4
C O M M I S S I O N E D B Y AW S
PAT H F I N D E R | T H E J O U R N E Y T O S E R V E R L E S S - F I R S T
Use cases
i R o b o t
COMPANY
iRobot
INDUSTRY
Consumer robotics and smart home
EMPLOYEES
1,000+
HEADQUARTERS
Bedford, Massachusetts
USE CASE
Serverless for the edge
KEY SERVERLESS SERVICES
IoT platform, API Gateway, compute, database, storage, data
streaming, ETL, query service, provisioning
Connected technology company and maker of the Roomba robotic vacuum cleaner, iRobot
wanted to scale its technology out without scaling its teams up. This required the company to
consider how it could deliver its products and services more efficiently, and how it could reduce
the number of services that it had to manage. Concurrently, iRobot needed to maintain or
improve reliability and get new features to market quickly.
The company saw serverless as the right fit: serverless met iRobot's technical needs while
allowing it to expand customer capabilities and enter new markets without increasing team size
or infrastructure. Today, the team size is about 25 people, all of whom get to focus on building
applications that drive business value because they don't have to worry about server and
infrastructure maintenance. Innovation remains at the heart of iRobot's vision. That innovation
is powered by the ability of development teams to run efficiently and increase the rate at which
they release new features. The company has been able to enhance its robots and expand into the
smart home market without expanding its team.
"Being serverless, we have so many fewer tasks to handle," says Ben Kehoe, cloud robotics
research scientist at iRobot. "Our operations burden has not increased as our fleet of connected
robots has vastly increased in size over the past couple years." Serverless, says iRobot, enables
the company to handle spikes in orders without additional personnel. On Christmas day, for
example, daily traffic spikes to 20x the average. In traditional architectures, that spike would
necessitate the allocation of extra engineering time and resources. With serverless, those spikes
become a 'non-event.'
