How BISSELL migrated 1,000,000 vacuum gadgets to a scalable IoT Platform on AWS

This publish was co-authored by Ramesh Chinnasamy, Principal Lead, IoT at AWS; John Rotach, Sr. Software Architect at AWS; and Mike Ply, Affiliate Director, IoT at BISSELL

Shoppers in nations all over the world are discovering worth in related vacuum cleaners to sort out life’s on a regular basis messes. As a part of the Web of Issues (IoT), related vacuums present shoppers enhanced performance, worth, and cleansing efficiency. Vacuum connectivity provides shoppers the flexibility to remotely management and customise vacuum cleaner operations reminiscent of beginning, stopping, or scheduling cleansing, monitoring location, monitoring utilization and battery cost, and total permits an improved cleansing course of.

BISSELL is a family-owned enterprise with over 145 years crafting premium cleansing merchandise that exceed client expectations. BISSELL affords a full vary of ground care and air therapy merchandise. Linked vacuum cleaners – significantly robotic vacuums –  are a rising phase of their enterprise. BISSELL has 1 million registered vacuum gadgets of their current IoT Platform powered by AWS IoT, and at the moment 25% of these gadgets are related within the subject. With growing adoption of related vacuum cleaners all over the world, BISSELL is anticipating 1 million gadgets to be related within the subject within the close to future. BISSELL needed to make sure its IoT platform may scale with this speedy anticipated development throughout product classes.

On this weblog, we focus on BISSELL’s IoT platform and the way BISSELL teamed up with AWS Professional Services and TensorIoT to create a scalable, high-performing, and cost-effective IoT platform in preparation for onboarding merchandise and bringing new merchandise to market.

Platform challenges

Previous to working with AWS Skilled Providers, BISSELL constructed an IoT platform on AWS with varied tradeoffs mentioned beneath. They used this platform to carry out end-to-end machine administration utilizing a wi-fi + cloud + app technique as proven within the picture beneath.

BISSELL constructed the platform utilizing options from varied suppliers however had challenges when attempting to innovate and add options together with:

  • Options acquired by way of varied companions restricted alternative for the BISSELL growth staff to “personal” the platform.
  • Restricted visibility and traceability of machine connection and information circulate inefficiencies inside the system.
  • Operational and efficiency points with operating over 450 lambdas to help information transformation and information circulate inside the system.
  • Lack of knowledge governance and information inconsistencies when storing machine information to Amazon DynamoDB.
  • Restricted automated testing for platform integration.

BISSELL engaged with AWS Skilled Providers and TensorIoT to assist deal with these challenges and convey frameworks and greatest practices to:

  • Lower time to onboard gadgets to the platform.
  • Enhance the soundness of system.
  • Simplify information evaluation.
  • Enhance API response time and total efficiency of BISSELL’s OT programs.
  • Enhance machine pairing success fee.
  • Enhance monitoring and logging to cut back manufacturing downtime.
  • Enhance troubleshooting of points.
  • Enhance automated testing.
  • Optimize utilization of AWS capabilities.

Answer walkthrough

BISSELL, AWS Skilled Providers and TensorIoT migrated the prevailing manufacturing platform to a brand new platform developed utilizing Connected Device Framework (CDF). They started by working backwards from the issues, developed a minimal viable product (MVP) function set, and migrated solely what was essential to have a strong base of CDF providers and different microservices to show out scalability and agility. They repeated this course of in subsequent phases by iterating on the bottom providers and labored backwards so as to add crucial functionalities. Including minimal safety guardrails and optimizing price and efficiency had been additionally key focus areas in these phases. These migration methods and deployment procedures created a seamless buyer expertise with minimal to no downtime.

Section 1 : Constructing baseline structure to fulfill scalability and agility

The section 1 aim was to place baseline parts into place and construct upon these parts to implement an MVP function set, that are widespread to all BISSELL related machine product classes and adequate to function a primary BISSELL machine end-to-end. These MVP options had been:

  • Machine pairing – permitting a consumer to pair a related machine with their account utilizing the BISSELL cell software.
  • Knowledge ingestion –an information ingestion pipeline that enables the platform to eat information reminiscent of machine cleansing schedules and fluid utilization.
  • Knowledge historical past – information storage and an API that enables customers to view machine historical past and metrics.

The staff began by working backwards from the prevailing IoT backend system, diving deep to find out which options had been important within the new platform and which had been unused or incomplete. Then, they produced an structure which met the wants of the platform within the near-term however is also prolonged for future use instances.

They first deployed foundational parts from Connected Device Framework (CDF), together with the Asset Library and Provisioning microservices. Subsequent, they did upfront work to attach BISSELL’s CI/CD system, which was key for the staff to develop shortly.

