Boingo Wireless Case Study

Boingo Wireless, Inc. helps the world stay connected. Its vast footprint of small cell networks covers more than one million DAS and Wi-Fi locations and reaches more than a billion customers annually, in places as varied as airports, stadiums, universities, and military bases.

Boingo has experienced fast growth over the past few years. “We had acquired a lot of companies in support of our broadband and media-services business expansion, and we needed the agility to quickly create turnkey engineering and R&D solutions to support that growth,” says Kishore Raja, the company’s director of strategic programs and research and development. “We didn’t want to have to hire hundreds of workers and set up new data centers, and then spend a lot of time building solutions that might not work.”

The company also wanted to upgrade its business intelligence (BI) solution. “Like any company, we need to understand our user base and how our services are performing, so we can make necessary adjustments and also better plan our activities going into the next year and beyond,” says Raja. However, the company’s existing BI solution did not deliver actionable insights fast enough for business analysts. “We had about 45 minutes of latency when we ran queries against a year of data, and it also took several hours to load the data records. That performance wasn’t acceptable.”

To gain more agility and improve BI system performance, IT leadership at Boingo decided to explore a cloud solution.

After considering cloud solutions from Microsoft, Google, and IBM, Boingo chose the Amazon Web Services (AWS) cloud. “We did our due diligence, but Amazon was the clear winner,” says Raja. “AWS offered the most services and the best support, in addition to the business agility we wanted. We saw the cloud as a way not only to reduce costs, but to create solutions in a much shorter time frame.”

After choosing AWS, Boingo moved 90 percent of its core services—including a big-data warehouse and dev/test environments—to the cloud. The company uses Amazon Elastic Compute Cloud (Amazon EC2) instances, with Amazon Elastic Load Balancing to distribute incoming application traffic across the instances. Boingo stores BI data in Amazon Simple Storage Service (Amazon S3), and it controls access to all AWS services and resources with AWS Identity and Access Management (IAM).

Boingo uses Amazon Redshift to ingest multiple terabytes of analytical data each hour from different sources, including account and authentication data from hotspot venues. The company also collects hotspot session and connection data and other operational data in Amazon Redshift. Using a data visualization tool, Boingo business analysts run standard weekly and monthly BI reports on top of Amazon Redshift. Additionally, Boingo takes advantage of multiple Amazon Kinesis Streams to ingest streaming data into Amazon Redshift. Boingo also uses Amazon Kinesis Streams to process data from thousands of wireless LAN controllers and place it into Amazon Redshift.

Using AWS, Boingo has the agility to quickly add new compute resources in support of business growth. “We can do things so much faster using the AWS cloud,” says Raja. “Procuring hardware and adding compute and storage took 20 times longer using our previous on-premises system. And there’s no maintenance for us, because the Amazon team takes care of everything behind the scenes. We didn’t have to hire new engineering teams to keep pace with our growth.”

Boingo has seen significant cost savings bymoving to AWS. “Using AWS, we are only spending about 10 percent of what we spent to support our on-premises data warehouse,” says Raja. “With those cost savings, we can put more resources into growing our business.”

Since migrating its big-data warehouse to Amazon Redshift, Boingo has also experienced huge performance gains for queries and data loading. “On our previous big data warehouse system, it took around 45 minutes to run a query against a year of data, but that number went down to just 25 seconds using Amazon Redshift,” says Raja. “And it only took us 20 seconds to load one million data records, compared to the several hours it used to take. We continued to see these performance increases even when we spun up new clusters and added new virtual machines to our existing big data cluster.”

Boingo is benefiting from faster data querying and loading. “It is much faster for our business analysts to get useful BI data with Amazon Redshift, because they can run analytical reports faster,” Raja says. “And the data compression in Amazon Redshift means that we can also generate a lot more data and more reports.” With these capabilities, Boingo can more easily identify service issues that could affect customers. “With the faster data analysis turnaround we have using Amazon Redshift, we can be more proactive and solve issues faster than we could before.”

The company’s analysts also gained better control of business data by using Amazon Redshift. “Our analysts get more control and direct access to our data, which they didn’t have before,” says Raja. “They previously had to talk to the engineering team to get access to the data, but now they can use standard SQL clients and existing BI tools to interact with the data in different ways.”

Boingo plans to explore additional AWS services. “We are looking at machine learning technologies so we can do predictive analysis on some of the managed services we’re going to deploy in airports,” says Raja. “We’re thinking about using Amazon Machine Learning for that initiative because of the great experience we’ve already had with AWS. Being on AWS has really helped us improve our day-to-day operations.”