Scale of Epic Proportions
Customer success is the true measure of a software service like Workday, and our customers depend on our ability to scale to meet their needs. But what is scale exactly?
It can be defined in many ways, such as transactional throughput, database volume and read/write throughput, monthly active users, web service traffic, or data I/O. But what matters most is that a system can process the peak workload demand of a given customer at performance speeds that meet both the needs of the business and the expectations of individual users. And at Workday, we believe true scale is achieved if that service is capable of handling the most demanding workloads of every customer—all at once, every day.
Here are some key metrics that underscore the capabilities of the Workday service:
Total employees managed. Across our human capital management (HCM) and talent software services, Workday manages employee records for more than 120 million people including active (our current 55 million users) plus retired, contingent, and former workers.
Total journal lines. Across our cloud finance and accounting software services, Workday services more than 75 billion journal lines across our customers.
Peak hires. Within our recruiting and talent management software services, over 50,000 hire transactions were executed in Workday in a single day for a single customer. Globally, we’ve supported the hiring of six million employees in a single month.
Transactions. Workday processes approximately 365 billion transactions per year, and this is growing at a rate of 40% year over year.
Payroll processing. Workday’s largest payroll customers process payroll for more than 500,000 employees.
Peak workloads. Globally, Workday has serviced requests for more than 1 million users per hour and has easily supported one customer’s employee self-service event driving in 80,000 users per hour.
Workday Architecture: Built for Scale
From the beginning, we architected the Workday platform on a set of principles that could meet the needs of our customers and rapidly evolve over time as those needs changed.
Workday’s technology is built on a series of abstractions—this encompasses the principles of API-driven design, but done at the system level. This approach allows Workday developers to independently evolve each layer of the service without impacting the layers above or below. We achieve the abstraction between our applications and the underlying transaction systems by writing our application as metadata instead of the traditional model of hard-coded software. By doing so, we give customers the ability to configure the service. This is a subtle but incredibly important distinction, and it means that every one of our customers that is live—from small businesses to the most complex global enterprises—has a Workday deployment that uniquely meets their needs.
Each customer has done this through configuration of the application, not with customized application code. And because of our abstraction approach, we can continuously update the entirety of the Workday application for every customer, on a single platform, with all customers on the same version. This is what enables us to introduce fundamental changes to confidently scale the service—from massively parallelizing transactional throughput for the enterprise, to improving application code efficiency and performing seamless data conversions—week after week.
Not All Workloads Are Created Equal
Given the global nature of our customer base and their business operations, the sun never sets on the Workday service. Many customer workloads are seasonal. Payrolls can run weekly, biweekly, monthly, or semimonthly — or on demand. Employees check in and check out at the beginning and end of scheduled shifts. Financial planning may be run monthly (or more frequently), but books are closed quarterly and yearly. Global corporations drive all kinds of activity at all hours of the day.
The Workday service works to better understand each customer’s unique timing and type of peak workloads, allowing us to autoscale specific types of processing resources to meet time-varying demand. In addition to the horizontal scale-out of the Workday platform (which also delivers improved availability), we can also scale out differently for different workloads--with nodes optimized to handle transactions, reporting, search, queries, scheduled jobs, integrations, and so forth. This allows us to apply the right mix of resources at the right time for a particular customer's current and future needs.
People are not the only ones who depend on a performant system. Because Workday sits at the core of an enterprise’s business technology ecosystem, it is often integrated with other systems, third-party web service APIs, and cloud services within that ecosystem that rely on Workday to be performant. Increasingly, these system-to-system integrations and API calls are driving transactions at speeds and volumes that are nearly incomprehensible.
Our system’s design has evolved to meet growing demand from both people and machines, just as we planned. By making our systems highly horizontally scalable, we have been able to scale and grow along with the user populations and workloads of our customers. For example, our current largest customer processes 20 times the daily transactions of our largest customer five years ago.
Today, Workday is trusted to process more than 30 billion transactions a month for the most globally diverse and complex organizations on the planet. As a result of how we architected the platform, we have scaled well and we will continue to scale as our customers keep growing in the years to come. The same principles for rapid evolution that got us off the ground years ago are still with us, ensuring that we are well equipped to quickly adapt, and manage increasing scale and complexity well into the future.