In 3 months, the staff accomplished growth and replicated the MVP options referred to as out above within the new platform. Keys to the staff constructing the total MVP options on this time had been:

  • Incorporating baseline performance “out of the field” from CDF.
  • A modular structure with minimal interdependency between parts.
  • Deep understanding and accompanying documentation produced from deep-dive periods round current platform.
  • Elevated developer productiveness attributed to the flexibility to face up and run particular person parts regionally on developer laptops.
  • A clear, well-understood and documented new code base.

Section 2: Iterate function set and optimize efficiency

The section 2 aim was to construct upon the MVP function set, including in options required by BISSELL’s extra advanced related merchandise. These options included:

  • Command and management – means for customers to start out and cease machine cleansing schedules.
  • Machine scheduling – means for customers to create cleansing schedules with the platform commanding cleansing schedules based mostly on these schedules.

Together with the above options, the staff wanted to herald different current functionalities reminiscent of the whole suite of cell APIs, machine APIs, and the machine administration portal to allow your complete fleet emigrate to the brand new platform.

Following the section 1 blueprint, the staff labored backwards from issues and held deep-dive periods to know performance within the current platform to find out what options to copy within the new platform. By leveraging the productiveness enhancements realized in section 1, the staff additionally accomplished the section 2 implementation in 3 months and was prepared to start full-platform testing and migration.

Different key non-functional options carried out  within the new platform:

  • Implementation of greatest practices throughout your complete code base in areas of safety and monitoring and logging throughout all microservices.
  • Visibility into function utilization and correlation with prices.
  • Capability to auto deploy copies of your complete platform stack to different environments (growth and check sandboxes).
  • Absolutely automated testing that included unit exams, integration exams (e.g. identification server interplay exams and API pairing exams to make sure right writing of knowledge to Amazon DynamoDB), and system exams (e.g. simulate precise consumer conduct at an HTTP stage) to confirm no deterioration in performance and to make sure interactions between modules labored as anticipated.

Section 3: Migrating the fleet with zero downtime

Migrating over 1 million registered IoT gadgets to a brand new cloud structure is advanced. Working with AWS Skilled Providers, BISSELL reviewed 3 choices for migration:

  • Iterative deployment: deploy CDF options as they’re accomplished.
  • Coordinated big-bang deployment: deploy the brand new cell software, machine firmware, and cloud answer on the similar time.
  • Cloud big-bang deployment: deploy the cloud answer as one full deployment.

Every answer posed completely different dangers, complexities, feasibility, prices, and timelines. The staff collaborated intently to reach at a balanced answer that greatest meet the wants of BISSELL.

The staff determined to maneuver forward with the cloud big-bang deployment. They might deal with remaining tech debt in publish go-live actions. BISSELL went dwell with the brand new platform with zero down time for shoppers and their name facilities acquired zero calls on points associated to going dwell.


The brand new BISSELL IoT platform based mostly on the Connected Device Framework (CDF), met all of its goal targets to scale on efficiency and prices. BISSELL diminished its AWS Lambda prices by 77 p.c, and its Amazon DynamoDB prices by 65 p.c; they diminished API response time by 50 p.c. Within the new growth surroundings, the event cycle time is now minutes as a substitute of days. In manufacturing, they’ll now resolve points in hours as a substitute of days. By way of applicable logging patterns and detailed alerts, they improved platform upkeep.

BISSELL’s collaboration with AWS has put the BISSELL IoT platform ready to scale previous its 1 million related gadgets aim. The platform offers a greater expertise for shoppers and helps them understand extra worth from their related cleansing gadgets.

To study extra about AWS IoT providers and options, go to AWS IoT or contact us. Additionally, study extra about how BISSELL constructed scalable, performant, and cost-efficient REST API for IoT Platform within the second publish of this collection, coming quickly!

Concerning the authors

Ramesh-Chinnasamy-2Ramesh Chinnasamy is a seasoned chief for International IoT follow inside the AWS Skilled Providers International Specialty Observe. He leads IoT options for our prospects in addition to develop greatest practices & packaged IoT options. He helps our prospects and companions to ship finish to finish IoT options from the sting by way of to the cloud to realize enterprise outcomes.
John Rotach-2John Rotach is a Senior Software Architect working within the Sustainability International Specialty Observe in AWS Skilled Providers. He has been with skilled providers for five years serving to prospects architect and construct IoT and sustainability functions on AWS. Previous to becoming a member of skilled providers John was a software program growth engineer who helped to launch the AWS IoT service.
Mike-Ply-2Mike Ply is a Affiliate Director of IoT at BISSELL, main groups centered on connecting IoT merchandise to offer thrilling and distinctive experiences. Earlier than becoming a member of BISSELL, he lead groups in healthcare to ship progressive digital options to enhance the well being of the neighborhood. He additionally labored in Automotive to outline a reusable software program structure for embedded gadgets. Mike is keen about bettering and innovating on digital experiences to simplify and enhance life’s on a regular basis duties.


Source link

Compare items
  • Total (0)