Many of the world’s largest and most complex organizations are Workday customers, with more joining our community all the time. That’s why our technologists are continuously focused on advancing the scalability and availability of our technology infrastructure.
Among them are Noah Arliss and Jason Howes. They are experts in the field of distributed computing, one of the many technologies that Workday uses to help evolve its infrastructure. Distributed computing brings multiple computing resources together over a shared network to do work that is too large for a single machine, such as running large-scale applications that manage many operations in parallel. The concept itself is over 30 years old but it’s being used in new ways, serving as the foundation for important innovations like the Hadoop open-source software framework.
Arliss and Howes, who joined Workday in November 2014, have one goal in mind: Grow a team tasked with uncovering new ways to extend the benefits of distributed computing for Workday and our customers (including our goal to reach zero downtime for Workday releases). We spoke with them to learn more about this technology and plans for their team based out of Workday’s Boston office.
How did you get interested in distributed computing?
Howes: Back in middle school, I was into bulletin board systems, which existed before the Internet. They let people connect their personal computers to servers using modems and establish communities where we played games and wrote on forums. The notion that you could communicate with people you never met through this invisible ether fascinated me. My curiosity drove me to study computer science at Cornell where I learned how computers communicate, and how to make them work together to solve problems using distributed algorithms.
Arliss: When I was in college at Brandeis University studying computer science I had a project to create a Beowulf Cluster―essentially building a cluster of machines to run tasks. So, a lot like Jason, the idea of connecting computers together and getting them to communicate was interesting to me.
Can you talk about how we’re using distributed computing at Workday?
Arliss: When we got here, Workday was already delivering system availability levels of at least 99.5% and effectively scaling to meet the needs of the world’s biggest customers with the help of distributed computing. That’s tough to beat. Yet we’ve taken on the challenge of finding new ways to employ distributed computing techniques to further evolve Workday’s architecture and ensure we are always delivering the most available and scalable infrastructure possible.
Some of the new methods include replicating all in-memory data, which increases availability and enables us to respond more quickly to resolve production issues. Distributed computing also lets us store data in a partitionable way so that we can simply add more hardware to handle more data or requests. Say for example a customer has 100GB of data today but expects to soon reach 1TB of data; distributed computing allows us to easily scale by adding more machines until we have the memory required for that full terabyte.
Where do you see distributed computing headed?
Howes: If you look at the software industry, you’ll notice several movements that are converging on a common theme: Software that has high availability and dynamic scalability built into its architecture from day one. You can see this movement in data fabrics (Ignite), big data processors (Hadoop), and NOSQL databases (Cassandra), all of which embrace parallel programming, eventual consistency, and event-driven architectures. These tools and concepts are paradigm shifts from traditional programming models but over time they will become standard elements of a developer’s toolset.
We’re also seeing that a lot of other industries beyond software are realizing how distributed computing allows them to do things they weren’t able to do before, thanks to its ability to handle thousands of requests per second in real time or make it fast and easy to query massive amounts of data. This is allowing major financial institutions to build better matching engines, and online retailers to scale on demand to better serve their customers.
The two of you were friends and colleagues before coming to Workday. What inspired you both to make the move?
Howes: I’ve been friends since college with a couple of engineers who now work at Workday, and hearing them talk about the types of technical challenges they were solving really intrigued me. Add in the idea of growing a team and it was the perfect opportunity for me. But I knew I couldn’t do it alone, which is why I wanted to bring in Noah.
Arliss: When Jason approached me the timing couldn’t have been better. To be able to come to a company that is clearly an innovation leader in the cloud, is achieving results, and doing it at scale was really appealing. Throw in Jason’s technical acumen and the opportunity to evolve the team and take this effort to new heights, and it was a no-brainer for me.
“We are growing an elite team of systems engineers who get to work on some of the most challenging distributed computing projects in the industry.”
You both work out of Workday’s Boston office. What’s that like, and how do you see your team there evolving over time?
Arliss: The office has a great vibe and is right in the heart of the financial district. It has a startup feel but with the benefit of being part of a large, established, and growing company. As engineers, we really appreciate that the entire Boston team is together on the same floor. This means we have direct access to our sales team and hear first-hand about our customers and prospective customers’ business pain points. This is incredibly valuable from an engineering standpoint. We’re not up in some ivory tower of academics—we’re hearing real-world customer challenges, and that’s huge for us as we’re writing code.
We are growing an elite team of systems engineers who get to work on some of the most challenging distributed computing projects in the industry. We want passionate, energetic people who love to learn and are willing to roll their sleeves up, dig in, and figure it out. It’s an important charter, and management has given us the support to do what we need to do. This allows us to try new things and take smart risks to drive innovation. I couldn’t ask for a better situation.