Deploying StarTree for Real-Time Enterprise Network Optimization at 7SIGNAL

Summary

• The development-to-deployment cycle will be reduced from 4 months to 1 month for UI updates that require changes to aggregations.
• Based on the results of StarTree’s proof of concept, we anticipate most queries will execute 2-10X faster.
• 7SIGNAL anticipates >50% cost savings for time series data storage and aggregation.

Scaling architecture for real-time network optimization

For businesses of all types, wired and Wi-Fi-enabled devices have increasingly become mission critical. Think hospitals relying on around-the-clock, medical devices and EHRs, machines at manufacturing plants that aren’t hardwired, or retailers using inventory management and point-of-sale systems. That makes constant connectivity a must.  

Yet the number of qualified Wi-Fi engineers on the market who can reliably monitor the performance of these crucial systems around the clock is relatively small. Many of the existing Wi-Fi observability tools only give companies information on connectivity at a given point in time or on certain access points.

7SIGNAL fills that gap, providing its customers with real-time enterprise monitoring of wired and Wi-Fi connectivity across their locations.

“We’re always there. We’re always watching,” said Joe Tennant, 7SIGNAL’s vice president of software development. That allows Wi-Fi engineers to know “what’s happening right now in a building across the town” that they may not be able to access. 

But with larger deployments and more expensive data aggregations, 7SIGNAL needed to improve the scalability of its infrastructure. In this case study, we’ll delve into how 7SIGNAL has accomplished its goals with StarTree’s help.

Client-side monitoring

7SIGNAL’s platform is made up of endpoint agents and network performance sensors. The platform monitors both the wired and wireless network as well as the wired and Wi-Fi clients and IoT utilizing it. The platform runs active and passive tests and delivers that information via the cloud to globe IT teams, help desk teams, and network engineers.

The company’s unique method of monitoring speed and connectivity from the client side, as opposed to the network side, enables it to recognize network inference caused by changes in the physical environment, such as radio interference, client roaming behaviors, inventory fluctuations, and more. This also gives it an improved ability to troubleshoot connectivity problems such as issues with DHCP or a captive portal. 

“Acting as a client gives us the whole picture of what the client’s experience is going to feel like,” Tennant said.

Real-time data monitoring

For a time, 7SIGNAL relied on a combination of AWS-provided services, but as their data needs scaled, they began looking for more optimized solutions. The aggregate data 7SIGNAL received tended to be at a lag of around 23 minutes, with additional time needed to crunch the numbers and send performance insights back out to customers. 

“In some cases, there can be up to a 20-minute delay in terms of what customers see in our dashboard, at least in terms of the aggregated data that we display in the charts,” said Paul Adams, a developer at 7SIGNAL leading the implementation of StarTree.

This lag impeded 7SIGNAL’s goal of providing its customers with real-time monitoring. “To be able to serve our customers most effectively, we need to know what is happening right now.” 

After 7SIGNAL reviewed a range of solutions, they settled on Apache Pinot, which is capable of ingesting vast amounts of data from a wide range of sources, including streaming (Kafka/Kinesis/Pubsub) and batch (S3/GCS/ADLS), along with SQL sources like Snowflake and BigQuery. But how could 7SIGNAL’s already busy team get the benefits of Pinot without taking on the significant time and cost requirements of deploying and managing it internally?

7SIGNAL came away from their search with one clear answer: StarTree.

Aggregation and speed

StarTree offers a fully managed platform built on Pinot to run sub-second queries on petabyte-scale data across all three major cloud providers. For 7SIGNAL, StarTree “really stuck out as being specifically tailored to our use case in that it is geared toward aggregations on huge quantities of time-series data with extremely low query latency and extremely fast ingestion capacity,” said Adams. With StarTree, data freshness delays dropped from 23 minutes to less than 5 minutes.

Before implementing StarTree, 7SIGNAL often had to partition its data when handling larger deployments. Now it can simply put all of that data together on a scalable and performant platform.

With StarTree’s services, “whichever combination of data points that we need to look at to identify Wi-Fi problems, I feel pretty confident … we should be able to get those answers,” Adams said.

Unlocking value

While 7SIGNAL is still in the implementation phase, it has already observed a 2-10x uptick in the speed of queries on data sets across the board.

StarTree has also expedited 7SIGNAL’s growth. “It used to be that if we needed to deploy a new widget, we had to build out an all-new aggregation pipeline to get the data specific to that aggregate set,” Adams said. In real terms, that can translate to a period of four months for development time as well as fully integrating historical data.

That development-to-deployment cycle has been reduced to one month, thanks to StarTree. “It’s just a matter of writing a new query and we’re ready to go,” Adams said. “It just really frees us up to be able to display any kind of data that we need to display, whatever aggregation it is.”

The team expects 7SIGNAL’s performance to far exceed its requirements thanks to StarTree.

“It’s super flexible. It’s very fast. The infrastructure is simpler,” said Tennant. Plus, it’s a relatively cheap solution, with greater than 50% cost savings for time series data storage and aggregation. “StarTree is a company that’s great to work with, and they’re very helpful.”

Future use cases: AI, algorithmic systems

As the company’s deployments grow, it’s excited about using StarTree to “open up a lot of capabilities that, at present, it doesn’t have,” Tennant said.

7SIGNAL intends to expand its use of algorithmic tools to process data and categorize it based on the type of Wi-Fi connectivity problem. StarTree will give it the freedom to assess a range of data sources in different categories or combinations.

As Tennant explains, “This offers us a lot of flexibility as we’re building out training sets as we’re developing different ways that we can use machine learning to help identify Wi-Fi issues.”

Discover StarTree Cloud

If this case study sparks ideas for your own data architecture and use cases, you’re encouraged to learn more about StarTree Cloud, its advantages over open source Apache Pinot. If you already have heard enough, feel free to schedule a demo or sign up for a 30-day free trial.