Building Reliable, Scalable, and Efficient Software with Open-Source

With Temporal, peace of mind is paramount. In addition to allowing for more efficient software delivery cycles, Temporal ensures users never have to worry about disruptive outages again.

Building Reliable, Scalable, and Efficient Software with Open-Source

With Temporal, peace of mind is paramount. In addition to allowing for more efficient software delivery cycles, Temporal ensures users never have to worry about disruptive outages again.

Building Reliable, Scalable, and Efficient Software with Open-Source

With Temporal, peace of mind is paramount. In addition to allowing for more efficient software delivery cycles, Temporal ensures users never have to worry about disruptive outages again.

interview with
Maxim Fateev

Co-Founder and CEO


In search of a stable and streamlined cloud applications

At first glance, Coinbase, Netflix, and Snap don’t seem to have much in common. Dig a little deeper, however, and it turns out they share something fundamental—all three companies rely on Temporal to operate and maintain reliable, streamlined cloud applications. 

Founded in 2019 by Maxim Fateev and Samar Abbas, Temporal is an open-source software company that helps companies build reliable, durable, scalable software—and keep their data workflows and processes running smoothly, even when the unexpected happens.

While Fateev grew up surrounded by journalists—his mother, father, and grandfather included—he preferred writing code. He can trace his interest in programming back to elementary school. When flipping through the pages of a popular science magazine, he spotted an article on how to program a calculator, and before long, he had figured out how to write the code on his own. 

He continued to pursue these interests throughout his school years. And in 2002, he took a software engineering position at Amazon, where he worked as the tech lead for a messaging infrastructure project that would eventually evolve into AWS’s Simple Queue Service (SQS), a solution that allows various software components to talk to each other stably and securely. Seven years later, Fateev worked as the tech lead for another landmark project: the first public version of AWS’s Simple Workflow Service (SWF), which later inspired Temporal.

Amazon is where Fateev met Samar Abbas, a software engineer who had recently joined the company from Microsoft. “We created the AWS’s Simple Workflow Service because we felt that it was a better way to build applications,” says Fateev. “Since we were at Amazon, we tried to make it very generic and scalable so it could be used as widely as possible.” 

After the Amazon SWF project, Fateev and Abbas took their careers in different directions—Abbas went on to join the Microsoft Azure team, where he created the open-source Durable Task Framework (which was later used as a foundation for Azure Durable Functions), and Fateev moved to the Google team that owned the company’s big data frameworks.

Fateev and Abbas reunited in 2015 when Uber opened a new office in Seattle.

There still wasn’t an approach to manage end-to-end procesess that could provide adequate scalability and reliability—and the need had only increased since they had worked together at AWS. Fateev and Abbas got to work and co-created the Cadence project, an open source, multi-tenant orchestration framework that’s scalable and reliable enough to handle millions of requests from Uber users all over the world.

“Cadence was based on the idea of what we did at Amazon, but it was a completely different implementation,” says Fateev.

“We released it as open source, because I believe that’s necessary in order for any infrastructure-level technology to be successful long-term. If it’s not, people won’t use it.”
No items found.
No items found.
Temporal Co-founders Samar Abbas (left) and Maxim Fateev (right)

Open Source at Uber

As Cadence picked up steam, Fateev and Abbas began to attract attention—a particularly unusual experience for Fateev, who tended to quietly pour himself into his creative and technical work. “I was not a start-up guy. I didn’t know any of the VCs' names. I never managed a single person in my life—I just worked as an engineer. We built Cadence as an open-source project because we believe it's useful for everyone, and we had a vague idea that maybe one day, if it was successful, we could make something else out of it. But then we started getting external adoption, and at one point our github repo was published on Hacker News.”

It was a top story and Fateev jumped into the comment section to answer questions about the technology, which drove significant traffic to their GitHub project. Then the VCs started calling. 

