Search DAS42.com

Get Started
Background

Zayo Group is a communications infrastructure company.


Its network runs over 16 million fiber miles, and its team has to stay ahead of problems on that network to keep customers connected.

DAS42’s relationship with Zayo started through QoS Networks, a company DAS42 already worked with before Zayo acquired it. During that migration, Zayo’s data team saw DAS42’s engineering and architecture work firsthand and brought DAS42 on for the broader Snowflake environment.

Thanks to Snowflake and DAS42, we can ingest and integrate data much more rapidly than we’ve done in our past, and we keep the solutions really simple. That makes us more efficient in the long run and much more cost-effective as well.

Cindy Charles, VP of Data Engineering and Architecture, Zayo Group
The CHALLENGE

Fragmented Data and Overwhelming Network Alerts.


Zayo’s network throws off data from everywhere: device telemetry, service tickets, inventory, sales.

Like most companies with that kind of footprint, the data lived in separate systems that didn’t talk to each other.

The sharpest version of the problem showed up in alerting. Zayo’s network generates thousands of alerts a day. Most of those alerts trace back to a small number of actual root-cause events, but without a way to connect them, every alert looked like its own incident. Zayo’s team was spending time triaging noise instead of fixing problems.

The Solution

Unifying Network Data to Identify Root Causes.


DAS42 built a model in Snowflake to bring Zayo’s alert data together with the device and network data needed to explain it.

Instead of treating every alert as a separate event, the model links related alerts back to the incident actually causing them and compresses that set down to what the team needs to act on.

DAS42 worked directly with Zayo’s team to define the architecture: bring the alert data into Snowflake, enrich it with what’s already known about each device, and collapse the noise into a working set of incidents.

The results

There are thousands of alerts that go off on a daily basis, but there’s really only a few root cause events behind them. What we do is we have built a model on Snowflake to enrich that data and really distill all of those alerts down to a really finite set of events that people can take action on.

Cindy Charles, VP of Data Engineering and Architecture, Zayo Group

The alert-compression model cut Zayo’s network incidents by 15%.

Thousands of daily alerts now resolve down to the root-cause events that actually need a person, so Zayo’s team spends its time on the problem instead of the noise.

Zayo also picked up speed and cost efficiency on the data side. Ingestion and integration run faster than before, and a simpler solution set has kept the ongoing cost of running it down.

Looking ahead, Zayo plans to build on the same Snowflake foundation for more advanced analytics, including Cortex AI, Document AI, and Snowpark – continuing to work with DAS42, which holds Snowflake’s Advanced Data Science certifications.