BreezoMeter: Delivering global environmental information with Google Cloud
About BreezoMeter
Founded in 2014, covering more than 100 countries, BreezoMeter provides real-time, dynamic, location-based environmental data on air pollution, pollen, wildfires, and the weather.
Tell us your challenge. We're here to help.
Contact usAbout More by SELA
More by SELA is a leading Google Cloud partner helping companies with their architectural, infrastructure, development, and automation work.
BreezoMeter uses Google Cloud to power its comprehensive environmental intelligence platform, helping to make the invisible visible, and mitigate the effects of air pollution worldwide.
Google Cloud results:
- Analyzes 1.8 TB and calculates ~10 billion of pollutant concentrations each hour on Google Cloud continually for accurate, up-to-date, hyperlocal environmental insight
- Easily provides API services with Google Cloud PaaS
- Democratizes data with BigQuery, allowing employees across the organization to extrapolate meaningful information
- Maximizes efficiency by only choosing the most appropriate resources for as long as they are needed
Measuring hourly air quality on 500 million locations
In 2014, a small Israeli startup came up with a very big plan. BreezoMeter wanted to provide actionable information on air pollution, an environmental hazard that is correlated with more than 8 million casualties per year. “We wanted to make this invisible killer visible with data and use those insights to make life better for the 80% of people who are affected by air pollution across the world,” says Emil Fisher, CTO and co-founder at BreezoMeter. “At first, people laughed at us. They didn’t understand what we were trying to do.”
Undaunted, Fisher, co-founder and CEO Ran Korber, and colleagues set about building a product that would take in data from a huge variety of sources to produce a data service that tracked air quality across the world. Eight years later, the mission to improve the health and safety of billions has stayed the same, but the reactions have changed. Now, instead of asking why they should use BreezoMeter’s data, organizations across the world are asking how they should use BreezoMeter’s data for their own products and use cases. Today, BreezoMeter captures air pollution, pollen, and fire data in more than 100 countries, covering billions of people.
“Right from the start, we knew that scalability and stability were key for our vision. Google Cloud gave us access to the vast resources we needed, but it also gave us the ability to power down those resources when they’re not needed, helping us to stay energy efficient.”
—Emil Fisher, CTO and co-founder, BreezoMeterThe company has built a suite of data products that add up to a comprehensive environmental intelligence platform. Striving to provide the most accurate and localized environmental data in the world, BreezoMeter works at the largest possible scale, advising government agencies, as well as multinational industry leaders. The company’s data is in every iPhone embedded into mapping and weather apps. Healthcare providers use its data to predict pollen and pollution levels in certain neighborhoods. Heating, vacuum, and air-conditioning (HVAC) companies use it to regulate their products and cut down on energy waste. Simply put, BreezoMeter is for any industry and any use case.
Commercial success is one thing, but it’s not the end goal. Since BreezoMeter’s launch, sustainability has taken on a new urgency, rising to the top of the global agenda for public and private institutions. Google CEO Sundar Pichai recently spoke about the ways in which Google is tackling climate change and why there are reasons to be hopeful. “We’ve seen a tectonic shift towards sustainability issues. That speech was a good example of that,” says Fisher. “I’m proud to be part of a movement that educates people about the environment.”
Sustainability is not an abstract ideal for BreezoMeter, it is embedded into every facet of the company from its wider mission, to the products it makes, right down to the technology it uses. At its inception, BreezoMeter chose Google Cloud to power its infrastructure, and the relationship has only strengthened since then.
“Right from the start, we knew that scalability and stability were key for our vision,” says Fisher. “Google Cloud gave us access to the vast resources we needed, but it also gave us the ability to power down those resources when they’re not needed, helping us to stay energy efficient.”
“BigQuery allows us to do things we couldn’t do before. We load and analyze ten terabytes a day right now, but it’s not just the scale, it’s also the support for new kinds of information, such as geographical data, which is very important for our latest products.”