“Amplify was the first VC we talked to,” he says. “Over the next few months, we considered our options and realized that if we really wanted to make this technology successful and put it in the hands of every developer, creating a separate company would be the best option. We had Uber’s support, but we wanted to build this technology for the rest of the world.”

Temporal’s early experiences were quite different from what most start-ups encounter because of the buy-in they received so early on. This included Amplify leading the company’s $6.75M Seed in 2019, followed by a $18.75 million in Series A  in 2020. The product was an easy sell because it provides immense value and is unique. Also, Fateev and Abbas had already proven themselves with the success of their earlier versions of the model. AWS’s Simple Workflow Service and Azure’s Durable Functions were both widely adopted, and Uber’s Cadence grew to 100 use cases within Uber in just three years. 

“The hardest part of a start-up is finding that magical product market fit,” says Fateev. “We had it before we even started the company. Having large-scale production at Uber for multiple years helped because it gave companies a reason to trust and be interested in Temporal.” 

Temporal attracted the attention of some big brands, including HashiCorp, Airbnb, Box, Coinbase, and more—driven mainly by word-of-mouth. From there, adoption continued to increase organically. Now the company has a strong base across the United States and within growing markets in Europe, Asia, and Australia. 

“I'm very excited about the mix of companies using our product,” says Fateev. “In addition to digital native companies, traditional enterprises have successfully implemented solutions with us because we’re able to help them easily maintain very large systems, or migrate from mainframes to the cloud. We have a very good value proposition for them.”

No items found.
No items found.
Of course, customer trust, viability, and funding aren’t everything. To be successful, it’s critical for a product to solve a pain point for its target market. In Temporal’s case, the pain points relate to reliability and productivity.

While the average person would expect software to be mostly business logic, the bulk of code actually does the secondary job of safeguarding workflows—to prevent data loss if a machine crashes or a network outage occurs. Writing all of that code is complex and time-intensive work, and the cost of leaving out that code is unimaginable. 

Most people know the pain of having their computer crash after forgetting to save a document on their hard drive—but cloud-based software like Google Docs has tackled this problem by adding autosave features. Fateev says that Temporal solves the same problem, but for the state of any applications. “With Temporal, software engineers can just write code as infrastructure failures don’t exist,” he says. “We have a magic auto-save button for their code state.”


A built-in safety net for developers

No items found.
No items found.

Reliability and durability for all

“We put in a lot of effort to make Temporal the most scalable, reliable, and economically viable way to run applications at scale,” says Fateev. “It’s good for small companies, and it’s good for the largest companies in the world.”

Temporal is open source and freely available for anyone to deploy and use and the company also offers a consumption based managed service called Temporal Cloud. While many organizations choose to run Temporal themselves, others don’t want to spend time and labor on building and maintaining it if they don’t have to. By using Temporal Cloud, these companies can pay to use a pre-built infrastructure supported by highly experienced specialists. 

Temporal Cloud is consumption-based, so customers only pay for what they use. The company also maintains high security standards. Temporal Cloud is set up so that users run their own code, and while Temporal stores their data, it can be encrypted in advance so that no one at Temporal ever actually sees it.

“That means that even companies with very strict data security requirements can use Temporal safely,” says Fateev. “We pass their security checks with no problem because they know they’ll never need to send us their customer data in clear text. Because of our security and our consumption-based business model, our service can be used by almost any enterprise.”

Currently, Temporal is the only software that delivers these reliability capabilities for developers. There are similar solutions, but none provide this type of security and ease for engineers. “We have no direct competition,” says Fateev. “When we talk to enterprises, we’re not offering to replace their database with a better database or their queue with a better queue. We’re saying, ‘Okay, you’re undergoing this massive implementation, and you’re working with a unsustainable solution that’s causing all sorts of problems—but you can use Temporal to fix it all.’” 

Fateev says Temporal’s main value proposition is increased productivity, whether users are developing, maintaining, or running applications. The product also improves visibility, which makes troubleshooting significantly easier. In most cases, engineers never even need to speak to anyone at Temporal because the software allows them to easily identify problems on their own. Some enterprises reported 5x faster project completion with Temporal than their legacy approach.

