Scalability
Specific to market data needs, InfoDyne’s TPS+Plus architecture offers a highly flexible and easily scalable set of exchange servers, object distributors, and application programming interfaces necessary to address the integration and performance demands of our customers. Since all TPS+Plus components are multi-threaded and mutexed to a high degree of granularity, they are capable of taking full advantage of multiprocessor hardware architectures and operating systems. What’s more, throughout the system, scaling takes place independently and dynamically by simply adding additional component instances while the system is running. This is also useful when upgrading a system since new releases are guaranteed to be backwards compatible allowing previous individual instances to be taken down and current releases to be brought back up dynamically.
With additional hardware and networking capacity, there is virtually no limit regarding the number of messages per second possible within a TPS+Plus system. Even with soaring messages rates and ever increasing demands for data consumption, TPS+Plus retains its ability to provide high throughput, high availability, and low latency market data. This is accomplished through service scaling (see below), load balancing, and single name space which allows the system to be deployed in a distributed manner and yet hide this scaling detail from subscribing applications.
However, with these soaring messages rates and ever increasing demands for data consumption, our customers can take advantage of the economical and technical solutions available in TPS+Plus for maintaining high throughput, high quality, and low latency market data. While the optimized multi-threaded design of TPS+Plus takes full advantage of multi-processor servers, the expected increases in message rates from data feeds, such as OPRA, may ultimately exceed the scalability and performance of even the highest performing platforms.
Service Scaling
Service Scaling allows users of InfoDyne’s TPS+Plus ticker plant technology to distribute processing for individual feeds horizontally across any number of physical server devices while preserving complete system functionality. Each server instance processes a portion of a given feed, and TPS+Plus middleware ensures that user applications continue to see each feed as a single, unified, service on the system without introducing any additional message hops.
In anticipation of expanding message rates, InfoDyne is raising the bar to insure that TPS+Plus can be scaled economically to provide the same easy access and high performance our clients have come to rely on. Customers should not have to replace their hardware, re-write their applications, or compromise on data quality because message rates are going through the roof. It’s always less expensive to simply add another box.
More than simple load-balancing which most vendor platforms support, this partitioned service allows InfoDyne to assign data feed processing to an array of servers as necessary to keep pace with the growing message rates. TPS+Plus middleware intelligently resolves requests on behalf of client applications and nothing is lost relative to quality of service. Most importantly, there is zero added latency. We also retain data blending, data models, option grids, chaining, underlying references and other capabilities that are simplified within a single server instance, but become complex when distributed across an array of servers.