A universal pain point for network engineers of even moderately sized networks is managing maintenance notifications. The problem isn’t always trying to tame the volume of these but also getting off notification lists that are no longer relevant and ensuring one is on those that are.
Following the successful work we have been doing with Euro-IX on the JSON Schema export for IXPs and the work we have been doing with PeeringDB to integrate the IXF and PeeringDB databases, we are proposing a new schema for networks to maintain and publish their maintenance notifications in a manner that we hope will become an industry standard.
You can see an exmaple of the schema using a fictitious IXP (but some real text) using the following links:
Please note that the ical format is not fully correct - just a demonstration for now.
Rationale - Provider View
- All networks announce service affecting maintenance
- Announcement / customer mailing lists are always out of date / inaccurate
- Notifications all have a common subset of data:
- start and end time
- description of services affected
- contact details of on call support team
- short after action report
- These are all typically unformatted emails
- This sucks and does not allow for automation
Rationale - Customer View
- Even medium sized networks receive a flood of maintenance notifications
- Quiet often, from old services
- Quiet often, from services that never made it past a sales enquiry!
- They usually do not receive the ones that actually matter
- Exceptionally difficult to find and understand upstream networks’ procedures for:
- announcing maintenance windows
- publishing planned maintenance windows
- understanding who receives announcements for maintenance windows
- Managing / sharing / scheduling work around floods of maintenance notifications is time consuming
- Cost efficiency and proper management requires automation