Querying the DFZ

RIPE RIS collects BGP data from collectors worldwide, providing visibility into routing dynamics. While this data powers RIPEstat and various third-party applications, operators and researchers often need analyses beyond what standard tools offer. On the other hand, using raw data MRT directly, is challenging. It is computationally intensive and requires custom parsing scripts. To cover this gap we are now publishing RIPE RIS data as indexed Parquet files, covering recent BGP
RIBs and updates.

As a result you can query routing data using SQL or datascience
tools rather than using MRT parsers. An analysis that previously took 10+ minutes
can now be done in seconds.

This approach has already proved valuable for our internal operations. When the
Baltic Sea cables were damaged, we could quickly explore the BGP impact. When analysis
the effects of the Iberian Peninsula power outage on BGP, we could investigate effects
within minutes instead of hours. This new data makes using route collector data
much faster and much more accessible.

This presentation will introduce the files and show how to use these new capabilities
through practical examples. I'll show how network operators can analyse bview/RIB/bgp snapshot
data and BGP updates this way - from investigating routing over time, to analysing stuck prefixes, or analysing the upstreams of a network as seen by collectors.

This data removes the traditional barriers of MRT processing, and makes BGP data analysis much more accessible, using much more tools and skills.