadata. In terms just hop3 may be a link between service providers. We can have a different value for hop3 to indicate its a provider or create a new Metadata that defines the LINK_TYPE as a METADATA tag. PM flows are always end to end.

Point to Multi-Point

A PM flow per A-Z and per CoS
This is a very common architecture for residential and business broadband/mobile backhaul services PM sessions can be grouped in Analytics to immediately understand Core vs edge impact.

Deliverable

Creating Templates

Data about Data
The Data is qualifying information and data is metric/KPI/flow/PM session ect. The model helps to identify how the data about the data fits into a particular 'context'. This is in fact turning data into information for insight/action/management/enlightenment! A TAG can be any arbitrary 'TAG name' that helps define the data set. We create tags that are meaningful to the topology, srvice, or business model.