The product is so user-friendly that developers with a wide range of experience levels can use Temporal. The product includes software development kits (SDKs) that support various programming languages, including Java, Go, TypeScript, PHP, .NET and Python, and Temporal is also working on a Ruby SDKs. “That’s pretty significant for some companies,” says Fateev. "Some shops might only use Python or TypeScript. So it was very important to meet developers where they are.”

No items found.
No items found.

Case study: Committed to community

It’s hard to pinpoint the exact size of the Temporal community because of its open-source nature; however, its dynamic and lively community slack channel paints the picture of a vital, growing group of users and followers. The developers who rely on Temporal for everyday tasks have been top-of-mind for Fateev since the very beginning. Trust and transparency are high priorities for Temporal, and the company always considers community impact when making business decisions. 

For example, the company has promised to maintain compatibility between the Temporal Cloud and any self-hosted instance. “If you’re running an application using a self-hosted instance of Temporal, you can also run it in the cloud and vice versa,” says Fateev. “If we ever add features that won’t work in open source, we’ll very clearly mark them. In the future, when we add the ability to migrate traffic from open source to cloud and back to open source without downtime, we don't want people to feel locked into our commercial offering.” 

Creation of a business around open-source software can be challenging because anyone can use the product free of charge. Successful open-source products, however, can become very valuable as word spreads and the user base grows. In Temporal’s case, the company was able to build a world-class, successful SaaS offering forf its open-source project. The more popular the Temporal project becomes, the more people will use the paid Temporal Cloud service.

“We had no marketing at all, and we still attracted all this adoption,” says Fateev. “But it took a very long time. When you choose open source, you can’t expect a huge community in six months. It took us six years, and we now have over 8,000 people in our growing community Slack channel. If you can go the route we took, where we built a product at Uber and then created a separate company, it's awesome. But from another point of view, we might not have been able to execute it if that was our plan from the beginning.”

Fateev credits at least part of Temporal’s success to the fact that their initial goal was simply to build a good, developer-friendly product. “Don't focus on the company, don't focus on monetization,” he says. “Focus on creating something that people value. I think what I love about America is that if you have something that people care about and are willing to use, you'll find ways to make money out of it. Look at Google.”

 Fateev still makes himself available to answer questions from engineers—but now, he’s got some extra help. To maintain a high quality of support as the company continued to grow, Temporal hired dedicated customer success representatives to assist developers using the open source framework. 

Fateev considers community feedback one of the most rewarding parts of his job. “I think we have the most positive communities you can find anywhere,” says Fateev. “People just are super excited to be part of our community, and they're very grateful for Temporal. We get a lot of unsolicited  ‘Thank yous.’”

No items found.
No items found.
Temporal's Replay Conference (left) and Temporal Meetup in Singapore (right)

Aiming for ubiquity

Looking back, Fateev is grateful that he decided to launch Temporal as a company. “There was so much potential,” he says. “I wanted to know what would happen if I stayed a little bit longer to see where it went, and it looks like that was a very smart decision. The nicest thing about Temporal is that the technology is still new enough that I’m able to invent things every day.”

If its high retention rates are any indicator, Temporal’s revolutionary work appeals to the company’s workforce, too. “I’m super grateful for the team we’ve built at Temporal,” he says. “We have a high quality of talent at all levels, and we certainly will put a lot of effort to keep it that way.”

Looking forward, Temporal’s goal is simple: focus on continual improvement, from security to approachability. Ultimately, Fateev believes that people will look at Temporal as a standard industry tool. “We’ll be successful when every developer knows about Temporal,” he says. “When we reach the point where it’s considered every time someone’s building an application, when it’s taught at schools and boot camps as part of the curriculum. When it becomes ubiquitous, we've won.” 

No items found.
No items found.
Temporal Company Offsite

No items found.
No items found.

No items found.
No items found.

No items found.
No items found.