—Uri Hellerman, Software Architect, BreezoMeterData analytics at speed and at scale
At its core, BreezoMeter’s platform takes in a huge amount of environmental data from a range of sources, applies sophisticated algorithms, and provides a set of APIs covering air quality, pollen, wildfires, and weather for a truly unified environmental data solution. These APIs handle millions of requests every hour for clients large and small across the globe. But air quality is not a simple thing to measure and it changes constantly. “Air quality is very dynamic,” says Uri Hellerman, Software Architect at BreezoMeter. “We have to recalculate the air quality for 500 million geographical locations around the world every single hour.”
With scalability and stability in mind, BreezoMeter built its platform with App Engine supported by Compute Engine virtual machines and Cloud Storage buckets. By choosing to be an early adopter of managed cloud services, the company avoided the costs of installing and maintaining infrastructure and allowed its developers to focus on building the best platform they could.
Over time, BreezoMeter evolved its infrastructure with the help of Google Cloud partner More by SELA, incorporating more automation with Cloud Functions and Cloud Run. But eventually, BreezoMeter grew so big that it strained against the limits of its existing setup. “We loved App Engine for years, it meant we didn’t have to hire DevOps specialists, but we had scaled to the point beyond what it was designed for, so we needed something different,” says Hellerman.
The next phase of BreezoMeter’s infrastructure was built around microservices orchestrated with Google Kubernetes Engine (GKE). This new architecture enables the company to scale up and down quicker than before, maximizing power while minimizing wasted resources. “We need to schedule bigger and more loads all the time. We use thousands of CPUs every hour, and that number will only go up,” says Hellerman. “Google Kubernetes Engine makes it easier and more cost-effective for us at the scale we want to work at.”
The most recent evolution has seen BreezoMeter’s move of its data platform to BigQuery, for analytics at unprecedented scale and speed. “BigQuery allows us to do things we couldn’t do before,” explains Hellerman. “We load and analyze ten terabytes a day right now but it’s not just the scale, it’s also the support for new kinds of information, such as geographical data, which is very important for our latest products.”
“We run billions of calculations every hour on Google Cloud, and we’re still growing, incorporating new kinds of data and building new products for our customers. We have gone from providing them with data to providing them with relevant actionable insights.”
—Emil Fisher, CTO and co-founder, BreezoMeterPower and control for practical sustainability
With Google Cloud at its core, BreezoMeter has continually adapted to and overcome every hurdle in its path, growing its customer base and providing more solutions for them. “We run billions of calculations every hour on Google Cloud, and we’re still growing, incorporating new kinds of data, and building new products for our customers,” declares Fisher. “We have gone from providing them with data to providing them with urgent, actionable insights.”
The sheer scale at which BigQuery operates has transformed the way BreezoMeter handles its data as well. Previously, captured data would be placed into Cloud Storage buckets soon after analysis to make way for new data, making historical analysis cumbersome. Now, the company simply stores its data in BigQuery, where it can be analyzed at will. “Using BigQuery as our data lake has democratized our data. It’s much easier to extract meaningful information than before,” says Fisher. “Anyone who knows how to write a SQL query can see the data and go to work on it.”
Working at scale is one thing, but Google Cloud helps BreezoMeter’s developers to work smart with managed services taking care of infrastructure operations while they concentrate on the core business challenges. They can maximize efficiency with innovations such as preemptible virtual machines to automatically shut down instances once their task is done, or use the latest, most powerful CPUs for faster calculations. Having this level of control not only helps the company save on costs, but also embodies the sustainable values it wants to spread.
Like any successful, sustainable organism, BreezoMeter never stops evolving and optimizing. After eight years, it is keen to develop more and better tools with Google Cloud in the years to come, not just because of its technical excellence, but because of shared values. “Our platform requires vast resources to run with thousands of machines in the cloud, so there has to be a very high standard of service,” explains Fisher. “But sustainability is non-negotiable for us. The commitments we’ve seen recently from Google Cloud, such as operating carbon-free by 2030, reassure us that we’re working with the right partner.”
Tell us your challenge. We're here to help.
Contact usAbout BreezoMeter
Founded in 2014, covering more than 100 countries, BreezoMeter provides real-time, dynamic, location-based environmental data on air pollution, pollen, wildfires, and the weather.
About More by SELA
More by SELA is a leading Google Cloud partner helping companies with their architectural, infrastructure, development, and automation work.