Burt Case Study
From a team of two entrepreneurs with an idea, to 30 employees across Sweden and the United States, Gothenburg-based Burt is an example of how keeping operations agile can lead to big success in a highly specialized market. The company started up in 2009 to serve businesses in the online advertising market, including marketers, publishers, and advertisers. Since 2012, the startup’s data intelligence and analytics platform has been focused on media companies in Scandinavia and New York, and now counts among its users several of the top 100 media groups in the world, including the Hearst Corporation and Bonnier. Through the platform, publishers can measure, analyze, and improve the impact of the advertising across their range of online deliverables, enabling them to make the most of a significant, and growing, revenue source.
It would cost hundreds of thousands of dollars for us to build and manage a data center, and we’d have to factor in time to order and install new hardware, impacting our ability to get to market quickly. We knew we wouldn’t have these constraints with AWS.”
Chief Technology Officer, Burt
Burt was in a position similar to many startups at the beginning of their journey: It needed flexible IT that would allow it to experiment cheaply and quickly get to market with new products and services. “From a technical perspective, we needed capacity to scale quickly as the amount of data we collected from websites expanded and we increased the services we delivered to customers,” says Burt Chief Technology Officer Carl Nelvig. “Plus, we needed a platform on which we could run various databases in order to crunch all the data. Having the flexibility to integrate with other technologies was key.”
The company’s options at the time of launch included starting up in the cloud or building its own data center. Apart from the significant capital expenditure required for building a physical infrastructure in-house, Burt also saw that it would require ongoing management at a much higher level than running a cloud model. “Both up-front and ongoing maintenance costs were, of course, a consideration,” says Nelvig. “To maintain agile operations, we needed an architecture that was simple to manage. We wanted the flexibility to give any member of our team the ability to spin up an instance as required, and we wanted to avoid the lead time and costs involved in building new hardware.”
Why Amazon Web Services
The cloud won out for Burt. “It was the obvious choice as far as we were concerned,” says Nelvig . “And at the time AWS was also the obvious choice. The breadth of its services is something AWS competitors couldn’t match.”
The firm uses a wide range of AWS technologies to power its analytics platform. Each day, Burt uses up to 500 instances of Amazon Elastic Compute Cloud (Amazon EC2) for its compute needs. Auto Scaling helps the firm automate the scaling of resources and minimize hands-on management time, and Elastic Load Balancing plays a major role in distributing traffic across instances—again, helping the Burt team automate as much as possible. Amazon Simple Storage Service (Amazon S3) provides object storage and Amazon Glacier is used for archiving. Data warehousing needs are fulfilled by Amazon Redshift.
Amazon Elastic MapReduce (Amazon EMR) and Apache Hadoop enable the company to process the vast amount of data related to scheduled reporting. And, with AWS CloudFormation—and, more recently, AWS CodeDeploy—Burt can automate deployments and eliminate manual operations further.
AWS CloudTrail delivers log files containing details of all AWS API calls, including identity, time, and source IP. And fully managed service AWS Config provides a comprehensive inventory of Burt’s AWS resources. Amazon Simple Notification Service (Amazon SNS) is used for internal reporting purposes. The firm has also started running code in AWS Lambda, and is looking forward to exploring more of its functionality.
Nelvig highlights the benefits of AWS Availability Zones (AZs). “In terms of regions, we mainly use Ireland, and increasingly the United States,” he says. “It’s extremely important that we are able to use three AZs per region.”
Burt engineers use a large number of NoSQL databases—as well as technologies such as RabbitMQ and Apache Kafka for transferring data—so any infrastructure it chose had to be fully interoperable with a range of technologies.
For Burt, launching with AWS has meant it has the flexibility to deliver its specialist analytics platform to the demanding media houses it serves. Nelvig says the ability to have more than one AZ per region was something not all cloud providers could offer and it’s a feature of AWS that’s vital to the company's operations.
“It would cost hundreds of thousands of dollars for us to build and manage a data center, and we’d have to factor in time to order and install new hardware, impacting our ability to get to market quickly,” Nelvig says. “We knew we wouldn’t have these constraints with AWS. We also knew we could build on top of AWS with other services as well as our home-built solutions.”
Nelvig says the AWS model is one that suits the business and aids experimentation and innovation. “Our use of AWS has evolved organically,” he says. “We started with one instance, and we’ve been adding, making mistakes, removing, adding again, and so on, to get to where we are now. Having this sort of agility helps keep management operations lean so we can focus on our core product. If we were running our own data center, we would be spending more time on infrastructure management and not on developing our offering.”
Burt is servicing several large media companies, with services to the organizations growing daily. Nelvig says the company is confident that AWS can provide the compute, storage, and processing capacity it requires to meet the customers’ needs.
“Today, we collect somewhere in the region of 100,000 messages a second, which is then processed by our platform,” he says. “Scaling up and down is a very important aspect of how we work. Working across Europe and the United States means that traffic is high nearly all the time. We try to automate as much as we can, so Auto Scaling saves us a lot of management time.”
Maintaining a highly available infrastructure is crucial to Burt. “Our system availability is good. Bringing data together for our clients is our business, so ensuring we don’t lose any of it is the most important thing for us," says Nelvig. "Using AWS, we’ve been able to maintain high availability to our customers. We’ve completed thousands of upgrades and deploys without impacting services.”
Nelvig points to the ISO27001 information security standard and AWS’ security credentials, which have helped the startup on its journey. He says: “If we had our own data center it would be more difficult for us to satisfy the compliance needs of our customers and convince them or our system’s security.”
As a result of working with Burt, some clients have adopted AWS technologies themselves. “Some of our customers run their own AWS instances. These publishers have highly specialized requirements and we can’t build everything for them, so we encourage them to use AWS too. It’s a situation that’s working extremely well for all parties.”
From a team of two entrepreneurs with an idea, to 30 employees across Sweden and the United States, Gothenburg-based Burt is an example of how keeping operations agile can lead to big success in a highly specialized market.
Benefits of AWS
- Ability to process approximately 100,000 messages per second
- High availability, completeing thousands of upgrades and deploys without impacting services
- Easily satisfy customers' compliance needs with AWS standard security credentials
AWS Services Used
Amazon Elastic Compute Cloud (Amazon EC2) is a web service that provides secure, resizable compute capacity in the cloud. It is designed to make web-scale cloud computing easier for developers.
Amazon Simple Storage Service (Amazon S3) is an object storage service that offers industry-leading scalability, data availability, security, and performance.
Learn more »
AWS Auto Scaling monitors your applications and automatically adjusts capacity to maintain steady, predictable performance at the lowest possible cost.
Learn more »
Redshift powers mission critical analytical workloads for Fortune 500 companies, startups, and everything in between.
Learn more »
Elastic Load Balancing
Elastic Load Balancing automatically distributes incoming application traffic across multiple targets, such as Amazon EC2 instances, containers, IP addresses, and Lambda functions.
Learn more »
Amazon EMR is the industry leading cloud-native big data platform, allowing teams to process vast amounts of data quickly, and cost-effectively at scale.
Learn more »
AWS CloudFormation provides a common language for you to describe and provision all the infrastructure resources in your cloud environment.
AWS CodeDeploy is a fully managed deployment service that automates software deployments to a variety of compute services such as Amazon EC2, AWS Fargate, AWS Lambda, and your on-premises servers.
Learn more »
Companies of all sizes across all industries are transforming their businesses every day using AWS. Contact our experts and start your own AWS Cloud journey today.