In-Depth

Profile: RTI Distributes the Data

Even a small company, such as Real Time Innovations (RTI), can turn a product into an industry standard.

• By Peter Varhol
• 09/01/2006

"We're geeks selling to geeks," remarked Stan Schneider, CEO of Real Time Innovations, Inc. (RTI), during a discussion about his company and business with FTPOnline.

While marketers might cringe at this characterization of a company and its products, RTI thrives on the notion that talented scientists and engineers can build software that provides unique value to those seeking to integrate data across large distributed networks. The company's primary product, RTI Data Distribution Service, offers a mechanism for using the publish-subscribe model to ensure that data gets from its sources to one or more destinations on large networks. This product can form the backbone of a distributed, data-driven application.

RTI was spun from a Stanford University robotics research group. Schneider had an academic relationship with Stanford—having managed one of its largest laboratories—that focused on intelligent mechanical systems. The company began producing tools for embedded software development. RTI's Scope product offered a way of analyzing the inner workings of embedded applications, and more importantly, provided a steady revenue stream for funding future product development.

Optimizing the Publish-Subscribe Model
RTI soon expanded into realtime data distribution using the publish-subscribe model. Moving asynchronous data between processes is a common challenge in software, and academic courses in computer science spend lots of time on the producer-consumer problem. Traditionally, you only have a couple of options for moving asynchronous data from one process to another. The most common solution is using the sending process to pipe the data into a queue, and the receiving process to poll for any new data in the queue. While this solution works, and you don't have a lot of other ways to move asynchronous data transfers, polling is an offensive strategy for software engineers because it isn't efficient. The polling model becomes more inefficient when you have multiple data crisscrossing the network, with multiple producers and consumers.

An emerging alternative for asynchronous data transfers is the publish-subscribe model, often shortened to "pub-sub." In this model, data sources, or producers, publish data to a known location on the network. For example, you could complete a memory-to-memory transfer, if high performance is required, or move data to your database or other persistent storage.

In pub-sub, these types of processes, with a need for data, can subscribe to the data through the messaging service. When published data arrives at the shared location, a message is sent to the subscribers. Subscribers can then obtain the data from the shared location and use it in their own processing. The pub-sub model improves upon the polling model, which frequently wastes processing cycles on performing null activity.

In addition optimizing the publish-subscribe model for use on large-scale networks, RTI Data Distribution Service also places unique emphasis on realtime performance—which refers not only to fast performance, but more importantly to predictable performance. This enables the product to accept and distribute data from devices, such as process monitors or data acquisition instruments that generate data at fast but predictable intervals. Why is this important? Increasingly, you need to make business decisions based on data that's coming directly from the factory or manufacturing plant. The ability to have up-to-the-minute data from the factory floor for business analysis is a goal of many enterprises today.

The Importance of Open Standards
RTI went a step further. Rather than offer up the RTI Data Distribution Service as a proprietary middleware solution, it decided to join a standards body and promote its technology as an open standard. In this case, the standards body was the Object Management Group (OMG), an organization with dozens of industry members that standardizes and supports a variety of software technologies.

What was the advantage in opening up its technology to other companies? Schneider explained, "We wanted to go to customers with a standard solution, rather than just another proprietary product. We wanted to give them the confidence of being able to get the same technology from multiple sources. There are others who offer proprietary solutions, and we didn't want to be thought of as just another proprietary vendor."

One competitor has adopted the RTI Data Distribution Service standard and offers a competing product. Schneider notes that by maintaining leadership in setting the standard, RTI keeps its competitive advantage. He said, "It's especially important in the markets we serve that such a solution comes as a standard."

He continued, "Government, defense, and others investing in a middleware solution all want the assurance that they won't be held hostage by using a proprietary product." Being a standard changed the rules of competition, but RTI knew that becoming the standard was more important than maintaining a proprietary approach.

Expanding on the Data Distributed Service
Recent RTI offerings expand the reach and capability of the data distribution system software. These offerings also include a data persistence product, based on the Oracle TimesTen in-memory database and offering realtime data storage and retrieval. Connectors to the enterprise Oracle database let you aggregate and analyze realtime data as a part of the enterprise data management environment.

RTI is also providing a better platform for the development of data-driven distributed applications. Using a development platform that includes application analysis, debugging, and UML modeling through a partnership with Sparx Systems, the company is providing you with a set of tools for more easily creating applications that move data between multiple nodes on a network.

RTI funds its growth organically, through its initial product offerings, so it has flexibility on how it defines its future. But clearly, data-driven distributed applications are increasingly important to enterprises. Any future that incorporates ways to more easily create this type of application is likely to be both in demand and profitable.

Real Time Innovations, Inc.
3975 Freedom Circle
Santa Clara, CA 95054
www.rti.com