Securing Pub/Sub Messaging: A Survey of Confidentiality-Preserving Solutions

In this article, we explore the impact of scaling publishers on message delivery latency in a distributed system. The authors use a simple pub/sub system with 60 subscribers and 5 messages per second from each publisher to test their hypothesis. They increase the number of publishers while keeping the number of subscribers constant, measuring the resulting impact on message delivery latency.
The study reveals that as the number of publishers increases, message delivery latency initially decreases but then stabilizes at a certain threshold. This behavior can be attributed to the increased competition for resources between publishers and the slower processing of messages in the system. The authors also observe that the impact of scaling on message delivery latency varies depending on the payload size of the messages being transmitted.
To better understand these findings, let’s consider an analogy. Imagine a busy train station with many platforms and trains arriving and departing simultaneously. As more trains arrive, the system becomes overwhelmed, leading to longer wait times for passengers. However, if the number of platforms is increased, the system can accommodate more trains, reducing wait times.
Similarly, in the pub/sub system, increasing the number of publishers can lead to congestion and slower message processing times as each publisher competes for resources. However, by adding more platforms (i.e., increasing the number of subscribers), the system can handle more messages and reduce latency.
In conclusion, scaling publishers in a pub/sub system can have complex effects on message delivery latency, depending on various factors such as payload size and the number of subscribers. By understanding these dynamics, designers and implementors can optimize their systems for better performance and reliability.