Visualizing Network Reliability in Real-Time with HEAVY.AI
One of the largest telecom companies in the United States offers wireless products and services to over 150 million subscribers.
This carrier's core network team is tasked with ensuring network reliability of voice, data, and VOIP calls. The team monitors this data as it streams in－terabytes per week from multiple platforms. Network reliability is critical to this telco, and monitoring and analyzing this data is paramount to sustain network reliability, and to continue to provide the level of service its customers expect.
Yet analyzing the massive scale of this data proved to be a daunting task. Even one of the largest telecom companies in the world had existing analytics solutions that were unable to efficiently handle the volume and velocity of it all, about 10 billion records per week. Network reliability reports took far too long to produce, meaning that a degradation or outage would take too long to identify and resolve, resulting in unhappy customers.
Network analysts couldn’t rely on these reports alone, which forced them to manually look through hundreds of millions of rows of records for trend analysis. The sheer volume of operational data meant that the team was unable to spot network anomalies, or identify network trends that consumed unusual or unplanned network resources.
“We're in the terabytes weekly of data generation and it’s just not humanly possible to parse through that, either with our own scripts or our eyeballs. We just can't see what all is in there,” says a Platform Engineer.
Performance & Ease of Use
For any major Enterprise, adopting new technology can be a challenge. This global telecommunications provider tried an array of analytics solutions, each promising to handle their unique data challenges. Setup time often took months. Yet one of their biggest obstacles was user adoption.
“Finding the right tool to address data challenges is a difficult task, there’s so much on the market and it’s hard to bring them all in for demos,” says a Sr. Manager of Technology. “We had significant challenges in finding a product that gave us the performance we were looking for, especially one with the entry level learning curve that facilitated people wanting to use the tool.”
Accelerating the Frequency of Decision Making
The global telecom carrier initially tried HEAVY.AI for its high performance, accelerated operational analytics, but quickly discovered that HEAVY.AI met more than just its performance needs. “Getting HEAVY.AI up and running, in a large environment like [carrier], was one of the better experiences that I've had in deploying a tool,” says the Sr. Manager. “The HEAVY.AI technical team was a joy to work with and that was key to us deploying so quickly.”
For HEAVY.AI and the carrier, that was only the start of the journey. They were soon using HeavyDB to process billions of rows of data in real-time, which allowed them to monitor network performance in HeavyImmerse and maintain its high standards of network reliability. Never before had this carrier been able to process and visualize such data volume with such velocity.
“The moment that I saw 6 billion rows of our data for the first time on a dashboard, it was like a muddy windshield was suddenly cleared, and I could see the whole road and all around me,” adds the Platform Engineer.
Data scientists and analysts were soon spending less time researching problems while arriving at solutions far more quickly. It also gave them time to deep-dive into their data to make more data-driven decisions, as well as to accelerate the frequency of their decision making.
What cemented HEAVY.AI as the solution of choice for this global telecommunication network carrier was the ease of installation and use, which helped drive adoption rates throughout the organization and created a single tool across multiple data centers and business units. “The usability of the platform really has far exceeded our expectations. We've had folks go in and, within a few minutes, started to create dashboards of their own. That sort of tool adoption rate is, in my view, kind of unheard of in this space,” quips the Sr. Manager. "Now everyone from engineers and planners, and managers on up to directors is using HEAVY.AI for the visibility it gives us into our data."
Network Reliability Reports in Seconds
With the help of HEAVY.AI, the carrier's network reliability reports now take mere seconds to produce, down from over 20 minutes. This allows its analysts to quickly identify customer impacting issues across the network. Engineers now derive insights faster as well, as they spot immediate problems and long-term trends in a few minutes, down from over 45 minutes.
With the ability to analyze massive data volumes at scale, and with growing user adoption, the company is discovering new ways of using HEAVY.AI. “So far, with HEAVY.AI, we've been able to solve multiple data problems across multiple platforms, and even across different environments. Instead of just network events and logs, we've actually done it with financial data, with spreadsheets, with smaller datasets that people may not normally look into for performance related data analytics,” states the Sr. Manager.
The telco is also leveraging HEAVY.AI's accelerated analytics to deliver WiFi services in several major U.S. stadiums. The core global telecommunication network team uses OmniSci to view network traffic and resolve stadium authentication and connection issues during short, two-to-three hour stadium events. “Now we have access to this stadium data in near real-time, so now we analyze what's going on at the event and we can provide the type of in-stadium service that our customers expect,” the Sr. Manager adds.
is also saving the global telecom carrier money by giving the company the ability to do data forensics and to identify misbehaving devices and rogue users that hog up inordinate resources. These users and devices derail their resource calculations and allocations. “It costs [carrier] money because we plan for a certain level of traffic or transactions. When you find out there's these rogue devices or rogue users, you have to go back and revisit your calculations, and you have to be able to hunt these devices down,” claims the Platform Engineer.
Accelerating Analytics Throughout the Enterprise
The telco carrier understands the huge opportunity in leveraging data in every aspect of its business, and this team is reaching out to business owners throughout the organization to explore those opportunities with HEAVY.AI. One of those opportunities is in preventative maintenance.
“We're basically looking at a five day rolling window, currently about nine billion rows, and we’re deep-diving into that data to prevent future problems and troubleshoot issues before they impact our customers,” the Sr. Manager explains.
To learn more about achieving telecommunication network reliability with HEAVY.AI, see how we support big data analytics in the telecom industry and download our modern telco data analytics whitepaper.
Also, download the full Telecom Industry Case Study to learn more about how HEAVY.AI supported this global telecom company.