The long road to a patent: Distributed search proves innovation comes in multiple forms
Thomson Reuters provides information so our customers can make informed decisions for their businesses and their clients; and the way our customers access that information is through our online products. Because of the vast amount of content we have available, we must both store all of that content and make it searchable for our customers.
The technology used up until the late 1990s to supply both the content storage and search abilities was based on large, expensive mainframe computers – including products like Westlaw and Checkpoint. This technology was proprietary in nature and had limits in its ability to scale to the needs of our business.
In 1999, these systems began to have stability problems. Not only could they no longer support the amount of content we had, but they also couldn’t sufficiently search our content nor allow for the concurrent usage our business required. In other words, the limitations of these legacy systems were limiting the growth of our online offerings.
Mark Bluhm, now senior vice president of Data Center and Infrastructure Services (DCIS) at Thomson Reuters, was selected to be part of a team that would review ideas to address the challenges. However, the ideas brought forward were based on the legacy systems and incremental in nature. Bluhm questioned the proposed solutions and their ability to scale to the needs of the business; so he was challenged to propose a solution himself.
Bluhm, who holds a PhD in computer science, envisioned a faster and more effective way to allow customers to search through our databases to find the exact information they need. He proposed a solution that used a vast amount of commodity servers that could exceed the ability of a single mainframe. Bluhm theorized that not only would this approach reduce the amount of time it takes to find a result, but it would also be able to grow and scale to a virtually unlimited amount of content while also providing more availability and stability.
Bluhm’s solution was based on emerging, and often, un-proven technologies. Everyone was skeptical including technology leadership, Bluhm’s boss, colleagues, and friends. Still, Bluhm believed in his idea. So much so that he convinced leadership to allow him to build a prototype on his own time. He worked nights and weekends, and when his prototype was complete – it proved the technology choices could achieve the result desired.
The beginnings of distributed search
Now that he convinced everyone, his idea had to be operationalized. So, Bluhm formed a small and extremely talented team to deliver the first search and content repository prototype for Westlaw.
Shortly thereafter, the largest database collection we had (public records) was running on this new system. Soon after, it was scaled to encompass many of our databases. This new system was named “Novus.”
This entirely new approach to search—what we now call “distributed search”—led to dramatically increased performance and decreased search time, and was a key building block that allowed us to create products like WestlawNext. It also allowed us to better leverage and share content across our products, forming the basis of the shared Novus platform. And, importantly, it reduced our reliance on expensive mainframe computers – a fact that transformed the way our data centers operate.
In April 2004—five years after Bluhm had the original idea for distributed search—a patent for the innovation was filed. In the subsequent 10 years, it withstood numerous legal challenges by other companies and Bluhm was the subject of a number of depositions.
Finally, in late 2014, Bluhm was awarded the patent for distributed search.
“This patent highlights how innovation is not just about new products. It can be about better infrastructures,” said Rick King, chief operations officer for Technology at Thomson Reuters. “This work was revolutionary and will save us millions of dollars per year. It freed us from mainframe and proprietary hardware and allowed us to create offerings that were inconceivable on the legacy platform.”
As for Bluhm, he has a few simple words of advice for others with big ideas: “Believe in your ideas. Surround yourself with like-minded and talented people. Sometimes the ideas that others believe can’t be done are the ones that are most important to pursue.”